ArALManertachnlictfgehc<itnbuacre/>l

of
Age
the
Iin

RIBA#' Phil Bernstein

© Phil Bernstein, 2022 Commissioning Editor: Clare Holloway
Published by RIBA Publishing, 66 Portland Place, London, W1B 1AD Assistant Editor: Scarlet Furness
ISBN 9781 85946 401 3 Production: Richard Blackburn
The right of Phil Bernstein to be identified as theAuthor of this Work Interiors designed and typeset by Studio Kalinka
has been asserted in accordance with the Copyright, Designs and Printed and bound by TJ Books, Cornwall
Patents Act 1988 sections 77 and 78.
All rights reserved. No part of this publication
Cover design: The First 47
may be reproduced,
stored in a retrieval system, or transmitted, in While every effort has been made to check the accuracy and quality
any form or by
of the information given in this publication, neither the Author nor
any means, electronic, mechanical, photocopying, recording or
otherwise, without prior permission of the copyright owner. the Publisher accept any responsibility for the subsequent use of
this information, for any errors or omissions that it may contain, or
British Library Cataloguing-in-Publication Data<br/> for any misunderstandings arising from it.
A catalogue record for this book is available from the British Library. www.ribapublishing.com

DOI: 10.4324/9781003297192
 ACKNOWLEDGEMENTS

RIBA reached out to ask whether I might be interested in writing a

When book about the implications of artificial
intelligence for the architecture
profession, I was surprised to learn that another far-better known pundit
(who shall remain unnamed) had declined a similar offer; apparently, the
prospects, to him, were simply too dire. However, as the science-fiction writer
William Gibson, of The Difference Engine fame, is purported to have said, 'the
future is already here, it's just not evenly distributed'. Such is the case with the
emergent technologies of Al, which lives on the smartphones (and thereby
in the pockets) of the same architects who will be made redundant by our
erstwhile robotic overlords.

What I did not realise as I started the project was that it would require me

to return to my computational roots from decades ago, when I was an

undergraduate student of an iconoclastic young professor at Yale, Roger

Schank, who had some of the earliest foundational ideas for teaching

computers to understand language. During my graduate studies, computing

was emergent but largely unacknowledged, and my instructor at the time,

Bob Frew, indulged my interest in using early computers (which were hardly
up to any useful graphic task) for project management. Late in my final
graduate year while on vacation in California, my uncle Edward Bernstein,
always one to have the latest gadget, gave me a week with his brand-new
Apple II Plus, on which was a new piece of software called VisiCalc, the first
spreadsheet. Suddenly, numeric modelling no longer required hard coding
and I surprised Bob with a finished project right after the holiday. Two years
later in San Francisco, the architect Plerb McLaughlin sent me several times to
Palo Alto, to research a new technology called 'expert systems', a provocative
but otherwise completely unrealisable technology. Around the same time, I
bought his firm's first personal computer not for generating drawings, but
-

to manage schedules and fee proposals. Computing to generate forms and

images, I came to realise, is irresistible, but digital tools have just as much, if
not more, agency in architectural process outside of design itself.

After dozen years in mainstream practice with César Pelli and his managing
a

partner, Fred Clarke, I joined Autodesk as a vice-president, and established

there many of the relationships in the technology world that I maintain today.
Jim Lynch, now a senior vice-president, is a close friend, sounding board and
Machine Learning

source of much and insight from the company. Sam Omans, now an
help
industry manager with Autodesk's Architecture/Engineering/Construction
business (and likely one of the few folks working in tech with a PhD in
architectural theory) has helped clarify ideas, locate information with the
Autodesk labyrinth and chase down critical images that illuminate the text.
Grace Liu, from the Autodesk Intellectual
Property team, was invaluable in
completing all the necessary image permissions there.

My thinking about the current state of Al technology owes a debt to

Mark Greaves of the Pacific Northwest National Laboratory, who has also
contributed the Foreword. During autumn 2020 I often met Mark, Steve
McConnell of NBBJ and the writer Cliff Pearson via Zoom. Our regular Al
salons, as we called them, replete with just a splash of scotch, did much to
clarify for me what can be the daunting trajectory of the development of

intelligent machines.

Here at Yale today I came to rely on our able architectural librarian in the Hass
Arts Library, Colwell, without whom I would have been unable to navigate
Tess
the university's vast, but often opaque, resources. Dean Deborah Berke has
been unfailingly supportive as I laboured to complete the manuscript while
we steered the School of Architecture through the global pandemic. And my
editors at RIBA Publishing -

ClareHolloway, Scarlet Furness, Liz Webster,

Richard Blackburn and Ramona Lamport are exemplars of professionalism
-

and patience, particularly with my peripatetic schedule. Clare was not fazed a
bit when I asked for my long-time friend and collaborator, the editor Andreas
Muller, to pitch in as an additional set of eyes on the project. Andreas, who
edited my last manuscript, offered regular, clear and very useful advice to
improve both the flow and logic of the argument.

A final note of thanks to my partner Nancy Alexander, who always creates

the space in which to do important work and the reminders for why it is so
important to stick to it. Without her unending support, none of these projects
would ever be possible.
 FOREWORD

I first started working with Phil, I held the one-dimensional impression

When of architecture that he mentions early in the book: architecture as a
discipline that translates desire and capital
occupiable space. Although
into
truthful and concise, this formulation makes it disturbingly easy for computer
scientists like me to view architects as people who mechanically execute a
semi-formal translation function. It took several Zoom-mediated and whiskey-lubricated
discussions of architecture and Al for Phil and his patient colleagues to
gently disabuse me of this blinkered and reductionist view.

Thechange in my own conception of architecture is not unlike the recent and

dramatic evolution of Al, which is why this book is so timely for both of our
professions. In the last decade, Al has rapidly advanced from meticulously
authored rule systems to the staggeringly complex world of deep learning
networks and self-supervised methods. Instead of relying on collections of
intricate rules
manually programmed for specific tasks, modern machine learning
systems now base their outputs on impenetrably complex patterns that result
from automatically analysing massive data sets. Al-generated design, which was
once mostly an academic exercise in combining rules in different ways, can now

produce creations that are far more subtle and compelling.

The need to reflect on architecture in the age of Al is therefore much more

acute than it was even five years ago. Although machine learning will not make
professional architects obsolete, neither will it have zero impact. The combination
of Phil's deep knowledge of the architectural profession and substantial Al chops
allows him to investigate the space between these two extremes and explore
how modern Al can affect the intricate information structures that underlie
the delivery of a constructed building. Phil also grapples with the elephant in
the room: can Al ever adequately comprehend the deeply human context of

places, replicate the architect's unique blend of formalism and creativity, and be
responsible for the safety and fitness of a building? To address this, Phil's analysis
goes beyond the usual reductionist critique and considers how modern Al could
not only make the overall value chain of architecture faster or more efficient,
but also result in a stronger architectural profession overall. I find this book
tremendously thought-provoking, and I hope you do as well.

Mark Greaves, Mercer Island, WA

vi MACHINE LEARNING

INTRODUCTION

A lmost half a century ago I was an undergraduate-aspiring-architect

studying at a school where the design studio was reserved for the elders
while the rest of us fulfilled our general requirements. An early interest
in computers had led me to a lower-level course called 'Introduction to
Computer Graphics', where, according to the syllabus, we were to spend the
first three drab weeks programming oscilloscopes to draw circles; exercises in
early computer-aided design (CAD) on the first graphic displays - which were,
of course, about 6 in wide, as seen in Figure 0.1.

Deciding that slog was not for me, I stumbled upon an exotic class in the
same department called 'Natural Language Processing', where, apparently,
we were going to teach computers to understand English. Our avuncular - if
prickly - professor, Roger Schank, explained that he had uncovered one of
the fundamental aspects of human existence by discovering the structure of
language understanding encoded within the mind. Our job was to translate
that theory into computer code. Of course, we were doing so on a 16-bit
predecessor of the IBM PC called a PDP-11 /45, with a whopping 256 kilobytes
of main memory. One afternoon, in the computer lab, with ten or more of us
working on the system, it burst into flames. 1

0.1:
A CATHODE RAV
OSCILLOSCOPE,
G. 1996, NOT
MUCH CHANGED
FROM ITS 1975
PREDECESSOR
(COURTESY
OF MAKEHAVEN
INC.)
 Introduction

0.2:
AIM EARLIER
VERSION OF
THE PDP-11
COMPUTER,
NOT ON FIRE 2

My project for the semester was to write a program that would accept input
from what was then called a 'newswire' –

a text streaming service derived from

teletype that delivered news from national sources –

about the particular topic

of oil tanker crashes and resulting spills, and then answer simple questions
about the same. Our solutions essentially 'hard-coded' the extraction of
meaning from English sentences and built semantic structures from which
the machine could perform what we thought was inferential reasoning, all
based on the thesis that we were digitising the thought processes also used by
humans. My solution, while adequate, was no early version of Google. It also
became clear by the end of the term that my days in the computer science
department were numbered, as the professor correctly observed that my
particular solution could not differentiate between a tanker accident and a car
crash. I beat a rapid retreat to the entirely analogue architecture curriculum
the next semester.
Machine Learning

Like my early academic experience, architects have always had an ambivalent

relationship with computation. In the early days of computer-aided design,
we worried that the ancient art of drawing would be lost. As the industry
3

transitioned in the 201 Os to 3D digital representation, and particularly building

information modelling (BIM), anxieties increased that tools would diminish
the designer's authorial agency when digital building components became
standardised and dispensed with ease into our designs by increasingly
powerful tools. Now that computation has become ubiquitous and machines
are learning to perform knowledge work through artificial intelligence, the

profession believes itself once again under threat. While the University of
Oxford's Richard and Daniel Susskind, exploring the implications of artificial
intelligence on the work of professionals, have suggested that 'we will neither
need nor want professionals to work in the way that they did in the twentieth
4
century and before', there is very little computational intelligence impinging
on the practice of architecture today. Yet we worry. Are we headed to a world

where computers largely replace architects altogether? Surely there are

reasons that this will not be the case?

Here is one example. When computers design buildings, the public will lose
out. Why? Because when an architect imagines a building, it is from its base
upwards. What you see is an imagined image of what is being created from bottom
to top. A real building shows us a holistic structure that starts from the very

foundations and works all the way to the top, from foundation to roof, and so on.
If we do not know the foundations of the building, we do not know what we are
being sold, and we will not understand what we are looking at. The bottom line is
that there is little or no aesthetic value in a building that has no soul, and so the
negative view of digital architecture is not justified.

Why? Because when architect imagines a building, it is from its base upwards.
What you imagined image of what is being created from bottom to top.
see is an

A real building shows us a holistic structure that starts from the very foundations
and works all the way to the top, from roof to foundation and so on. If we don't
know the foundations of the building, we don't know what we're being sold, and
we won't understand what we're looking at. The bottom line is that there is little or
no aesthetic value in a building that has no soul and so the negative view of digital

architecture is not justified.

I see two reasons why digital architecture is lacking aesthetics. First, it is solely
a technical process. If you try to make the process of creating a building into

something beautiful, a lot of work will go into that, and you won't save much in
time. And second, digital architecture uses software as the sole element in the
building process, which we know is unreliable.
Software is so far from being easy
to understand, we wouldn't recognise
even software. We can't see a word of
it as
the message it's trying to send. So, it's not easy to tell from looking at a building
whether it was designed by a computer or not. The work of design is not something
we can simply throw into the computer and have it turn out like a BMW.

Digital architects try to make the process of creating buildings into something
beautiful, but ultimately they don't deliver. There are other issues with digital
architecture. To make the building more "real", you need software to manipulate
the material the building will be made from. In the past, materials were never
digitally designed, so when a digital architect took the material out of a computer,
the software had no idea how to work with it, and the result was a "coffin" with
"free" space at the base. Digital architecture only takes you to the model first, and
once it's built, it will only keep it in the model, and create something completely

different with it. Digital architecture can give you better clarity, but at the cost of
aesthetic. It has never produced anything that looks like a real building.

Here's dear reader, that you're still with me after that brief italicised
hoping,
diversion, which was generated entirely (and unedited by me) with a
technology called GPT-3, the third generation of what is called a Generative
Pre-Trained Transformer, an artificial
intelligence (Al) system that conjures
original text in response to a prompt, which in this case was '(w)hen

computers design buildings, the public will lose out'. 5 GPT-3 is my oil tanker
program run in reverse –

instead of merely absorbing a narrative to somehow

understand, it creates one. It runs on computers orders of magnitude more

powerful that were programmed by much more talented people.

The sample output above has an uncanny similarity to coherent English, the
broad framework of an 'original' argument and even cites concrete examples.
It has the maddening quality of bare intelligibility, as the more you think you
understand it, the more obscure it becomes. GPT-3 technology is the current
generation of machine intelligence that'learns' language by ingesting huge
amounts of text from the internet and 'teaches' itself underlying semantic
structures. This is the same strategy that the mostly reliable Google Translate
uses to translate a web page from English to Spanish, but greatly accelerated

by rapidly advancing machine learning (ML), Yet both my early effort and that
of GPT-3, above, lack real coherence, and the more time spent reading the
text above, the less sense it seems to make. GPT-3 is certainly a more efficient

approach than mine of 1976, but without the sweeping philosophical assertions.
Yet, the possibilities here are intriguing. Could a computer design an entire

building well? The computer scientist, Mark Greaves, who contributed the
Foreword to this book, describes the advances in natural language generation
with tools like GPT-3 as having 'fluency and expressivity':

>>
Using modern techniques, machines are starting to successfully
ML

perform creative, original tasks in domains like language that were

once uniquely the realm of humans. There have certainly been limited

achievements based on more traditional Al which have been called

‘creative', such as the famous 'hand of God' move played by Deep Blue
in its chess match against Garry Kasparov. But these are quite rare,..
These systems seem to exhibit a level of creativity and expressiveness
and linguistic artistry that machines hadn't reached in the past. And,
in the realm of game playing, ML-based AlphaGo has shown real
creativity as well. 6 «

Creativity and coherence do not equate with competence, however, and

therein lies at least part of the answer to the question that this book will
explore. To wit, how should the profession of architecture consider and
respond to futures made possible by advances in artificial intelligence, the
so-called 'Second Machine Age?7 Having moved deliberately, if reluctantly,
through the eras of both CAD and BIM, can we propose a willful, designed
route –

professional strategy that acknowledges the inevitability of

a –

a preponderance of intelligent machines in every dimension of design,

construction and built asset operation while maintaining a proper role for
human architects?

This is not a new question, at least as far as the relationship of machine

intelligence and its
implication for Design with a capital D is concerned –

always the first place that architectural theory visits when struggling with a
8
big problem. Unexamined, however, are the implications for the practice,
rather than the result, of architecture. If designers solve, as described by Peter
Rowe (quoting Horst Rittel) 'wicked problems', 9 with open-ended beginnings
and fixed conclusion, competent practice requires heuristics across a
no

broad spectrum of technical and aesthetic issues. This would seem to be

a strategy that is exclusively human. However, computers are increasingly

able, empowered by machine learning, to learn these techniques, and when

they do professional strategies and methods
so, –

and the value of designers

themselves will be inalterably transfigured.
-
For the practice of architecture, the implications of this change are nascent.
Machine learning algorithms are evaluating mortgage applications, reading
routine X-rays, inventing never-before-seen strategies for playing board

games, and even getting dangerously close to composing coherent ideas.

However, Al-based approaches to the design generation are only now
becoming apparent and none are commercially viable. As Daniel Susskind
predicts in his second book, computers are increasingly becoming capable
of tasks, as opposed to entire jobs. 10 Across the spectrum of services that
architects provide, there are ample opportunities for the automation of tasks.
Does that mean architectural work will be replaced,
by contrast augmented,
oder

by capable computers? Rather than wrestle with the larger question of

whether we are to be wholly replaced by machines, perhaps a more intelligent
route can be found where computers assist in the critical, but more mundane,
aspects of practice: those that drive project delivery, technical precision and
performative predictability.

At thisjuncture, the architecture profession itself is not under existential

threat, but neither has it developed a strategy for the inevitable advent of
learning machines. That strategy should have an expanded remit beyond our
usual worries about agency as designers and look more broadly at how
our

powerful computation may affect our roles in the formulation, conception,

delivery and use of buildings, and thereby the essential value of the profession
of architecture itself. To that end, this text will examine three aspects of
practice in the upcoming age of computer intelligence that in combination give
us a view towards our fate:

1. How the processes and methods of practice may change.

2. What those changes will mean for our relationships to the systems of
delivery in the built environment.
3. What opportunities there may be to refactor and improve the results of
our efforts.

In concert it will attempt to propose a strategy for the profession going

forward into the age of machine intelligence.
 FRONT MATERIALS

PROCESS

ARCHITECTURE
IN THE AGE
OF MACHINE RELATIONSHIPS
INTELLIGENCE

Understanding the implications

of a disruptive tool like artificial
intelligence means exploring
the relationship between the
enabling technology, the agency
RESULTS
of architects in deploying that
tool towards new ends, how
that agency might change the
architect's role in the systems
of project delivery, and CONCLUSION
finally how the value of the
architect's services might change
accordingly. Each of the following
chapters is a different take on the END MATERIALS
combination of these questions.
ACKNOWLEDGEMENTS III

FOREWORD V

INTRODUCTION VI

1.1 TOOLS AND TECHNOLOGIES 2

1.2 WHAT IS ARTIFICIAL INTELLIGENCE [AI]? 13

1.3 PROFESSIONAL INFORMATION AND KNOWLEDGE 21

1.4 AI AND PROCESS TRANSFORMATION IN DESIGN AND BEYOND 32

1.5 SCOPES OF SERVICE 45

1.6 DELIVERY. MEANS AND METHODS 58

2 ECONOMICS, COMPENSATION AMD VALUE 72

2.2 LAWS POLICY AND RISK 82

2.3 THE DEMAND FOR PROFESSIONALS 92

2.4 EDUCATION, CERTIFICATION AND TRAINING 102

3.1 THE OBJECTIVES OF DESIGN 116

3.2 CREATING,CONSUMING AND CURATING DATA 126

3.3 TASKS, AUTOMATED 136

3.4 LABOUR OF DESIGN 148

3.5 VALUE PROPOSITIONS AND BUSINESS MODELS 155

4.1 CONCLUSION 168

BIBLIOGRAPHY 173

REFERENCES 176

INDEX /

CREDITS 184
PROCESS
01

» HOW WILL THE PROCESSES AND

METHODOLOGIES OF ARCHITECTURAL WORK
CHANGE UNDER THE PRESSURES (AND
POSSIBILITIES) OF INTENSIVE AUTOMATION?
WHERE IS OUR WORK BEST AUGMENTED OR
INEVITABLY SUPPLANTED? «

DOI: 10.4324/9781003297192-1
 TECHNOLOGIES
AND
TOOLS
1
1.

» COMPUTERS HAVE BEEN DEPLOYED OVER

THE LAST TWO DECADES LARGELY IN THE
SERVICE OF MORE EFFICIENT DEPICTION AND
REPRESENTATION. ARTIFICIAL INTELLIGENCE
SHIFTS THE FOCUS FROM REPRESENTING A
DESIGN TO REASONING ABOUT IT DIGITALLY
-
IN SOME CASES WITHOUT THE NEED FOR
A HUMAN OPERATOR OR EVEN MUCH HUMAN
INTERVENTION. WHAT DOES THIS MEAN
FOR THE EVOLUTION OF THE INSTRUMENTS
AVAILABLE TO MAKE ARCHITECTURE? «

DOI: 10.4324/9781003297192-2
TAXONOMY OF TECHNOLOGIES

Various scholars have written the history of technology in the architecture

profession and it is beyond the remit of this project to reprise that history in detail
here.1 It is possible, however, to see that history organised along two axes: the
progression of technologies that comprise the toolkits of architecture, and the
taxonomy of uses these tools support. Artificial intelligence defined broadly as
-

the ability to perform complex cognitive tasks in ways that produce results akin to
the human mind is an enabler of the various tools described in this taxonomy.
-

On the first axis, a vast simplification of the history of technologies in

our discipline could start as far back as 2150 BC, with a statue of the
Mesopotamian ruler, Gudea, that depicts a drawing of an architectural
plan
sitting on his lap, marking the putatively first evidence of analogue drawing
2

of design. For at least the next 42 centuries architects abstracted their ideas
a

by depicting design, explicitly or implicitly at scale, on pieces of paper. Mario

Carpo maps this process in digital terms by suggesting that drawings and
allographs notations, as vectors of design information, are actually a way of
processing data with the very limited 'central processing unit (CPU) cycles'
made available by manual graphics. With minimal expenditure of energy
3 -

the act of drawing two parallel lines on a piece of paper the architect could
-

imply the construction of a three-dimensional, tectonic, materialised wall in

the actualised building, and memorialise that rich idea for reference by the
entire enterprise responsible for making it.

1.1.1: GUDEA,
WITH A PLAN
DRAWING IN
HIS LAP, C.
2150 BC 4
Machine Learning

COMPUTER-AIDED DESIGN

In the penultimate decade of the 20th century, architects moved in earnest

into their firstforay of digital, computer-aided design (CAD, sometimes
referred to as computer-aided drafting), whereby the mechanical process of
inscribing lines, arcs and circles on paper was replaced by creating those same
lines on to a virtual plane by inputting digital lines. While the vast investment
in computers and plotters made this transition feel monumental, it was in fact
more of a translation of existing techniques of drafting into virtual form, the

object of which was still the production of drawings but more precise. The
-

evolution of technologies of large-scale printing devices (like plotters) moved

in parallel with the electronic drafting tools (like AutoCAD©).

The move to computerised drawing, while ushering in an era of curvier

buildings that were
suddenly easier to draw and to a lesser extent also
-

easier to build by virtue of the more precise geometry afforded by CAD,

-

did little to address the informational gap between the definition of a

design, its intent by the architect and the builder's ultimate responsibility
for its construction, a divide defined in the Renaissance by Alberti in De
reaedificatoria. According Carpo, Alberti set out the proposition that
5 to
architects draw and builders follow those drawings without deviation. CAD
gave architects an opportunity to draw faster, with more graphic consistency,
and even reuse certain representations (like AutoCAD© blocks) across multiple
sheets of a drawing set or even multiple projects. Yet, despite the added
informational power of this data, the quality of work itself did not improve:

» In short, the ontological gap between design intentions, their

notation through construction drawings, and their material
implementation leaves an inevitable grey area of undecidability,
argument, frustration, litigation, and liability where all kinds of ad-
hoc personal interventions, approximations, improvisation, bullying,

persuasion, implorations, machinations, and subterfuge take the place

of construction drawings and specifications, and haggling becomes the
design instrument of choice. «
6

Into this gap, 20 or so years later, came the next leap

representational in

technologies: building information modelling (BIM). In theory, BIM

represented a flip of the traditional allographs strategy for architects: a 3D
representation of the building was constructed in virtual digital space, from
which those venerable drawings would be extracted as 'reports'. Gudea's floor
 1.1 Tools and Technologies

plan, rather than representing the diagram of a design that otherwise lived 1.1.2 AN
EARLY CAD
in his head, would now be an extraction from full-scale digital replica living
DRAWING
in computer memory just another view of the relevant data. Every member
-

BY PELLI
of the design-to-build team could, in theory, add information to that model CLARKE PELLI

to complete their respective work. In practice, however, incompatibilities in ARCHITECTS

process and outcome, the adversarial nature of building and the centuries-old
allure of drawings have made BIM a tool used largely for production of even
better working drawings than CAD. Its epistemological value as an organising
principle for the informatics of building is almost completely ignored.

The last 10 years have seen explosive digitisation of many aspects of modern
life, and design and construction have been no exception. Powered by the
ubiquitous availability of the massive storage and CPU power of the cloud,
and the ability to deliver those capabilities virtually anywhere through the
internet, the architecture/engineering/construction/operation (AECO) industry
is adopting a variety of computational tools, if peripatetically.

Designers have an array of

»
modelling
»
rendering
» data management
»
analytical platforms.
Builders, armed with digital tablets on the project site:
»
rely on digital versions of what used to be the exclusive domain of paper

» usedata tools to manage the array of administrative transactions that

comprise the construction process, deploy drones and LiDAR (Light
Detection and Ranging) to document construction progress.

Building owners:
» demand digital documentation of completed projects in lieu of rolls of

post-construction drawings
» assemble data from sensor networks in their building control systems
to optimise building performance.

THE DIGITAL INTERSTICE

All told, the AECO industry seems to have entered a 'computational

interregnum' of sorts, where various processes across the delivery continuum
are becoming digitised. As a result, the variety of programs, platforms, data
types and supporting hardware becoming as varied and disorganised
is - -

as the disaggregated building industry itself. In 1970, Nicholas Negroponte

anticipated this state when he suggested that any transition to computation

first imitates directly the analogue process it proposes to replace. As a result, 7

we are likely to see more years of excitement and confusion as the building

industry wills itself into its digital future at the same time that it gives up
8
the simple 'interoperability' afforded by paper-based, analogue processes.
Desires for some sort of broad theory of global data exchange will remain

strong, but unrequited, outstripped by a software market eager to address

new opportunities and the inherent complexities of various processes that

necessarily comprise the delivery chain.

The building industry is typically not well-enough organised, nor can it compile
enough market clout, to adopt fresh technologies or innovations soon after
their introduction. It often has to wait until hardware, software or business
models are sufficiently mature for architects, engineers and builders to adopt,
adapt and improve such systems for their use. Such was the case with CAD
platforms, which were originated by the aerospace industry and had to be
ported down to personal computers sufficiently inexpensive to be in reach
of AECO customers. Similarly, modelling platforms such as BIM or high-
resolution rendering eventually appropriated the tools of manufacturing and
movie-making once those technologies were within economic reach.
9
 1.1.3:
A LIDAR SCAN
OF CONSTRUCTION
IN PROGRESS,
MAPPED AGAINST
A BIM DATA SET.
THE CONDUITS
(IN GREEN)
AND DUCTING
(IN BLUE)
ARE VIRTUAL
ADDITIONS TO
THE DIGITAL
SCAN FROM THE
MODEL.

MOVING TO MACHINE INTELLIGENCE

In an exception to this otherwise reliable rule of thumb, architects did lead

the charge with the earliest versions of machine-driven design: scripting. As
the first proponents of technologies, such as McNeel Grasshopper (and, later,
Autodesk Dynamo), architects defined and elaborated the idea of what is now
known parametric, or generative, design, where computerised scripts drive
as

geometry engines (like Grasshopper controlling Rhino). However, after 20

years of such work, scripting capabilities are largely deployed in the service of
relatively minor problems, such as shape generation or fenestration geometry,
rather than systematic alternative generation or analytical evaluation. While
10
later-generation tools like Hypar have begun to accelerate the idea of
generative design, the dearth of analytic tools to evaluate alternative designs
generated by scripting has generally limited their use to form generation.

Machine learning and artificial intelligence, the most recent tools on our
timeline, likely to follow the typical path in order to reach architects.
are

While major corporations are already absorbing AI/ML capabilities into their
core operating strategies, most artificial intelligence available to architects
11

is delivered through their smart phones, while we order dinner online or

request a ride to the office. A few promising start-ups and other experiments
are testing the technology on various tasks on the construction site (e.g.

worker safety checking through computer vision) or project administration

(e.g. managing vast swathes of change orders and requests for information),
suggesting that we are still in early days.
8 MACHINE LEARNING

1.1.4:
AUTODESK
SPACEMAKER
AI'S
DESIGN AND
EVALUATION
INTERFACE

_.._____
,,..=--­
..,.,.,....,._,.,,,,,..,,..,_.

-
..,.

However, there are indications that this wait may not last much longer.
As I write this chapter in November 2020, Autodesk, my former employer,
announced the acquisition of Spacemaker, an Al-driven design generation tool
that evaluates site and building constraints and generates preliminary design
solutions (see Figure 1.1.4). That tool comprises a combination of design
representation, evaluative analysis and an Al infrastructure that learns best
results by interacting with its human decision-maker/operator.

As such, it may be an indicator of where the next generation of modelling

representation, beyond BIM, may be heading. The acquisition cost was $240
million, or 25% more (in 2002 USD) than the acquisition of Revit Technologies,
which started the BIM revolution in AECO in earnest.12

A TAXONOMY OF USE
In earlier work I have proposed a taxonomy by which the vast array of
digital tools emerging might be categorised, irrespective of their underlying
technologies.13 In that analysis I suggest that the tasks of the building
enterprise, as supported by computation, fall broadly into four categories:

1. Representation (the depiction of authorial ideas).

2. Analysis and simulations (evaluation of those idea to understand their
performance and implications).
3. Realisation (the translation of those ideas into built form).
4. Collaboration (the distribution and management of information across
the enterprise).
As we consider the implications of machine learning and artificial intelligence
in this context, those definitions require some additional refinement.

Representation
Drawings, text, images and physical models were the representational tools
of thepre-digital age, followed by CAD and, eventually, BIM and parametric
design. In a world of increasingly digitised data that
might be consumed by
smart machines, mathematical models and other data sets depicting a built
asset like sensor data coming from a building control system should also
- -

be considered as representational. Further, as early forays into parametric/

generative design suggest, algorithms that generate design are increasingly an
important part of the representational process.

Analysis and Simulation

Computation is well-suited to
examining data that results from representation
in order to understand and evaluate it. Extended by the putative power of
artificial intelligence, this capability might more accurately be described, as
suggested by Agrawal, Gans and Goldfarb, as prediction: using (representational
and other) data to evaluate the implications of design decisions and predict
outcomes and implications of their underlying logic and decisions. 14

Realisation

Design data is the

logical underpinning of digitised construction processes.
This isparticularly true as construction, assisted by AI/ML, is automated and
the processes of building are absorbed by machines that learn how to build.
Those robots will need design data to guide them on the job site. Similarly,
gathering information that documents the construction process LiDAR scans -

of construction process that are mapped into BIM by intelligent systems are
one example is another aspect of digital project realisation.
-

Collaboration

From the of internet-based data management to today's common

early days
tools such as BIM 360, Procore or even Google Docs, project teams have
needed to organise, transmit and manage digital assets irrespective of format.
As we transit the digital interregnum, however, there is an increasing need
toindex, locate and understand all this data, much in the same way that
modern search engines find information without user concern about format
or location. Overtime, digital design and construction will demand federated
1.1.5:
RELATIONSHIP
OF project informatics across the enterprise, where data structures in differing
TECHNOLOGIES formats are connected to form a coherent whole with distinct parts. It is highly
AND TOOLS
OVER TIME unlikely that standard data formats or interoperability protocols will allow
all this data to become useful acrossthe varied processes of construction.
Machine learning algorithms may well be our only strategy to make sense of it.

THE EVOLUTION OF TECHNOLOGIES AND TOOLS

Each of the technologies described above analogue drawing, CAD, BIM, and
-

ultimately AI/ML -

will change the way work is done and the tools deployed to
accomplish it. The intersection of technology types and the evolution of the

resulting processes and procedures they enable is described in Figure 1.1.5 .

Computational power, reliance on automation, depth and breadth of data

resolution and precision increase dramatically as our profession has moved
from the upper left to the lower right on this grid. Certain aspects of design
autonomy will be refactored as machines take over some but not all of
- -

the tasks of the architect. Figure 1.1.6 describes some of these possibilities,
mapping the evolution of our four categories of technologies against both the
data richness of the design enterprise (the top curve) and the automation of
design process (the bottom curve).
 1.1.6:
HUMAN VS
MACHINE
AUTOMATION
It is suggested here that as tools become enabled, there are parallel
more

increases in the granularity and complexity of resulting information, along

with the potential, in the digital era, for computational automation. During
the drawing era, for example, informational complexity increased only with
a similar
change in the technical demands of construction itself, with the
introduction of modern construction systems and complex delivery models.
Automation was not available at all. The advent of CAD allowed for additional

geometric complexity, more extensive documentation (in theory) and some

automation through standard component libraries and scripting tools inside

the CAD programs themselves, such as AutoLISP in early AutoCAD©. Those
early scripting tools allowed the functions of AutoCAD© to be recorded and
repeated; their successors in Rhino controlled dimensional parameters of the
geometry model and gave the designer an ability to generate new forms.

A significant jump on both curves is apparent with the advent of BIM,

however, representation switched from the abstraction of drawing to
as

virtual 3D models inparallel with parametric BIM families and data sets, and
generative design approaches through mature scripting that will memorialise
processes and procedures once dependent on human intervention. While
scripts within CAD simply manipulated geometry, tools like Dynamo allow a
designer to parametrically manipulate both the components of the design
(like the
size of windows) and their relationship to the overall building (like
the location of the windows within an exterior wall). In the interregnum, we

will see extensive generation of digital data and the episodic automation of
various processes that would be otherwise disconnected.

The transition to broad-scale AI/ML will greatly enhance both the amount and
the value of its precedent digital sources, serving as data lakes for intelligent
machines to learn from. At the same time, computers will start to train
themselves to perform machine-automated tasks. Those same machines will
teach themselves about the relationships of the heterogenous data sets that
comprise projects and create an 'interoperable' constellation of AECO data.

These trends potentially daunting but entirely necessary

portend a - -

trajectory for practice as the complexity of buildings and the power of

computers advance in parallel. In combination, sophisticated digital modelling
enhanced by machine intelligence will likely draw together the disparate
forces of design, construction and building operation in a single, consolidated
effort, coalescing around the data about the enterprise of building, from
inception to demolition. As Carpo speculates:

the
unprecedented power of digital simulations, one may surmise that
� B ; Given

at some point virtual models may become perfect duplicates of and substitutes
for, the buildings they represent embodying and enacting all and every aspect of
-

them. Their designers could then make a digital model just as builders would once
have made an actual building, and the final translation from model to building
would entail no intellectual (or informational) added value whatsoever. « 15

The information about and knowledge necessary for building would, as a

result, be dramatically transformed. But how?

ARTIFICIAL
IS
WHAT
2
1.
?
(AI)
INTELLIGENCE

»
THE IDEA THAT COMPUTERS MIGHT
DRAMATICALLY AUGMENT THE CAPABILITIES
OF HUMANS -
OR POSSIBLY SUPPLANT
US ALTOGETHER -
IS MANY YEARS OLD.
BEFORE SETTING OUT THE OPPORTUNITIES
AND THREATS OF AI FOR ARCHITECTURAL

PRACTICE, THIS CHAPTER WILL SKETCH A

BRIEF HISTORY OF THE TECHNOLOGY AND
POSIT A TAXONOMY BY WHICH ITS PRESENCE
AND FUTURE MIGHT BE UNDERSTOOD AND
PREDICTED. &#x00AB

DOI: 10.4324/9781003297192-3
 Machine Learning

ORIGINS OF GOOD OLD-FASHIONED AI

The computer scientist John McCarthy is generally credited with coining the
term 'artificial intelligence' in 1956, suggesting that computing machines
could somehow mimic the functions of the human brain. And long before
computing was within reach of mainstream practising architects, Nicholas
Negroponte and others were exploring the idea of digital design. As early
as 1964, Walter Gropius acknowledged that there might be a role for these

new machines in the profession, suggesting that 'it will certainly be up to us

architects to make use of them intelligently as means of superior mechanical

control which might provide us with ever-greater freedom for the creative
process of
1
design'.

In 1958, the psychologist Frank Rosenblatt put forth a theory of'perceptrons'

that was the precursor of today's modern neural networks. Rosenblatt posited
that it wastheoretically possible to represent visual information by'teaching'
a crude
digital facsimile of a human neuron, and thereby encode human
knowledge in accessible form. A decade later, the theory was challenged by
MIT's Marvin Minsky, who suggested that'deeper' models (with more layers of
such neurons) would never yield reliable results, presaging an argument in Al
strategy that survives, in part, to this day.

1.2.1:
NICHOLAS
NEGROPONTE’S
URBAN 5
SEEK, BY THE
ARCHITECTURE
MACHINE
GROUP, AS
EXHIBITED AT
THE JEWISH
MUSEUM, NEW
YORK, 197θ
 1.2 What is Artificial Intel igence (AI)?

At the same time Nicholas Negroponte was exploring more practical questions
in his MIT
lab, 'The Architecture Machine'. An architect by training, Negroponte
experimented widely in the early uses of technology and design, anticipating
our use large screens, video and cameras, machine intelligence and
of tools like
immersive environments. His efforts anticipated early strategies for artificial
intelligence, positing the possibilities of'an intelligent environment that we
would all eventually inhabit and that would eventually surround all of us'. 2

While Negroponte examined questions of the representation, generation

and manipulation of 3D space, other work considered core to understanding
human thought probed what was called natural language processing:
computers understanding and creating the written word. Computer scientists
3

attempted to translate emergent theories of human cognition into software,

creating computational and semantic structures to extract understanding
from text and reason inferentially from that resulting knowledge. This
established what has since been called 'GOFAI' or 'Good Old Fashioned AI', 4 as

attempts were made to build models of human cognition.

By the mid-1980s, AI research and software companies were attempting to

extrapolate these theories into commercial 'expert systems', in which the
knowledge of particular domains were painstakingly encoded by human
programmers into algorithms in order to transfer detailed domain knowledge
into acomputer that could, theoretically, replace its human counterpart.
These efforts rapidly hit the limits of both the efficacy of theory and the
processing and storage capacity of hardware. The 'AI Winter' ensued, in which
attention(and funding) waned for many years and the promises of AI would
seem unfulfilled. 5

BEYOND PERCEPTRONS

By the 1990s, computers were getting faster, cheaper and more available, and
a different strategy for AI emerged: neural networks and machine learning.
Benefiting from the vast computing power and equally gigantic storage
-

capacities of the cloud, AI systems began to be based on a digital emulation

-

of human memory, encoding information and relationships in increasingly

complex layers that could be indexed and accessed like hyper-intelligent
databases. 6 The power of computation revived the theory of the perceptron.
The definition of an expert system could shift from 'human-encoded
understanding' to 'computer-generated expertise' through programs that
'learned' from examining enormous data sets. Machine learning programs
could, by virtue of their ability to process vast amounts of example data,
1.2.2:
AI-BASED
X-RAY
EVALUATION 7

create statistical correlations that

approximated learning. Rather than
8

somehow simulating the mechanics of human understanding, the goal of

machine learning Al became software that could deeply learn rather than
-

be
told about -

the world.9

Nowadays, 'deep learning' comprises much of the academic and commercial

work in Al and is often conflated with the broader definition of the term.
Cloud computing infrastructure, fed by vast data sets coming from an array
of internet-based sources, along with significant progress in the underlying
learning algorithms, has brought AI into daily use, from smartphones to
Google Translate. Computers are now world champion chess players,
competent radiologists and credit scorers. Later generation ML programs
can learn the rules of a game like backgammon by simply playing millions of
simulated contests and testing what works best

As capable as these systems appear, they do not actually understand

anything, but have rather built semi-reliable statistical correlations of
information relationships. In highly constrained contexts (where, for example,
there is a well-defined training set, such as a game or a set of specific images),
ML systems are surprisingly effective, and particularly today in the realm of
natural language, as Google Translate demonstrates. As a result, much of the
work in AI/ML today pursues this empiricist strategy. What the AI/ML systems
entirely lack, however, is a rudimentary understanding of how the world works
or anything remotely resembling common sense.
There is scepticism that without such an understanding of context, Al systems
cannot be truly useful or reliable, and therefore if they will challenge the
existence of, say, architects. The psychologist and entrepreneur Gary Marcus
concludes that these 'narrow intelligence' strategies are flawed:

» (l)t is a fallacy to suppose that what worked reasonably well for

domains such as speech recognition and object labeling which -

largely revolve around classification will necessarily work reliably

-

for language understanding and higher-level reasoning. A number of

language benchmarks have been beaten, to be sure, but something
profound is still missing. Current deep learning systems can learn
endless correlations between arbitrary bits of information, but still go
no further; they fail to represent the richness of the world, and lack
10
even any understanding that an external world exists at all. «

Marcus argues that the work should consider a return to the original
motivations of the Al field simulating human cognition and combine the
- -

data collection and analysis capabilities of deep learning systems with new
models of perception and inference. While some progress has been made on
this front, such systems do not exist today and will depend on the digitisation
11
of new theories of understanding and, particularly, causality.

TOOLS OF INTELLIGENCE FOR ARCHITECTS

As AI becomes more capable, today's architect is presented with a range of

potential sources of intelligence to deploy in the service of the craft:

» her talents, skills and experience (as certified by, for example,
12
own

their professional registration)

» an
array of hard-coded computer programs that achieve specific ends
(such as energy analysis or structural engineering)

» machine learning systems (which might learn from data coming from
theirdesign projects, or even sensors within finished buildings and
provide insight), and
»
ultimately the speculative prospect of cognitive systems that can
reason within context (only seen in science fiction today).

The latter three, based on technology, are summarised in Figure 1.2.3 .

1.2.3:
COMPARISON
OF AI
TECHNOLOGICAL
TYPES

The computers available to architects today are more adept at direct problem-
solving than what Stanford Anderson once called 'problem-worrying', resolving
14
the goals of the problem while simultaneously creating the design, evocative
of both Negroponte's and Peter Rowe's interest in heuristics as a strategy
for solving 'wicked problems'. Hard-coded software single-mindedly solves
15

specific problems; your cost-estimating system will tell you nothing about
the fire exiting required of your design, nor is it capable of learning how to
do so. Emerging AI/ML systems, now being applied to problems of the built
environment, may be able to evaluate or even predict issues in a specific
context, but certainly are nowhere near ready to design entire buildings,
heuristically or otherwise. And, as of this writing, devotees of cognitive
16

systems have spent decades building 'real world knowledge' as the basis of a

next generation inference system, with limited success. 17

While questions of human consciousness are not considered here, it is fair to

describe intelligence artificial or otherwise as the ability to amass, organise
- -

and reason inferentially about heterogeneous collections of knowledge in

context, Marcus believes this robustness cannot be currently achieved by
today's Al systems, which are unable to 'reason flexibly and dynamically about
the world, transferring what is learned in one context to another, in the way
that we would expect of an ordinary adult'. 18 The work of architects surely
qualifies as in Marcus's terms, as he suggests that
'higher-level reasoning'
'where there is nocoherent, causal understanding of basic concepts, there
may be way to engineer robustness in complex real-world environments'.
19
no

Early attempts, however, are beginning to emerge, with Autodesk's

Spacemaker Al acquisition as a primary example. That tool uses a combination
of modelling, analytical algorithms and Al to generate and evaluate planning
alternatives, and then 'learns' about best practice by compiling results
garnered from both the result of analysis and the choices of the human
designers selecting the best options.

MACHINE CAPABILITIES

Al expert Mark Greaves describes the capabilities of current Al systems within

a continuum from evaluation (understanding the implications of information),
to simulation (using information to approximate similar circumstances in a
different context), through to generation (creation of new ideas). 20 I will add to
his categories, which are generally based on Bloom's Taxonomy of learning 21
(to which we will return later), a fourth, earlier capability of lesser profundity,
'understanding,' where the system can locate, access and deploy information.

Current capable empiricist AI/ML systems are excellent at evaluation (when

systems read routine X-rays, for example) and even, in some circumstances,
simulation (demonstrated best by extremely competent chess programs).
However, examples of real generation, beyond the occasional flash of brilliance
in the use of a hitherto never before seen Go move, 22 are limited, and there are
no circumstances where computers are capable of generating a set of original
ideas that comprise the design of something as complex as a building.

1.2.4:
MODIFIED
GREAVES’S
MODEL OF AI
CAPABILITIES 23
This is, however, only the current state of affairs and is a function of empiricist
Alsystems that can only 'deduce' based on massive correlations of data. It
is likely that, over time, empiricism will give way to emulation of cognition
as philosophers, neuroscientists, computer scientists and their commercial
counterparts build ever more capable machines that move toward general Al,
what Pedro Domingos has called the Master Algorithm, in the service of what
24

is now becoming known as 'artificial general intelligence', or machines that

can both learn and reason about the world in context. Today, we are far away

from such functionality, but it remains the grandest goal of Al development.

When Marcus's 'robustness' meets Greaves's generative capabilities, architects

(and most of the labour force, as Daniel Susskind has suggested) have much
to worry about, since '(q)uite simply, if we cannot count on our AI to behave
reliably, should not trust it'. 6 So, until then, the work of human architects is
we

to orchestrate the combined tools of their talents, an array of software tools

including BIM and emerging deep learning tools into coherent and valuable
practice in anticipation of the day when cognitive platforms are readily
available. By then, one hopes, the profession will have a firm grip on both the
technologies available and the means to direct them.
KNOWLEDGE
AND
INFORMATION
PROFESSIONAL
3
1.

» PROFESSIONAL KNOWLEDGE IN

ARCHITECTURE IS AN INTRACTABLE

CONSTRUCT, GIVEN THAT MUCH OF THE

COMPETENCE AND DECISION-MAKING BY
PROFESSIONAL ARCHITECTS IS BASED ON

TRAINING, INSIGHT AND, MOST IMPORTANTLY,

JUDGEMENT. TECHNICAL KNOWLEDGE IS
NEITHER WELL-ORGANISED NOR EASILY
REACHED. ASSUMING AI MAKES BOTH THE
CAPABILITIES AND KNOWLEDGE BASE OF
PROFESSIONALS MORE WIDELY ACCESSIBLE,
PROFESSIONAL KNOWLEDGE WILL NO LONGER
RESIDE EXCLUSIVELY IN THE MINDS
OF HIGHLY TRAINED ARCHITECTS. DOES
PROFESSIONAL DESIGN TRANSFORM? «

DOI: 10.4324/9781003297192-4
Machine Learning

PROFESSIONALS AND PROFESSIONALISM

In The Future of the Professions, Richard and Daniel Susskind's treatise on the
potential demise of the
professions in a world of increasingly capable artificial
intelligence, you can find the following complete, if slightly tongue-in-cheek,
definition of why society has created and empowered a class of professionals,
architectural or otherwise (emphasis added):

acknowledgement of and in return for their expertise,

» In
experience and judgement, which they are expected to apply in
delivering affordable, accessible and up-to-date, reassuring and
reliable services, and on the understanding that they will curate and
update their knowledge and methods, train their members, set and
enforce standards for the quality of their work, and they will only
admit appropriately qualified individuals into their ranks, and that
they will
always act honestly, in good faith, putting the interests
of clients a head of their own, we (society) place our trust in the
professions in granting them exclusivity over a wide range of
socially significant services and activities, by paying them a fair
wage, by conferring upon them independence, autonomy, rights of
self-determination, and by according them respect and status.
1 &#x00AB;

There is a social bargain defined here: running our world would seem to

require both extraordinary expertise leavened by public trust, so we ennoble a

small group of highly educated, certified and supposedly well-paid individuals
to do very special things. Architecture, it seems, as the profession that
translates desire and capital into occupiable space, falls under this rubric. And
while some have that this power and influence has failed the public
suggested
it is intended to serve, architects remain a crucial, if challenged, component
2

of the equally challenged building industry ecosystem. That system, explored

more thoroughly in Chapter 1.6 is largely organised in the service of making
,

products to wit, buildings and other physical infrastructure yet architects

- -

as professional players within it have a distinctly different role: individuals with

expertise who provide judgement and take personal responsibility for results
of that judgement. In the systems of building delivery, this means that, unlike
almost everyone else (builders, subcontractors, product suppliers, fabricators),
professional architects provide services, not things. And as a result, they take
personal, rather than corporate, responsibility for their actions.
 1.3 Professional Information and Knowledge

The philosopher Donald Schön, who studied how professionals learn and
deploy expertise, has suggested that architects and other professionals work
in a way that is distinct from less institutionalised careers by virtue of what
he calls `reflective practice', or the ability to apply insight and make decisions
through the implicit understanding gained with `extraordinary knowledge in
matters of human importance'. What Schön called tacit `knowing in place'
3

we might call intuition, or the heuristic approach that Rittel describes as

4
necessary to solve the `wicked problems' of design. We will examine later the
question of whether an Al could even achieve such `extraordinary knowledge',
but for purposes of this specific exploration of professionalism, let us stipulate
that the architect's synthetic role, by virtue of her professional responsibilities,
is not well-replaced entirely by either empiricist Al of today or perhaps even
tomorrow's cognitive systems that could somehow `learn' all the procedures
and processes of practice. If Schön is correct, there is something about
professional knowledge that will lie beyond the reach of those systems.

However -

and with Al, it seems there is always a `however' -

the formulation,
design, procurement, construction and
operation of a building is rife with
procedural and data-driven tasks, ranging from calculating quantities
to modulating temperature and humidity. As such, the systems within
which those processes operate are sure to be influenced, if not partially
transformed, by computing. The question for architects is where,
autonomous
how and what will tomorrow's Al-assisted architects really need to know?

WHAT IS PROFESSIONAL KNOWLEDGE?

In autumn 2020 I participated in an online panel of practitioners, educators

and students to explore career prospects for graduates, during what we
hoped was the latter stages of the global pandemic. Late in the discussion
there was an exchange between the principal of a local firm and a well-
respected, left-leaning dean of a New York school, comparing their respective
expectations of professional know-how of recent graduates. Unsurprisingly,
the practising architect wondered why his recent hires knew so little
about `how a building goes together' (a familiar refrain) or the processes
or procedures of practice. The dean wondered why she should prepare

graduates for a profession that currently seems so unsuited to the challenges

-

social, economic, environmental of the times, and suggested that her

-

job was to graduate students not in order to support the practice of today,
but rather to radically reform it in and for the future. Unwilling to miss an
opportunity to triangulate one of my favourite hobby horses, I asked the
practitioner how the profession simultaneously demand that graduates
can

know so much more -

and be delivered to them

by the schools as `mini-architects'
while paying them so little. The response was telling: in this local
-

firm, reaching licensure meant a bump in pay of $1,000, or slightly less than 2%
of what I expect he pays his least experienced staff. 5

So, what do you need to know to be an architect? The vantage points of

the panellists can be a good starting point: what do educators think a
young architect needs to know to begin her career, versus the authors of
professional licensing examinations and requirements? Of course, the timing
does not correspond exactly, since licensure in the UK and US requires

professional experience after education, but nevertheless a comparison of

the two constructs is instructive. For purposes of this examination, I compare
a curriculum with which I am deeply familiar (that of the Yale School of

Architecture) and the post-graduate licensure competencies as outlined by the

ARB Part 3 examination in the UK and the comparable standard in the United
States: the Architectural Registration Exam (ARE) administered by the National
Council of Architectural Registrations Boards (NCARB). 6

At Yale, we require our `Part 2' candidates (those getting a post-baccalaureate

Master of Architecture degree) to be capable in the following curricular
7
categories:
» Architectural Design traditional studio pedagogy in the creation of
-

a building design, focused primarily on the conceptual and schematic

phases of the work.

» Visualisation depiction, through analogue
-
and digital means, of
ideas of building design.
» Environmental Systems the design and integration of structural,
-

mechanical, electrical and other technical subsystems of building.

»
History/Theory understanding the historical
-

and theoretical
platforms of the discipline of architecture.
»
Building Technology seeing and performing the technical
-

manifestation of building, including integration of complex systems,

creation of construction documentation and field construction

experience.
» Urban Design and Landscape -

understanding how buildings operate in

the context of cities and sites.
» Professional Practice -

introduction to the profession, including ethical,

legal, project management, business and project delivery aspects.
The comparable list of competency categories from the NCARB ARE is both
more extensive and somewhat incompatible, with six major categories and
8
almost 100 sub-competencies:
» Practice Management -

how do you run an architectural practice?

»
Project Management how do you manage, coordinate and lead
-
a

project through the stages of delivery?

»
Programming and Analysis -
how do you set up a project to be ready
to design it?

»
Project Planning and Design -
what do you need to know to design
a complete building that meets all applicable regulations and
requirements?
»
Project Development and Documentation once the design is set, -

how do you properly detail, document and transmit that project to

the builder?
» Construction and Evaluation -

once the design is complete, how do you

support and administrate construction, and evaluate the building after
its completion?
The ARB, as a somewhat more parsimonious examiner, has fewer categories
and only half as many sub-competencies (50): 9
» Professionalism -

how do you function ethically as a professional, and in

a practice?
» Clients, Users and Delivery of Services -

what is your role in

organising, leading and managing a project?
»
Legal Framework and Processes -
what are the legal and regulatory
frames of practice and building, and what is your responsibility for
them?
» Practice and Management -

how do you run an architectural

business?
»
Building Procurement what are your roles and responsibilities
-

during contactor procurements and subsequently, construction?

An informal mapping of sub-competencies against curricular categories yields

an interesting disconnect, particularly as regards UK standards, suggested by
the comparison in Figure 1.3.1 .

It would seem that issues of practical implementation, and particularly as

regards the business and processes of professional practice, are emphasised

strongly by the ARB, and assuming that architecture curricula are generally
1.3.1: aligned around the world(having all been derived from a similar source in the
COMPARING Beaux Arts) there isgenerally a dramatic shift in emphasis as a student leaves
CURRICULUM
the studios of schools and enters those of practice, with the resulting tensions
AND
COMPETENCIES, illuminated during our panel.
US AND UK
FOR NCARB,
An architect is trained on curricular platforms established by educators and -

ARB AND YALE

PART 2 accreditors like ARB to reflect an understanding of required aptitudes on
-

the one hand, while professional licensing establishes a parallel, if different,

level of minimal professional competence on the other. As such, both are
10

rough proxies for what our discipline believes an architect needs to know to
practise. And while the American standard seems to emphasise the synthetic
act of design itself in comparison to its British counterpart, an understanding
of performance, practice and technical issues, as indicated by the emphasis
on environmental systems, building technology and professional practice,

appears to be, on the whole, a more important gauge of whether an educated

student can be properly certified to protect the public's health, safety and
welfare -

the raison d'être for professional certification in the first place.

In a recent exploration of the implications of Al, the computer scientist Stuart
Russell suggests that a fundamental characteristic of human reasoning, and
with it, of human knowledge, is our ability to deploy our understanding in
the service of actions hierarchically with `dozens of levels of abstraction'. 11 He
further invokes the Aristotelean concept of practical reasoning, the idea that
knowledge and actions are rationally deployed in the service of achieving a
12
specific goal. For architects, those goals might be as limited as `make sure

this door swings in the right direction' to as lofty as `try to stir the soul with
this spatial experience'. The knowledge necessary to accomplish the former
requires rudimentary understanding of how doors work, perhaps in the
a

context of a building regulation, while the latter likely demands the collective

insight from a lifetime of work. Either way, each goal shares the common
requirement that the architect reference proper, current and relevant
information and apply her judgement in its use.

A talented designer deploys her skills in managing these `dozens of levels

of abstraction' in ways both poetic and technical, and much current work
in architectural Al research, particularly in the academy, is focused on the
former. Training machine-learning systems on thousands of beaux-arts
floor plans in order to generate new options or using Al strategies to create
novel aesthetic solutions is, of course, a valid avenue of exploration. It is not,
however, the place where such systems are likely to have the most immediate
nor important implications for practice in the near term and thus not the
focus of the balance of this examination.

Developing a complete epistemology -

aesthetic or technical -

is a task for

others, but for the purposes of this

examination we might therefore conclude
that architectural knowledge ranges from detailed technical information,
through an understanding of procedures and processes, to insights gained
from constant reapplication, refinement and synthesis during the course of
career in the profession. Where might Al fit in to such a construct?

HOW MIGHT PROFESSIONAL KNOWLEDGE CHANGE

UNDER AI?

The emphasis of licensing regulators on the more prosaic aspects of practice

and professional competence would seem at odds with both the curricular
emphasis of architectural knowledge as created in the academy and the need
to train architects to face an uncertain future of new conditions, technologies
and responsibilities.
1.3.2: AN
AI-GENERATED
IMAGE

It is not, however, incompatible with the current, or even immediately

forthcoming, capabilities of artificial
intelligence systems, which are very
adept collecting, indexing
at and referencing vast amounts of very `thin' data.
By `thin', I mean epistemologically `shallow' information: information that is
largely devoid of machine-generated meaning or insight. Today's Al systems
can absorb vast amounts of digital information, mostly words and images, but

with scant real understanding of their underlying meaning or implications.

Despite Mario Carpo's assertion that a completely digitally indexed world
makes knowledge universally accessible and therefore the need for
13
techniques like human reasoning or the scientific method obsolete, we have
yet to see machines with even the faintest idea of what a building actually `is'
-

which will require the evolution of cognitive intelligence in order to usefully

deploy the vast potential catalogue of building knowledge.

A brief demonstration tells this story well. Consider the image in Figure 1.3.2 .

Two of today's supposedly `best' Al capabilities are natural language

processing and image indexing and generation. Researchers at the Allen
Institute for Al, who work on understanding human implications of Al systems,
have built generator that creates images from its understanding of a short
a

descriptive phrase. The picture here is the result of the phrase `building in a
city', not a difficult or particularly complex challenge. You can judge the result
for yourself and try your hand at Al-created images on the Allen Institute
website. 14

Stipulating that computers in general, and Al in particular, are excellent

at finding, sorting and cataloguing information and accessing it through
correlation and statistics, 15 I would assert two related potential implications of
the intersections of architecture, digital knowledge, and Al:

1. Most relevant -

andcertainly no surprise to architect readers the

-

world's architectural knowledge is dissipated, poorly organised and

1.3.3: LEXUS/
virtually inaccessible with ease. There is
architectural version of, say,
no
NEXUS LEGAL
the MEDLINE index, which cross-references all medical research, or (here RESEARCH

in the United States) Lexus/Nexus, which provides access to the entire SYSTEM IN
THE UNITED
history of American legal cases. Proper medical treatment would be almost STATES.
impossible without the former, nor could common law jurisprudence like in THERE IS NO
the US or UK progress without the latter. COMPARABLE
SYSTEM FOR
2. Perhaps, therefore, the first role of any Al system aimed at the building ARCHITECTURAL

industry could be getting our data, which are increasingly digitised and DATA

sorted.
 At the point where the of architectural data, ranging from
numerous sources

building product manufacturer's specifications to LIDAR scans of downtown

sites in London, are accessible,
generation Al systems can not only begin
next
to emerge, but proliferate to the
point where both those systems and the
data they catalogue can be of use. As I have asserted elsewhere, 16 measured
performance based on data and analytics likely to become a much more
are

important part of design and construction as it digitises, and a first role of Al

would be to help rationalise the informational platforms necessary to make
that happen.

In doing so, one of the greatest yet untapped resources available to today's
architects may become available to their successors, to wit, the digital project
data, terabytes of which reside on servers in today's offices across the worid,
that are the artefacts of project work. As a practitioner in the 1980s and 1990s,
despite all
our digital drawings and other data, we relied on memory to inform
1.3.4: us, with only our brains to connect our hard-won experience on previous
DIGITAL 17
INFORMATION
projects with decisions we needed to make on our current jobs. While
SOURCES AS today a human architect might have to scan a multitude of digital models to
SUGGESTED BY determine a best practice or trend illuminated by that data, an Al is well-suited
THE BUILDING
to gathering and evaluating such information from a firm's archives.
VENTURES
INNOVATION
NETWORK
That same principle may apply to another intractable issue of architectural
knowledge, the interoperability of information across digital systems. Figure
1.3.4 describes a potential array of today's digital information that could be
rationalised by Al into new outputs. Al processes are well-suited to indexing,
cross-referencing and correlating such data, and as such could become an
implicit interoperability tool for AEC information, and thereby begin to build a
more coherent informational platform for subsequent systems disciplinary -

and computational to evolve. A parallel concept in computer science, known

-

as `glue code', is a precedent. Glue code is comprised of computer instructions

that operate at a low level in larger systems to connect disparate parts of a

larger program, passing data from one subroutine to another. While hardly
exotic, Al may be the glue code of architectural knowledge of the future.

NEW MODELS OF KNOWLEDGE

The broad outline of argument to be built further in the balance of this

an

text should be apparent: there is much to do and a long way to go before Al

becomes even useful for architects, much less an existential threat. Let us
assume that utility is preferable to destruction, that architectural processes

(like many processes of the industry) are increasingly dependent upon

computation, and that truly useful systems must rely on that data. There will
Al

likely be time, perhaps

a a decade hence, where computational and
or more

epistemological coherence will combine for architects, and the days of disparate
standards, incompatible digital processes and inaccessible insight will end.

That work will likely grow from the priorities of practice, which, as argued
above, are largely concerned with the more practical, procedural and prosaic.
And as other parts of the building delivery process, examples of which include
feasibility studies, precision cost-modelling, construction automation and
autonomous digital building operation, evolve through increasing digitisation,
architects will need to understand how to manage and access information and
deploy it inthe service of the new responsibilities and professional obligations
that will result. Where today's architect relies on passing familiarity with an

ever-increasing pool of information combined with professional judgement and

intuition, tomorrow's will likely need the intervening capabilities of Al to design.
BEYOND
PROCESS
AND
AI
4
1.
DESIGN
IN
TRANSFORMATION

» EMERGENT ARTIFICIAL INTELLIGENCE

TECHNOLOGIES GENERATE AN ARRAY OF
OPPORTUNITIES FOR DESIGNERS IN AN
INCREASINGLY DIGITISED CONSTRUCTION

INDUSTRY, WHILE SIMULTANEOUSLY

INTRODUCING TREMENDOUS UNCERTAINTY
IN DEFINING THEIR ROLES AND
RESPONSIBILITIES. IN A yilORLD OF
ACCELERATED AUTOMATED PROCESSES, HOyy
MIGHT THE CURRENT TECHNOLOGIES AVAILABLE
TO THE PROFESSION TRANSFORM AS AI
BECOMES MORE CAPABLE? «

DOI: 10.4324/9781003297192-5
My first real job in an architect's office was in the pre-digital era, long before
computers became ubiquitous in the profession. The managing partner of our
small practice in North Carolina returned to the studio one summer afternoon
after a demonstration of a new technology called 'computer-aided drafting'.
While he deeply sceptical
was of the entire idea, and especially the cost, he
noted that watching the plotter create a drawing was mesmerising, 'like a real
1
draftsman working on one part of the drawing and then another'. He also
made it clear that no machine was going to be replacing anyone there laying
down plastic lead on mylar sheets in our office in the foreseeable future.

My old boss was channelling architects' early but persistent uncertainty

towards technology. While a decade earlier the visionary technologist Nicholas
Negroponte and his colleagues were working at MIT on The Architecture
Machine', mainstream practices like my employer were years away from
anything more sophisticated than a word processing system.2

Somewhere in the space between Negroponte's research and my first drafting

one can find the dual anxieties at the heart of
job tracing flashing details,
this ambivalence: the conviction that our work as architects is a uniquely
valuable contribution, paired with the paranoia that capable machines will
mercilessly replace us -

the source of our profession's angst about machine

intelligence and its putative disastrous effect ondesign process. However, as
Stanford computer scientist, Roy Amara, is purported to have said, 'We tend

1.4.1 AN
EARLY PEN
PLOTTER,
C. 198θ
Machine Learning

to overestimate the effect of a technology in the short run and underestimate

3
long run,' which is certainly the case with architecture's
the effect in the
current concerns about artificial intelligence. The technological circumstances

today are radically different from those of four decades ago, but the
ambivalence justifiably remains.

Resulting concerns about the implications of Al on practice the gamut

run

from design theory to employment economics. Neil Leach wonders whether

Al-enabled computers can be creative, 4 while Antoine Picon gingerly embraces
the opportunities of machine-assisted design and simultaneously worries
about Al restructuring the labour force:

» Until recently, one assumed that automation would impact only
poorly qualified jobs. This might not be the case. Architecture will be
probably among the most severely hit disciplines. The reason for this
high degree of vulnerability is that architecture is among the most
formalized of all the arts. The mechanical part is stronger than in
other domains, and hence the traditional position of the discipline on
the threshold between art and technology.5 «

Labour economist Daniel Susskind {quoting the Governor of the Bank of

6
England) refers to this phenomenon as 'the massacre of the Dilberts'
and
challenges the value proposition of professionals more generally:

argue that the professions will

» (W)e undergo two parallel sets
of changes. The first will be dominated by automation. Traditional
ways of working will be streamlined and optimized through the
application of technology. The second will be dominated by innovation.
Increasingly capable systems will transform the work of professionals,
giving birth to new ways of sharing practical expertise. In the long run,
this second future will prevail, and our professions will be dismantled
7
incrementally. «

There can be a wide gulf between theoretical speculation and the realities
of daily practice, so now is an opportune time to bridge the two, lest the
dismantling begin in earnest. Beyond theorising about the possibilities of this
new technology ignoring in the hope it will pass us by or fighting the inevitable
-

automation of knowledge work we should examine the relationship between

-

design process and machine intelligence to determine how, if at all, they can at
worst co-exist and at best be mutually complementary.
 1.4 Ai and Process Transformation in Design and Beyond

PROCESSES AND PLATFORMS

As the building industry has become increasingly digitised, an emergent

taxonomy of data inputs, algorithmic processes and potential Al-assisted
outputs is emerging. As the industry relies more and more on computers, it
creates large pools of digital information. Those data are made accessible
via so-called common data environment platforms that organise, index and
reference the resulting information. They systematise both inputs and outputs
created by the collection of current automated process tools common to
the industry CAD, BIM, analytical engines, spreadsheets and so forth. As
-

Al tools become available to AECO, a new set of autonomous process tools

will emerge that create results without the direct input and control that
were necessary for their automated predecessors. The combination of all

three process platforms common data environments, automated process

-

tools and autonomous processes will generate a new series of potential

-

outcomes that are 1.4.2:

likely to radically change design process and the architect's
INPUTS,
responsibility for results.
PROCESS
TRANSFORMATION
See Figure 1.4.1 for a diagram of these relationships. AND OUTCOMES
Today's building projects generate huge piles of digital information, as the
tools of building are increasingly computer-based. While in the 1980s and
early 1990s the most relevant digital output was created by either CAD or
word processing tools (primarily by AutoCAD© and WordPerfect, respectively),
some element of almost every part of the design-build-operate-use continuum

is digitised today. The ease with which the architecture profession shifted

entirely to remote work during the Covid pandemic demonstrates that design
work can be largely digitised and is likely to remain so, permanently.

As we saw in the previous chapter, today's artificial intelligence systems,

based largely empiricist platforms, must consume vast amounts of digital
on

information to reach some level of basic competence. Unfortunately, these

data exist today in a wide variety of forms and formats, and are not nearly as
heterogeneous as the records of thousands of games of chess or millions of
pictures of cats.

However, it is possible to categorise digital building information in ways

suggested by current practice and the likely evolution of digital techniques
and processes, in order to develop intelligent input strategies for the futures
of Al in design. Those categories include:

Design representations in the form of models, drawings, studies,

project management documents and other metadata generated as
result of the creation of asset design, coordination and construction.
Representational information is generated by the designers, builders and
building operators to formally memorialise intent. Building information
models (BIM) fall into this category and are important for other reasons

that will be explored later.

Reality capture information that documents existing physical

reality, including topographic and GIS data, as well as video records,
photographs (2D and 3D) and point cloud data from scans of existing
conditions and construction in progress, all of which combine to
translate thephysical world of building into accessible, digital form.

Evaluation, simulation and analytical data, including reference data

sets that support design, engineering and construction management,

analytical models for determining performance and simulations that use

design representations as input to project behaviour of the project as
it is being developed. Structural engineering analysis software, energy

modelling and computational fluid dynamics models are good examples

of what Andrew Witt calls 'technoscientific' models. 8
1.4.3:
A BUILDING
INFORMATION
MODEL (BIM)

1.4.4:
A LIDAR
SCAN

1.4.5:
DIGITAL
ANALYSIS AND
SIMULATION
 1.4.6:
AUTODESK
B36θ
COST
MANAGEMENT
TRACKING
SYSTEM

1.4.7:
AUTODESK
TANDEM
DIGITAL TWIN

Control and coordination information generated by process control,

monitoring and tracking protocols, such as construction management data
systems, project control websites and cost estimating/management systems.

Asset operation, systems performance and use information generated

from building management control systems and other sensors that run
the asset, and internet of things (loT) information depicting the interaction
of users with the asset itself. As built assets contain more and more digital
infrastructure of all sorts, the resulting data will accumulate as a record
against which future designs can be developed.
 In combination these data sources are incoherent, held together only by
the abstract concept of'the project' enterprise that originates, realises,
operates and uses a built asset. Individually, however, they are well
organised, indexed, consistent and thereby entirely accessible as data
sources for machine intelligence systems, and in some cases may be
generated by such systems themselves.

PROCESS TRANSFORMATION

At the centre of the transformation of data streams into new, digitally enabled
outcomes are both existing and new software, platforms and computerised
procedures that work in combination to move the work of architects to its
digital future, comprised of three elements.

1. Common Data Environment: Streams of input coalesce in an

environment called the Common Data Environment, or CDE, which is

formally defined as 'agreed source of information for any given project

or asset for collecting, managing and disseminating each information
9
container through a managed process'. That information structure
creates an index referencing system for organising, locating, versioning
and deploying various digital artefacts of the asset creation process
(referred to above as'containers') and as such provides referenceable
'containers' (using CDE terms) for the various input streams and data
collections that accumulate during project asset creation.

The designers, builders and operators of a project then have two types of
tools to either create or support their respective roles in the creation of a

building: those that are 'automated' and those that are 'autonomous'. It is a

distinction that will be important to define, during the balance of this book,
how Al tools offer different opportunities and threats to the architect.
- -

2. Automated processes: Automated tools are those digital instruments with

which most architects are familiar today and fall directly in the category
of'algorithmic' machine
intelligence. Any software tool today that, as
an outcome of direct human manipulation of its capabilities, generates
results and data based on a specific set of inputs could be characterised
as 'automated'. For purposes of this discussion, we can use the example of
BIM to illustrate this point.
 40 MACHINE LEARNING

An architect developing her design using a BIM tool is largely in complete

control of the data creation process. She decides that her project needs doors,
walls, windows and floors, and creates those elements digitally by instructing
her BIM software to do so. Each of those elements has been algorithmically
encoded to have specific characteristics; a door, for example, has a certain
constrained relationship to a wall and in that sense has specific tectonic
‘intelligence’ that the architect deploys as she creates the model. The process
of deploying that door is automated as she instantiates it into her BIM, and its
representation in various modalities – the plans, elevations, sections, details,
schedules – is automatically populated in those spots for her. Most traditional
software tools available to the architect today are such ‘automated’ processes.

3. Autonomous processes: Artificial intelligence tools, however, are distinct

from their automated counterparts in that they process and generate
results without the direct intervention of the designer, operating
autonomously. When our BIM-enabled architect above searches online for
product information, for example, the search engine’s AI-driven process
combines what it has ‘learned’ as a machine about not just the realm of
building product information on the web based on previous searches but
what her specific interests might be, based on a model of her previous
searches and its conclusions about the objectives she had in initiating the
search in the first place. As such, the product search, and all such tools
that today are based on AI/ML systems, are ‘autonomous’ and distinct
from her BIM authoring experience.

A NEW GENERATION OF OUTCOMES

With increasingly large, better organised data sets accessible to AI-based
systems, we can speculate on the likely set of autonomous opportunities that
architects will see in the next decade as such systems become more capable
and available. These categories, that we will call ‘autonomous outputs’ for now,
form a speculative framework from which we can begin to build strategies for
the implications for the profession, and might include the following:

Design task automation: Procedures and protocols that require the direct
intervention of the designer as likely to be autonomous in the future.
While most code checking is a manual process today, that procedure
can be supported by submitting a digital model to a code-checking tool
that uses AI to evaluate code compliance, combining a more traditional
‘architectural’ model with a technoscientific counterpart.

ML_11_02AW.indd 40 11/02/2022 17:01

Tomorrow, that Al-based code checker could be lurking in the background 1.4.8:
UPCODES
of a BIM process, anticipating code-related problems as the designer
CODE-
creates her scheme. Similar autonomous protocols might reach across the CHECKING
entire delivery and operation life of a building. SOFTWARE

Alternative generation, exploration: Design alternatives were once

created entirely by direct manipulation of design information, like the
cardboard models shown in Figure 1.4.9 Today this process is assisted by
,

scripting, a form of'automating' the control of certain digitally controlled

parameters of a model to vary its characteristics and thereby create a
variety of solutions. Those scripts are sometimes combined with analytical
software such as energy analysis to evaluate and optimise the results.
- -

Scripts create the parameters by which designs might be generated in the

future by Al.

Construction automation: As robotics accelerate with Al-assisted systems

for control, enhanced by computer vision and bolstered by the advent
of industrialised construction processes borrowed from manufacturing,
construction tasks once performed exclusively by human workers will be
augmented, and in some cases,replaced, by autonomous devices on the
job site. An excellent recent example is the PictoBot, an Al-driven robot
 that works alongside an accompanying human supervisor to paint interior

walls, otherwise unassisted. While the patterns, locations and finishes

of such work might be specified in the architect's building information
model, translating the designer's intent for a surface with those particular
characteristics requires a context-dependent operation that can only
mechanised effectively through some sort of computational intelligence.

Autonomous building function: If the architect's and engineer's initial

B1M is a first functional
descriptor of a building that might include the
performance objectives of its systems (and the design of the control
infrastructure that might implement that performance), the operating
air, water, waste and signal systems of that building generate another
digital collection of data that, in concert with real-time analytics, can be
used to calibrate and optimise those systems. Companies that provide
such building infrastructure, such as Johnson Controls, build not just,
for example, an air distribution system but also the digital controls for
the system that communicate with Al-based monitors that memorialise
and optimise the system output and use of energy (thereby reducing
carbon). Of course, the resulting analytical data sets, interpreted byAI,
can also provide insight into the design of subsequent buildings and their
component equipment.

1.4.9:
MANUALLY
CONSTRUCTED
STUDY
MODELS, C.
1993
 1.4.1¸:
SCRIPTING
TOOLS TO
GENERATE A
BUILDING
ENCLOSURE

1.4.11
PICTOBOT, AN
AUTONOMOUS
PAINTING
ROBOT, AS
PROPOSED BY
E. ASADI, B.
LI AND I.
CHEN 1 � B8;
A VIRTUOUS LOOP

The relationship between today's largely automated processes used by

designers and builders (like BIM) will, over time, give way to the autonomous
opportunities of Al-based processes, and in doing so transform both the
inputs and outputs of the building process. As computational platforms gain
independence from their human masters, they themselves will generate
additional sources of data in a potentially virtuous, self-reinforcing data loop.
Properly guided, this cycle might bring the industry many of the advantages of
productivity, efficiency and effectiveness that the designers, builders and users
of buildings alike so desire. Doing so requires those same players to be directive
about the generation of digital information and its intelligent use, with clear
ideas of how Al might translate vast oceans of data into useful knowledge.
SERVICE
OF
SCOPES
5
1.

» DIGITISATION WILL CONTINUE TO

TRANSFORM THE WORK OF ARCHITECTS,
AUGMENTING AND SUPPORTING SOME
ACTIVITIES AND REPLACING OTHERS. WHILE
TODAY’S COMPUTERS DO SO IN A WAY THAT IS
LARGELY PROCEDURAL -
ACCELERATING WORK
AND MANAGING COMPLEXITY -
AI SYSTEMS
OFFER DIFFERENT OPPORTUNITIES AND
THREATS TO THE PROFESSION IN THE SCOPE,
BREADTH AND STRUCTURE OF THE SERVICES
THEY OFFER THE INDUSTRY. AS AI SYSTEMS

EVOLVE, SO WILL THEIR INTEGRATION TO THE

DAILY WORK OF ALL PROFESSIONALS AND,

EVENTUALLY, EVEN ARCHITECTS. «

DOI: 10.4324/9781003297192-6
Machine Learning

JOBS VERSUS TASKS

When Richard and Daniel Susskind argued in their 2015 book, The Future of the
Professions, that artificial
intelligence would eventually replace society's need for
professionals, architects were understandably concerned. Ours is a precarious
profession, smaller, less politically powerful and certainly less remunerated
than our equally threatened brethren in law or medicine, who will certainly put
up a bigger fight before allowing themselves to be automated out of existence.
The Susskinds declared, right at the outset of their treatise, that

»... we are on the brink

of a period offundamental and irreversible change in
the way that the expertise of specialists is made available in society. Technology
will be the main driver of this change. And in the long run, we will neither need
nor want professionals to work in the way that they did in the twentieth century
1
and before. «

They observed that society established professionals to dispense expertise,

and in doing so those professionals horde expert information, control the
access to it and deploy it in ways that can be easily automated by an intelligent

machine. Radiologists will give way to algorithms who can more patiently and
accurately read diagnostic images; attorneys will no longer be needed to search
documents for evidence, prepare routine legal arrangements or even represent
clients in disputes; and architects will not be needed to design, document or
help build projects.
2

As I write in 2021, we are six years past this declaration of extinction, with no

significant encroachment by super-intelligent machines on the work of any

of these disciplines. Best not get too comfortable, however, as technological
change comes much more slowly to the building industry in general, and
architecture in particular. 3

The encroachment on architects and the serviceswe provide is probably

better described in Daniel Susskind's subsequent book, where he argues

that it is
likely that discrete tasks rather than entire jobs will be eliminated by
computers. He extrapolates from the work of economists David Autor, Frank
Levy and Richard Murnane, called the'ALM hypothesis', drawing a distinction
between entire jobs versus tasks, and routine versus non-routine tasks. Routine
tasks (like many of those automated by mechanisation during the Industrial
Revolution) require what ALM theory called 'explicit' knowledge, which is easy to
document, explain and repeat. These are tasks based on 'implicit' knowledge -

requiring, for example, the creativity and judgement of professional architects -

 1.5 Scopes of Service

which will not be automated because it is impossible to capture such work with
4
rules expressed with logical expressions such as algorithms.

Susskind projects this thinking on to the current era of emergent Al to argue

that machines may well encroach into the domains of implicit knowledge:

» The temptation is to say that because machines cannot reason like us, they
will never exercise judgement; because they cannot think like us, they will never
exercise creativity; because they cannot feel like us, they will never be empathetic.
And all that may be right. But it fails to recognize that machines still might be able
to carry out tasks that require empathy, judgement, or creativity when done by a
human being -

by doing them in some entirely other fashion.

5 &#x00AB;

Today's empiricist, deep-learning based systems are beginning to emulate

creativity, but certainly not to the degree that, for example, one might be willing
to entrust that algorithm to replace wholesale the professional judgement of an
architect. In the meantime, it is more important to examine, through the lens
of today's professional services, structures where Al might affect the tasks of
professional work.

MACHINES LEARNING ARCHITECTURE

If today's most capable intelligent machines are based on various strategies for
deep learning, we can evaluate their capabilities by what they can be trained
to do. Mark Greaves's machine capability taxonomy, described in Chapter 1.2 ,

was derived from canonical reference known to many teachers, Bloom's

a

Taxonomy of learning, which defines a hierarchy of capabilities that build from

so-called 'lower order thinking' (like memorisation) to 'higher order thinking'
(like creating something new). Bloom created this approach as a guide to
choosing pedagogical strategies in the classroom, and his students, Anderson
and Krathwohl, subsequently refined the hierarchy to the terms very familiar to
teachers today,
6
and compiled in comparison in Table 1.5.1 .

It can be argued that the most successful deep learning systems today -

the
ones that invent new game strategies and thereby annihilate their human
opponents, credibly translate from English to Japanese or even compose
music or paintings have somehow climbed Bloom's pyramid, having gone
-

far past remembering or even testing data to 'creating' new concepts. Within
certain extremely limited contexts, like the specific rules of the game of Go,
for example, or the learned' patterns of thousands of paintings, perhaps this
is true. A machine programmed with a rigorous set of rules, however, can be
1.5.1: ANDERSON/
LEARNING BLOOM KRATHWOHL GREAVES
TAXONOMIES
CD DESIGN
H CREATE FORMULATE
DEVELOP

X
DEFEND
LU EVALUATE JUDGE
Q
X SELECT
o

X
UJ
COMPARE h
ANALYSE DISCRIMINATE <
UJ
CC
TEST o

CHOOSE
H APPLY DEMONSTRATE
SKETCH/SOLVE H
OT

CLASSIFY
UJ UNDERSTAND LOCATE
a
cc TRANSLATE
o

cc
ID
DEFINE
S DUPLICATE
o REMEMBER
MEMORISE

said to have a ton of explicit knowledge, and as such is 'creating' in only a very
limited way, particularly since the measures of success winning the game or
-

identifying a tumour on a radiograph are so specific.

-

So let us assume, at least for the next few pages of this argument, that
Susskind's thesis of task automation is the most likely implication of Al on
architectural practice in the foreseeable future, and that selected tasks within
the services that the profession provides may well be augmented, accelerated
or even replaced by an intelligent computer. Susskind asserts that those tasks
are easy to identify: they serve explicit goals that can be easily measured (to
determine success) and there needs to be a lot of data for the machine to learn
how to achieve the goal. 7 One might argue that a wide array of architectural
tasks might fit this bill, including questions such as: 'Does this project meet the
firesafety code?' or 'Does this ceiling plenum accommodate all the building
services?' With enough data and proper training, could a computer achieve
these goals? Or, even better, not just answer the question but generate the
required design solutions to meet those needs?
Why not? That stuff either fits in the ceiling plenum or not, and those doors
either have the right fire rating and swing in the correct direction, or they do not.
Assuming our deep-learning Al could study enough ceilings and exit corridors, it
should be able to learn right from wrong, and correct from negligent.

If
only it were that simple, we could start building architectural intelligence into
machines right away. However, there is another dimension to task automation,
what Stuart Russell calls hierarchical planning and management:

Intelligent behavior over long time scales requires the ability to plan and
»

manage activity hierarchically, at multiple levels of abstraction all the way

-

from doing a PhD (one trillion actions) to a single motor control command
sent to one finger as a part of typing a single character in the application cover
letter. 8 «

In order to achieve those that require explicit knowledge

specific goals, even

and have clear, measurable outcomes, an intelligent machine must be able to

deploy a hierarchy of (automated) tasks in an integrated order to reach that
goal, and in doing so assure that the tasks work in concert towards the defined
objective. Very few of the obligations of the architect today, even those reliant
only on explicit knowledge, can be automated in this way, and as such it is
unlikely that large swathes of design service will be satisfactorily automated in
the near to middling future.

ARCHITECTURAL SERVICES

Modern architects deliver their services in a prototypical continuum that begins

with project definition and extends through design to construction. In the UK, such
services are well-defined by the RIBA Plan of Work, 9 and in the United States as
'Basic Services' by the AlA's Owner-Architect Agreement B101. 10 Each are compared
in the diagram in Figure 1.5.2 .

In general, each scope prescribes a route through a standard set of tasks that can
be described in in categories like Project Definition, Design, Production and so forth.
Each phase of the work is comprised of a series of subtasks that differ by phase and
are modulated based on the expectations, deliverables and
professional standards
that govern the architect's services. These subtasks themselves can be categorised

into general buckets like Practice Management, Project Management, etc, and they

span across the phases of service. A rough mapping of a sample of such tasks,
aligned with service categories, can be found in Figure 1.5.3.
1.5.2: If we look at these task components through a lens that combines Susskind's
RIBA AMD
'clear goals and lots of data' criteria and Russell's task hierarchy, and includes
AIA SCOPE OF
SERVICES the ALM's tasks that require implicit knowledge, we can start sorting the service
work of architectsby likelihood of empiricist automation.

Let us call any task component that can be easily defined with a measurable

goal and executed through explicit logic as 'procedural', those that require an
intelligent integration of procedural tasks to reach a goal, even a measurable
one, as 'integrative' and those that are inherently creative, subjective and/
reliant on implicit knowledge as 'perceptive'. Figure 1.5.3 attempts to categorise
each component on this continuum, from procedural through integrative to
perceptive, depending on the work necessary to complete each task component.

As the coded bars suggest, there is very little that today'sarchitects do, even
at this relatively detailed level of examination, that can be characterised as
easily automatable. In fact, much of the technology of today is procedural
(including every piece of software we use), all of which is deployed in the
service of higher order tasks they accomplish. Eliminating humans from the
architectural equation is going to require an enormous jump in capability,
climbing the Bloom Taxonomy while combining those capabilities to accomplish
hierarchically complex objectives. This suggests that architects would better
spend time strategising which procedural aspects of practice might best benefit
from autonomous processes of Al, rather than worrying that our work will be
replaced wholesale by capable machines.

AI-SUPPORTING SERVICES

We can excerpt a few

sample tasks from the list in Figure 1.5.3 in order to test
this thesis that selected
responsibilities in the project process are more suited
to autonomous technology than others. Choosing a few examples that are

primarily procedural by using our classifications above, I speculate on how Al

systems might work in concert with their human architectural counterparts.

What this quick sketch problem suggests is that, at least in the near term, Al

systems will be limited in scope, require enormous amounts of what is currently

unavailable data, and likely augment, rather than eliminate, the central jobs of
architects. This is a reassuring conclusion in the near term, but bears further
consideration as Al systems evolve, in theory, from empiricist to cognitive
capabilities.
1.5.3:
TASK
COMPONENTS
IN
TRADITIONAL
SCOPES OF
SERVICE
1.5.4:
TASK COMPONENT GOAL METRIC PROPOSED AI
AI
POSSIBILITIES PRACTICE MANAGEMENT
ACROSS TASKS
OBTAINING WORK CAPTURE AS MANY WIN RATE MONITORS SUSPECT AND
NEW PROJECTS PROSPECT PROJECTS,
AS POSSIBLE DETERMINES KEY
CHARACTERISTICS, TAGS
PROJECTS THAT ARE MORE
LIKELY TO BE OBTAINED.

GETTING, OPTIMISE STAFF EXAMINES STAFFING

ASSIGNING, THE USE OF UTILISATION ASSIGNMENTS OF ALL PAST
MANAGING STAFF ACROSS PROJECTS, PROPOSES STAFF
STAFFING ASSIGNMENTS BY TASKS, IDENTIFIES
EMERGENT PROBLEMS.

MONITORING ASSURE PROFIT EVALUATES ALL PAST

PRACTICE FINANCIAL AND OVERHEAD PROJECT MANAGEMENT
FINANCIAL HEALTH RECORDS, SUPPORTS
HEALTH FEE PROPOSALS, FLAGS
PROBLEMS IN OPERATION
PROJECTS

PROJECT MANAGEMENT AND COORDINATION

MAINTAINING CONFORM PROJECT HOURLY RATE, BASED ON PAST

BUDGETS AND TIME, STAFF TARGET PROFIT PERFORMANCE, MONITORS
SCHEDULES AND FINANCIAL BY PROJECT, PROJECT OPERATIONS,
RESOURCES ARE UTILISATION PROJECTS OVERRUNS AND
ALIGNED CONFLICTS, PROJECTS
PROFIT, RECOMMENDS
CORRECTIONS

COORDINATING ASSURE DELIVERABLE MONITORS TIMING AND

CONSULTANTS CONSULTANTS SCHEDULES DETAIL OF CONSULTANT
AND OTHERS PERFORM THEIR SUBMISSIONS, RECOMMENDS
WORK ACCURATELY TIMING OF SUBMISSIONS,
AND TIMELY FLAGS DISCONTINUITIES

IDEATION, DESIGN GENERATION

GENERATING EXPLORE THE NUMBER OF RECOMMENDS VARIABLES

ALTERNATIVES SOLUTION VIABLE BASED ON PROJECT TYPE,
OPPORTUNITIES ALTERNATIVES EVALUATES PREVIOUS
SELECTED SCHEMES,
EXCERPTS SUCCESSFUL
SUB-COMPONENTS, DEFINES
SOLUTION SPACE
TASK COMPONENT GOAL METRIC PROPOSED AI

IDEATION, DESIGN GENERATION

EVALUATING FIND THE BEST NUMBER OF HELPS EVALUATE TRADE-

AND SELECTING ALTERNATIVE VIABLE OFFS, DETERMINES
ALTERNATIVES SOLUTIONS ALTERNATIVES PROMISING PATTERNS IN
ALTERNATIVE GENERATION,
PROVIDES OBJECTIVE
EVALUATION OF SOLUTIONS

TECHNICAL PRODUCTION AND ANALYSIS

DETERMINING CONFIRM VARIANCE EXAMINES DESIGN

CONFORMANCE TO THE DESIGN FROM BRIEF DELIVERABLES AND MAPS TO
THE BRIEF CONFORMS TO THE PARAMETERS PERFORMANCE PARAMETERS
PERFORMANCE [AREA, BUDGET, OF THE BRIEF AND FLAGS
TARGETS IN THE VOLUME] DISCONTINUITIES;
BRIEF RECOMMENDS REMEDIATION
BASED ON PAST SOLUTIONS

EVALUATING ASSURE THE TARGET GENERATES ESTIMATES,

AND MANAGING PROJECT VALUE OR MONITORS CONFORMANCE,
PROJECT COSTS CONFORMS TO THE CONSTRUCTION RECOMMENDS COST
CONSTRUCTION COST TARGET ALIGNMENT STRATEGIES
BUDGET BUDGET

COORDINATING ASSURE SYSTEMS VALID IDENTIFIES SPATIAL

SPATIAL AND WORK IN 3D CONFLICTS AND CONFLICTS, SOLVES SIMPLE
TECHNICAL SPACE IN INTERFERENCES PROBLEMS, ELIMINATES
SYSTEMS CONCERT INVALID CONFLICTS,
SUGGESTS SOLUTIONS FOR
IMPORTANT CONFLICTS

REVIEWING ALIGN PERCENTAGE OF EVALUATES INPUTS FROM

CONSTRUCTION CONSTRUCTION CONSTRUCTION CONSTRUCTION SITE,
PROGRESS PROGRESS COMPLETION, IDENTIFIES INSTALLATION
WITH TIME ACCURACY OF DISCONTINUITIES,
AND PAYMENT INSTALLATION COMPUTES VALUES OF
SCHEDULES INSTALLATION, ASSESSES
CONSTRUCTION COMPLETE

CLIENT AND REGULATORY MANAGEMENT

COORDINATING ASSURE THE PERMITS AND EVALUATES CODE

WITH REGULATORS PROJECT APPROVALS CONFORMANCE, IDENTIFIES
CONFORMS TO ACHIEVED DISCONTINUITIES,
REGULATORY RECOMMENDS VARIANCES
CONSTRAINTS
 PRESENT, NEAR FUTURE AND BEYOND

Empiricist Al systems that subscribe to the Susskind definition are beginning

today's
to appear in 'BuildTech'
marketplace, and many focus on narrowly
drawn procedures that demand clearly measurable goals, explicit logic
and plenty of data. The table below describes a selection of some of these
companies that are emerging as of mid-2021:

1.5.5:
COMPANY GOAL LOGIC DATA
AI-BASED
START-UPS IN SPACEMAKER AI OPTIMISE MULTI- ADJUST BUILDING SITE PLANS AND
ARCHITECTURE, UNIT BUILDING DIMENSIONAL ANALYTICAL
2021
2θ21 CONFIGURATIONS PARAMETERS TO OUTPUTS
ON A SITE OPTIMISE USE AND OF SPACE USE
CONFIGURATION

PLANIT IMPACT MEASURE AND REDUCE TRANSLATE ENERGY MODELS,

ENERGY AND WATER BUILDING USAGE DATA, SITE
USE IN A BUILDING CHARACTERISTICS INFRASTRUCTURE
INTO ENERGY INFORMATION
AND STORMWATER
INPUTS AND
OUTPUTS

JOIN ORGANISE, ANALYSE AND COST ESTIMATING

OPTIMISE, AND MANAGE COST HISTORY, LOCAL
CONTROL PROJECT INPUTS AND ECONOMIC
CONSTRUCTION COSTS PROJECT HISTORY CONDITIONS,
FROM CONCEPT TO TOWARD COST MARKET
COMPLETION TARGETS INFORMATION

ENVELOPE DETERMINE THE TRANSLATE ZONING CODES,

MAXIMUM ALLOWABLE BUILDING ZONING PARAMETRIC
BUILDABLE AREA ON REGULATIONS IN BUILDING
AN URBAN SITE TO BUILDING TEMPLATES
CONFIGURATIONS

As these systems become more capable, collecting data and building complex,
correlative data structures within their neural networks, it is likely that their
logics will expand to a wider range of targeted tasks across the architect's
responsibilities. Russell suggests that new ideas were often attributable to
'the three ineffable I's: intuition, insight, and inspiration'. 11 Procedural Al will
augment these critical (perceptual) components of professional judgement,
making the architect's services increasingly reliant upon, and validated by,
analysis and data. One can imagine a day where the architect, having fully
explored a range of options for the configuration of site including the resulting
-

performance data about rental area, storm water draining, zoning conformance
and even construction cost -

can recommend with greater confidence a

decision that they chose with the assistance of a procedural Al. These systems
will remain, however, limited to the lower rungs of Bloom's ladder of higher
order thinking analysis, with perhaps a touch of evaluation since creation
- -

will continue to require the integration of a wide range of information, decisions

and competency. Empirical Al is highly unlikely to reach the top rung.

Beyond the near-term future, the architect's services will need to respond to an

evolving set of new expectations and constraints, including:

λ the automation of construction

λ the increased use of data across all enterprises including those of
clients, and
λ
higher order design aspirations of social equity, environmental justice
and epidemiological safety.

Few of the required services that architects will need to provide to address
these needs can be found in the traditional methodologies of today, be it
through the RIBA Plan of Work or the AlA's definition of'Basic Services'. Deeper
analytical insight, deployment of broad data evaluation and coordination of the
data-driven tasks of a design team with varied (and ever-increasing) numbers of
consultants will require architects to integrate the Als that will support this work,
in the same way in which they manage their engineers today. The challenges of

design tomorrow will be best faced and conquered by people, masters of the
ineffable I's, whose ideas will drive the spaces, buildings and cities of tomorrow,
even if we reach the distant goal of Domingos' Master Algorithm.

1.5.6:
SPACEMAKER
AI, RECENTLY
ACQUIRED BY
AUTODESK FOR
$24θ MILLION
METHODS
AND
MEANS
DELIVERY,
6
1.

>> ARCHITECTS OPERATE IN THE SYSTEMS

OF PROJECT DELIVERY WITH SPECIFIC
RELATIONSHIPS AND OBLIGATIONS TO THEIR

CLIENTS, THEIR CONSULTING COLLABORATORS,

BUILDERS AND, ULTIMATELY, THE PUBLIC.
THEY DEPLOY A VARIETY OF TECHNICAL
SKILLS AND TOOLS, IN COMBINATION WITH
HEURISTIC STRATEGIES THAT COMBINE SKILLS
OF JUDGEMENT, INTUITION AND LEADERSHIP,
TO FULFIL THEIR RESPONSIBILITIES. HOW
DOES THE ROLE OF THE ARCHITECT AS A
PROFESSIONAL CHANGE AS TECHNOLOGY MOVES
TOWARD AUTONOMOUS COMPUTING? <<

DOI: 10.4324/9781003297192-7
SYSTEMS OF DELIVERY

The term 'project delivery' comprises two central aspects of making buildings:

1. The constellation of clients (who create demand for buildings and then
operate and use them), architects (who design them) and constructors
(who procure, fabricate, assemble and build them).
2. How those players are arrayed in a set of professional, informational,
financial and legal relationships defined by their respective roles,
responsibilities and ability to manage risk.

These two factors combine in typical ways according to a formal template

known as a project delivery model.
1
In a perfect world, an appropriate
delivery model would be determined that matched the demands of the project
and the capabilities of the participants. However, since power dynamics and
politics play as much a role in such decisions as technical considerations, the
choice of a delivery system is not always perfectly suited to the players or the

project itself, introducing informational discontinuities at the beginning of a

project that often last for the duration, to little good effect.

PROJECT DELIVERY TYPOLOGIES

1.6.1:
A SELECTION
OF TYPICAL
DELIVERY
MODELS SEEN
TODAY

Over the course of the 20th century, and particularly as architecture evolved
into a bona fide profession, the key players in these delivery models
developed prototypical roles. Clients look to convert capital into a physical
asset, but lack the technical capability to do so, so they hire architects to
define their needs and contractors to translate that definition into a building.
For a variety of reasons that includes the misalignment of interests, these

systems yield unsatisfactory results and as such there has been extensive
2

experimentation in reforming them.

Machine Learning

As described in Chapter 1.3 , that misalignment, combined with the

prototypical roles of the architect and builder, sets each in opposition to
the other. Architects deliver professional services in the form of judgement
and are therefore largely in the business of creating and dispensing ideas
(the design), whereas contractors deliver products, and as such are primarily
tasked with making things (the building). These distinctions, in this context,
are important. If architects have the role of 'proposing ideas' and contractors
somehow 'disposing them', tensions will surely result. However, as the built
environment becomes increasingly digitised, the divide between designer and
builder feels unnecessarily artificial– serving neither the architect, builder, nor
their mutual client. While the production of many of the products that drive the
economy– automobiles, airplanes, consumer appliances, electronics– has been
digitally optimised for decades, the construction industry has trailed far behind.
3
This gap is beginning to close with the industrialisation of construction.

Thus, the autonomous opportunities of artificial intelligence

are likely to have

an impact on both the services provided by architects in delivery systems and

how those services are converted into physical artefacts of the built world.

The challenges of delivery dynamics in modern construction have been

addressed by a variety of strategies, ranging from the reconfiguration of
relationships between design, cost estimating, construction and suppliers
as seen in models like public-private partnerships (PPP) and recent

experimentation with integrated project delivery (IPD), to attempts at

revolutionary means of joint design/construction representation like BIM– or
even a combination of these together. The heart of the problem, however, was

identified as early as 1963 in a report prepared for the UK construction industry:

>> The basic decisions of construction control are

often incomplete
or unduly rushed because necessary information is not available
sufficiently ahead of time, or is not complete enough. On many
occasions members of the construction team could, but do not,
ease this problem by supplying the data that would facilitate the

preparation of fuller and more useful information by others. Building

construction is remarkable among industrial activities for the lack of
detailed information about how it proceeds. Until more is known there
can be no basis for improvement. « 4
 1.6 DELIVERY, MEANS AND METHOD

Each of the players working to deliver a project requires information at

a specific level of resolution, at a certain time, and with a precise degree
of completeness. Each creates, consumes, deploys and/or distributes
this information in order to fulfil their obligations, make a profit and not
assume unmanageable risk. Projects are thus delivered in a constellation of
interdependent yet incompatible collections of information structures and
responsibilities. In the digital age, various technologies lurk below the surface
of delivery, hoping to smooth and improve it. Al is just now arriving to join
that crowded field of suitors. We can expect Al to augment– yet also possibly
replace– the capabilities of the architect, and to generate and organise the
informational structures that connect architecture to construction.

DESIGN INTENT

At the centre of these questions is the definition of the architect's fundamental

responsibility to the delivery process itself. That duty is best understood

through a concept of 'design intent', defined by the RIBA Plan of Work as '(t)he
means by which the design team describes a Building System in a manner that
5
allows a specialist subcontractor to design the system'. If you are wondering
who, exactly, is the designer of any given system in a building, you are not
alone. Architects and their collaborating consultants (engineers and the
like) set out, at a level of detail necessary to be clear about their intentions
(whatever that may mean), information about how the design should look and
operate once it has been completed. There are numerous intermediate steps
necessary before 'the design' is ready to be realised in the field, including the
specifics of materials and assemblies, the exact procedures and processes of
building (called the 'means and methods' of construction) and numerous other
decisions made by the contractors, both big and small. Broadly generalising
about the nature of modern construction on both sides of the Atlantic, it can
be charitably concluded that construction has become too
technically complex
for a single entity – designers or builders – to be
wholly responsible for it. Even
the smallest project has a coterie of subcontractors, product manufacturers
and material suppliers, and cost is always top of mind. In the most optimistic
characterisation of the ideal arrangement under these rules of engagement,
the architect sets out her ideas in sufficient detail to
guide the builder, and the
contractor figures out all the specific particulars. Much is lost in translation.

It was not always such. In the late 18th century in the UK, and as long as 100 years
later in the US, the architect was wholly responsible for all aspects of construction.
Higgin and Jessop described the project delivery model in Figure 1.6.2 .
1.6.2:
PROJECT
DELIVERY IN
THE UK, 18TH
CENTURY 6

Alas, worries about and failures of project coordination and management

ofbudgets, along with the increasing complexity of urban- and industrialage
construction, brought things more into current alignment by the 19th
century with only slight elaboration in the 20th, as seen in Figure 1.6.3
7
.
,

Note the appearance of the quantity surveyor (to manage costs) and the
main contractor (to procure and coordinate the work), both roles that reflect
current practice. As these models evolved, the level of granularity and
resolution of information necessary to build ever increased, and it became
apparent that as the project's interface with the procurement process, the
contractor and her minions – including subcontractors who would fabricate
building systems – was best to 'finalise' the design itself.8

As the building industry has moved from drawings to BIM, the inherent
tensions of such a system were exacerbated rather than calmed
by the
availability of 3D information. Architects that
they had neither the
complained
expertise nor the fee to provide extreme construction detail in their design
intent BIM data, and builders declared that the resulting BIM deliverables
were unsuitable for building. So despite the insertion of a technology designed

to increase transparency and collaboration, the age-old pathologies persist.

Yet there are other forces at play that may, through technology, finally close
this ancient divide. In a market that is increasingly pressed towards more
1.6.3:
PROJECT
DELIVERY IN
THE UK, 19TH
AND 20TH
CENTURIES 9
efficiency and productivity, and dissatisfied with the results of traditional
delivery approaches, the move to digital fabrication and industrial processes
in construction is inevitable. This is particularly true as automation moves the

industry towards what is now being called 'industrialised construction' and

accelerated using manufacturing strategies for building. There is an obvious,
10

resulting demand for higher quality digital deliverables to smooth the path.
However, will architects and their consulting engineers deliver them? And can
Al play a part in closing the professional divide?

PROFESSIONAL RELATIONSHIPS

This question must be examined in the context of all the obligations of the
architect, informational and otherwise. To do so, we can turn the traditional
delivery model diagrams found in Figure 1.6.1 inside-out to look at the specific
connections of the architect in any such structure, depicted in Figure 1.6.2 .

With our hypothetical architect in the middle of her relationships, we can see
four distinct roles required (see Figure 1.6.4).

An agent of the owner, who acts as the client's intermediary in the process,
generating the design and stewarding it, armed with descriptions of her
'design intent', through construction and acting as the owner's proxy to assure
the building conforms to that intent.

A leader of the design team, who orchestrates and integrates the work of
various consultancies in the service of creating a coherent, coordinated and
accurate design which will be passed along to the builders.

A guide to the builder, to articulate the goals of the project and help the
construction team to interpret, clarify and ultimately review and approve
the design intent on behalf of the client as it is translated into more detailed
information to support construction. A subcontractor, responsible for a given
building system, will often create very detailed information in support of the
fabrication of that system (shop drawings) but the architect must review and
approve such proposals before the fabricator may begin.

A protector of the public, including both the specific users of the project as

well as those with whom the architect has nospecific contractual obligations
but nonetheless is responsible for the health, safety and welfare of those who
inhabit her design.
 1.6.4:
ARCHITECT’S
RELATIONSHIPS
IN DELIVERY
MODELS*

*
For further detail on this diagram,
see Figure 2.2.1.

The architect thus either generates, reviews or is subject to the implications

of the vast amounts of information that swirl around even a simple building
enterprise. That information, despite efforts to rigorously standardise it,
manifests in a wide variety of formats, versions and levels of detail, and must
be understood, coordinated and often translated from source to recipient to
be useful. An immediate opportunity of Al – one that is yet unexplored as of
this writing – would be to try to 'understand' the relationships between these
data and help deploy them in a structured, accessible and efficient manner.
 RELATIONSHIPS, DELIVERY AND COMPUTATION

So perhaps Al will day rationalise the ebbs and flows of digital information
one

that follow the relationships and obligations of the architect as a project is

delivered. If enough data can be organised and made accessible, this would seem
a prime opportunity for the next generation of empiricist, deep-learning systems

that are so good at finding patterns and connections between data points.

But how might Al, in the immediate future, either augment or eliminate the
jobs of architects as defined in these four roles? Some early speculation is
summarised in Table 1.6.5 .

1.6.5:
EXAMPLE
OF AI
IMPLICATIONS
FOR THE
DELIVERY
ROLE OF THE
ARCHITECT

The importance of guiding the development and outcomes of emergent Al

systems is apparent in the contrasts between augmentation and elimination
scenarios. While autonomous algorithms may help architects to generate
alternatives and evaluate them, it is technically a very short step to allowing
those systems to make final decisions about the design, essentially leaving the
architect (or someone else) as just the operator. Less likely is the possibility
that the coordination of the design team itself would be entirely replaced
by automation, as this task is as much about human leadership as technical
integration. The move towards industrialised construction creates demand
for far more data than just traditional 'design intent', in the form of higher
resolution design data and procedural templates for sequences of fabrication
and assembly. These may be automated by virtue of intelligent algorithms
in the future and have the potential to eliminate huge swathes of the
architect's role– and compensation – as a result. Finally, in what is perhaps a
central existential threat that will be examined in Chapter 2.3 the architect's
,

fundamental responsibility for the public's health, safety and welfare, as

defined by the understanding, interpretation and implementation of planning
and safety codes, is being automated by companies like UpCodes even today.
Once that job is done by computers, the need for professional certification by
a person– rather than a machine – is called into question.

MEANS AND METHODS

So it appears now that design, construction, and building operation processes

are rapidly digitising; that the
obligations of the players in delivery are likely
to evolve accordingly, particularly those of designers and builders; that
these processes are creating a lot of data that could be consumed by hungry
Al systems looking to learn how to (charitably) help; and that the biggest

gulf in the digital divide is between design and construction, especially as

construction becomes more like manufacturing.

So what is the most useful focus for Al in delivery from the architect's
perspective? Automation algorithms good at memorialising processes (like
are

how to count the number of windows in a building), empiricist autonomous

processes are good at examining data for patterns and, eventually, cognitive
systems might stitch the two together. Bridging the gap from design intent
to construction execution is the most likely target of the combination of all
of these Al-based technological options, in that it will require a broad set of
capabilities beyond a single technology and could, potentially, improve the
overall delivery of projects in the most dramatic and immediate way.

As BIM was beginning to focus the attention of the building industry on

process revolution, the labour theorist, Paolo Tombesi, put forth a proposition
about the changing role of the architect called 'flexible specialisation'. He
argued that themaking of a building– in essence, the underlying delivery
model– was comprised of a constellation of processes and obligations which
required the technical and synthetic design skills of an architect, and as such
the profession should move away from slavish dedication to the design of
objects to a broader, more dynamic set of responsibilities in project delivery
11
(see Figure 1.6.6 ).
While Tombesi did not speculate beyond the technological possibilities of
BIM, his argument extends easily to circumstances where certain data and
procedural aspects of his process nodes are automated in part by Al, and
the architect assumes the role of deploying these capabilities to orchestrate
and integrate the results. Further, the variety of data flows implied by
delivery structure of flexible specialisation demand a rationalising platform to
organise, normalise and monitor that information, a powerful potential use of
Al's capabilities.

Finally, beyond the immediate interdependent processes of construction lies

an extensive supply chain of both raw and processed materials. The architect
and historian, Andrew Rabeneck, has gone so far as to argue that designers
and builders are mere 'contingent players with the new political economy'
12
of the capital flows of construction commodities such as steel or cement.
Control of the intellectual property of industrial production was, according
to Rabeneck, assumed by those who took on the risk of development of
those systems and, as a result, 'industrial control over scientific knowledge
and processes grew steadily among material and product manufacturing
13
companies'. He extends this argument to suggest that this shifting of
knowledge and understanding of the supply chain and its components is part
of the larger loss of value and influence of architects themselves.

As the means and methods of construction are increasingly industrialised, it

is possible that Rabeneck'spurported professional descent of architectural
prowess may accelerate, further widening the gap between intent and
execution. This possibility strengthens the argument that the tools of Al
–which could significantly enhance the architect's understanding of the

processes, procedures and informational requirements of construction, and

allow that insight to inform and support design – can be deployed to either
augment or eliminate the architect in the systems of delivery.
1.6.6:
TOMBESI’S
CONCEPT OF
FLEXIBLE
SPECIALISATION
RELATION-
02
SHIPS

» THE BUILDING ENTERPRISE DEPENDS

UPON A SERIES OF NETWORKS ACTUALISED
TO YIELD RESULTS. PROFESSIONAL
TEAMS OF CONSULTANTS AND BUILDERS
ARE ONE SUCH NETWORK; THEIR RISK
RELATIONSHIPS, ECONOMIC EXCHANGES AND
DATA NETWORKS ARE OTHER EXAMPLES.
HOWEVER, A THIRD SET OF NETWORKS -

THE NEURAL NETWORKS THAT COMPRISE AI/

ML SYSTEMS -

WILL SOON OVERLAY THESE

STRUCTURES AND TRANSFORM THEM. «

DOI: 10.4324/9781003297192-8
VALUE
AND
COMPENSATION
ECONOMICS,
1
2.

»LIKE ALL TECHNOLOGY, AI IS LIKELY

TO PROVIDE NEW OPPORTUNITIES FOR
VALUE CREATION AND PRODUCTIVITY WHILE
SIMULTANEOUSLY ELIMINATING THE NEED
FOR SOME TASKS PERFORMED BY HUMAN
ARCHITECTS. THE DANGERS OF TRADITIONAL
PRESSURES OF COMMODIFICATION MIGHT BE
OFFSET BY NEW CAPABILITIES, AND WITH
THOSE CAPABILITIES, NEW OPPORTUNITIES
TO DELIVER VALUE -
AND INCREASE
PROFITABILITY -
IN PRACTICE. «

DOI: 10.4324/9781003297192-9
Architects operate within the larger economic models of the building
industry, which has long struggled to create consistent value propositions
for its participants. Most buildings are produced under an economic dictum:
achieve the end product by optimising for the single variable of lowest
first cost. Clients often select architects by arbitraging fees for the lowest
price; architects then choose their consulting engineers in the same way.
Contractors are often chosen based upon lowest bid, passing that logic down
the entire supply chain to the far reaches of building product manufacturers,
fabricators and suppliers. The value of the resulting project to the participants
in its creation or, ultimately, the client, is not reflected in the economic deals
that actualise it. As artificial intelligence changes the capabilities, obligations
and outcomes of project architects, how might the economic propositions
evolve accordingly?

CANONICAL MODELS

As came to be understood during the first Industrial Revolution, technology

disrupts patterns of employment and pay. Mechanised farm equipment
displaced plough drivers and their oxen, and industrialised looms put hand
weavers out of business. This is a version of 'creative destruction', as defined
by the Austrian Schumpeter in the mid-20th century, where
economist Joseph
new economic systems often catalysed by new technologies destroy and
- -

replace their predecessors, presumably improving economic performance

1
for everyone while doing so. And although these improvements may prove
temporary, as new competitors touting newer technologies enter the fray, the
concept of creative destruction advancing the marketplace can be paired with
the so-called 'canonical model' of employment described by Daniel Susskind,
where 'it was impossible for new technologies to make either skilled or
unskilled workers worse off; technological progress always raised everyone's
2
wages, though at a given time some more than others'.

If artificial intelligence is thecatalyst for creative destruction in the practice of

architecture, then one can conclude that design jobs will be eliminated by the
capabilities presumably assumes. The canonical economic model argues,
it

however, that those jobs are more than replaced by new ones necessary to
support the creation and support of the new technologies. So you might have
lost your job working in an architect's office doing zoning and code analysis
and planning studies, but you (or someone else) will surely be hired by one
of the many Al-based companies creating software to do that particular
task. And, in theory at least, the quality of the resulting design work created
3
Machine Learning

by the surviving practitioners will improve by virtue of the tremendous new

capabilities of software-assisted design, and the cost of your house or school
will decline as a result.

Those challenges are heightened by the possibilities of AI-related

replacement. Susskind further argues in A World Without Work that it is not
entire jobs that will be supplanted by AI, but rather specific tasks within those

jobs. He suggests that technology will continue to spread the gap between
skilled labour (not easily replaced by machines) and unskilled jobs (where
many tasks may be automated). Architects are generally considered skilled
workers, but there are broad swathes of our jobs subject to such automation,
as suggested in Figure 1.5.3 .

The challenge, as always, is converting that resulting potency into actual value
that is reflected in the economics of practice. Technology notwithstanding,
it has always been difficult for architects to both improve the quality of our
services and the amount of money we are paid to provide them.

EFFICIENCY AND EFFECTIVENESS

Two relatively recent technological shifts in practice demonstrate this

difficulty. As computer-aided drafting tools such as Auto CAD© came to the
fore, largely in the early 1990s, the capital investment in both hardware and
software was significant. To defray these costs, firms often would charge
4

clients an hourly fee for 'CAD services' that was treated as a project expense
not unlike travel or blueprinting. Over time, however, clients got wise to
this idea and refused to pay these charges, arguing that the benefits of the
computer's precision and efficiency accrued to the architect, not themselves.
Of course, this point of view did not reflect the greater accuracy or consistency
of CAD-generated deliverables, nor the increased complexity of design
solutions they were able to create, and architects (once again) failed to convert
the improvement in the quality of their services to an increase in their fees,
which continued to be pressured by lowest-first-cost competition.

The second technological wave brought along by BIM had similar, if

structurally distinct, results, at least here in the United States. Figure 2.1.2
documents performance of the American architectural profession in the two
years on either side of the great economic crisis of 2009. Net revenue for
all AIA firms in 2005 was approximately $27 billion, and the profession had
returned to 96% of that revenue number ($26.4 billion) by 2013, signalling a
 2.1 Economics, Compensation and value

2.1.1:
A HAPPY CAD
OPERATOR AT
HIS STATION
IN THE
OFFICES OF
CESAR PELLI
& ASSOCIATES
[NOW PELLI
CLARKE PELLI
ARCHITECTS],
C. 2000

recovery. If net revenue is arough proxy for work produced, then that same
amount of effort was produced by 16,000 fewer employees (about 11 %) at the
same time that BIM adoption rose almost four-fold:
2.1.2:
COMPARING
REVENUE, BIM
ADOPTION AND
EMPLOYMENT IN
TWO RECENT
YEARS 5

What the data seems to suggest is that these architects increased their
productive capacity, presumably by use of newer technology, by over 10% by
producing the same amount of work with far fewer people. Anecdotal data
from practitioners indicate that while work volume increased steadily after the
Crisis, fee multiples stayed depressed, suggesting that these numbers may
not completely reflect the productivity gains of BIM technology. A more careful

analysis mapping profits, BIM adoption and a proxy for productivity (net fee
revenue per employee) in Figure 2.1.3 indicates that productivity accelerates

with BIM adoption but, sadly, profitability is unrelated. A missed opportunity to

be sure.
2.1.3: So, here is a case where technology (BIM) brought new capabilities
new
PRODUCTIVITY, and efficiencies, yet it appears the profession did not directly benefit in any
PROFITABILITY
AND BIM tangible, economic way. If AI has similar implications, essentially shrinking
ADOPTION 6 the commoditised value of the architect's services, a radically enfeebled
profession is likely to result. However, there are at least three strategies
-

automation, analysis and prediction made possible by the technical potency

-

of emergent AI that could help us to avoid this fate.

PRODUCTIVITY REDEFINED

The analysis above isinvented proxy for the putative improvements in

an

architecturalproductivity in the correlation between technology adoption {the

instrumentation) and the number of staff positions necessary to generate a
certain fee volume. There are no generally accepted measures for determining
such productivity, a result of both the intractable nature of the design process
and a general lack of attention of researchers to such questions, particularly
those in the professional associations. In my professional practice courses,
7

I pose this question slightly differently: if you are a manager of a design

process, exactly how long does it take to have a good idea, and then produce
it? If design is a process of solving Horst Rittel's 'wicked problems', then the
'wicked' nature of the process itself makes it difficult to precisely answer
this question, making the resulting projections of time, effort and expense
similarly intractable.

So perhaps any intelligent strategy for deploying AI tools for increased

productivity should take a different tack, in three dimensions automation, -

analysis and prediction by combining the capabilities of AI as defined by

-

Greaves in Chapter 1.2 Figure 1.2.4

, .

Automation is intelligently replacing linear processes currently performed

by humans with AI-assisted algorithms that might learn those capabilities
as they 'see' more examples.

Analysis is creating more sophisticated means to measure and

understand the results and implications of a design decision.
Prediction is combining the capabilities of automation and analysis to

project a final result of such decision, and in doing so learning, by virtue

a

of selections made by human designers, how to both select and optimise

a result.

The diagram in Figure 2.1.4 sketches this relationship by examining a process

realm ripe for technology-driven productivity: checking a building design for
code compliance.

The hypothetical designer of this AI-enabled future is struggling to validate

thebuilding code compliance of her design, and is deploying a progressively
capable set of machine learning algorithms, which we will call a 'Code
Evaluation System', that combine both cognitive and empiricist strategies.
Early in the life of this system she has created a design, probably in BIM,
so the components can be easily identified spatially and typologically. The
system first identifies all the relevant life safety components relative to, say,
fire exiting, in her model; in this case, the doors and stairs, all of which must
be properly sized, oriented and configured. If the AI missed a component, the
architect identifies it manually, and the AI identifier learns over time to better
find and understand those elements.

Once this
system has learned to find all the pieces of the building's existing
components, it can move to 'Part 2' of this process applying evaluative
-

algorithms to analyse the actual behaviour of each of the elements of the

2.1.4:
THREE STEPS
TO AI-
ENHANCED
PRODUCTIVITY

design. Do the doors swing in the right direction? Do they have the proper
fire ratings? Are the stairs properly configured, within protective enclosures,
large enough to handle the number of occupants? This is the intersection
of automation, analysis and evaluation in Greaves's terms, indicated in the
diagrams as 'Part 1' .

Our machine learning Code Evaluation System is starting, with a combination

of rules-based algorithms and practice with its human operator, to learn
about the interaction of architectural elements necessary to make a safe
building. By extending this system with additional simulative capabilities -

an

overlay of fire
spread, smoke distribution and occupancy behaviour it can -

move from the evaluation of individual elements to the behaviour of an entire

building, evaluating the comprehensive performance of that building under

the simulated conditions of duress. This is the intersection of analysis and
prediction with simulations, in Greaves's terms, or 'Part 2' of the evolution of
our system.
In the ultimate manifestation of this proposed system, 'Part 3' in our diagram,
the system has assembled a sufficiently robust understanding of the
combined characteristics of safe conditions in a building so that it has limited,
but ever-expanding, capabilities to generate a series of safety solutions from
the preliminary model presented by its architect master. If a number of such
architects are training such a progressively capable system, it will 'learn' over

time to optimise the answers.

The system proposed here has several salient characteristics:

» It is working on a discrete problem that is vexing, technical and

measurable in result.

» It is solving problems that do not implicate the architect's design

abilities.

» If it works well, it saves time, effort and brainpower for the architect,
who can then, in theory, either convert the resulting work cycles into
profit (by simply cashing them) or by applying them to improve the
'wicked' characteristics of the design itself. Either way, this architect
has converted the capabilities of her new AI-enabled design assistant
into value, economic or otherwise. 8

BEYOND PRODUCTIVITY AND PERFORMANCE

A sufficiently effective set of AI tools, capable of both optioneering and

accurate prediction, brings the architect's value proposition into an entirely
new realm, entirely decoupling compensation from the value of time sold and

allowing designers to be paid on the basis of the resulting performance of

their buildings. This is a tricky but potentially lucrative strategy where design
services are no longer a commoditised transaction in the delivery chain, but rather
tied closely to project behaviour. The resulting promises should not be bound
to issues of beauty, experience or aesthetics, but rather to the sorts of technical
behaviour that computers could be trained, overtime, to accurately predict.

Productivity and process strategiesnotwithstanding, another immediate value

opportunity in a world increasingly dependent upon AI is the fuel for that
process itself called 'the new oil'. The building industry
9
-

data, sometimes
is notorious for lack of data standards, repositories or means for shared
a

access across itsparticipants. Yet there is a hierarchy of provocative data that

will accumulate across the various dimensions of the building enterprise itself,
described graphically in Figure 2.1.5 .
 For our purposes, the centre of this universe is the data collected within the
architect's office itself, which in sufficient quantity might be useful to the firm
in improving the capabilities of its AI systems. However, since such systems

require enormous amounts of data, it is more likely that architectural data per
se would be collected among the profession itself. Beyond that, training and
reference data for machine learning platforms could be aggregated at project
level, say by building type,or even across the design-to-build marketplace.

Eventually, all such data could exist in the context of the overall building
industry, including design, construction, procurement and operational
information, ultimately referenceable across all the players implementing AI.

2.1.5:
DATA
REALMS
WITH
EXCHANGE
VALUE IN
AECO
There is a tremendous economic value in such a proposed data reserve,
both as a source of
insight and
employment, Training systems requires
as AI

coherent, consistent, 'clean' data and it has been argued that conforming data
requires as much, if not more, effort than building the AI software itself.
10

As the 'canonical model' suggests, curating and managing this data is a likely
source of employment for architects,
new roles emerging from the resulting

creative destruction.

Such data should not, however, be simply gifted to the industry, as it has
tremendous inherent value as well as potential risk of misuse. The creation

of a data trust, mediated and managed by an independent, non-profit

third party, could create the necessary platform for what will likely be huge
amounts of digital building information that will be created in the future,
11
fodder for capable AI systems.

ALTERNATIVE ECONOMIC MODELS

If Susskind is right, the next decade will see an array of architectural tasks
augmented, and in many supplanted,
cases by AI systems. Whether entire
jobs will be replaced as a result is the
subject of some debate. Mastering
a game like chess or Go with very specific rules, precedents and a highly
-

constrained context is hardly a precondition for managing the various

-

tasks, responsibilities and synthetic processes that comprise the abilities of

any capable architect. It may be that, in the aggregate, the production and
analytic functions of practice will be replaced by autonomous computation
to an extent that fewer architects will be required to design the world's
environment. However, conversely, perhaps, released from the drudgery
of the mundane, the power of designers will intensify sufficiently so that
the influence and importance of architects and architecture will expand by
virtue of technology rather than be diminished by it. Careful and purposeful

planning at the intersection of technological innovation and process

improvement is the only way to make such an outcome a reality.
RISK
AND
POLICY
LAWS,
2
2.

>> ARCHITECTS HAVE TRADITIONALLY

BEEN HELD TO A COMMON LAW STANDARD
OF REASONABLE CARE, BASED ON THEIR
RESPONSIBILITY TO EXERCISE GOOD
JUDGEMENT IN CONTEXT. AS THOSE
RESPONSIBILITIES EVOLVE WITH MORE
INTELLIGENT TECHNOLOGY, HOW MIGHT LEGAL

STANDARDS, PUBLIC POLICY WITH REGARD

TO BUILDING AND THE PERFORMANCE OF THE
ARCHITECT CHANGE ACCORDINGLY? <<

DOI: 10.4324/9781003297192-10
While we await a wave of
intelligent machines to change the architectural
profession, AI/ML technology is already in wide use to set insurance rates,
evaluate radiographic images, determine eligibility for loans and other
government benefits, even capture suspected criminals with facial recognition.
And given the inscrutable nature of these algorithms, which 'teach themselves'
to generate results from pools of data (like hundreds of thousands of
insurance claims, or even millions of portraits of potential criminals scraped
from the internet), it is impossible to unpack the underlying logic by which
they make decisions. Worse, that logic is barely comprehensible to the
professionals that deploy it.

Attorney Michele Gilman represents indigent patients fighting for their

rights in the opaque US health system, a bureaucracy made even more
impenetrable now by its use of AI to allocate resources. This was exhibited in
a recent case in Baltimore, Maryland, as described in a recent article in MIT
Technology Review:

» Not until they were standing in the courtroom in the middle of
a hearing did the witness representing the state reveal that the
government had just adopted a new algorithm. The witness, a nurse,
couldn't explain anything about it. 'Of course not they bought it off
-

the shelf,' Gilman says. 'At least she's a nurse, not a computer scientist.
She couldn't answer what factors go into it. How is it weighted? What

are the outcomes that you're looking for? So there I am with my

student attorney, who's in my clinic with me, and it's like, "Oh, am I
1
going to cross-examine an algorithm?'” «

This vignette elegantly captures the range of challenges that the regulators,
the public, the courts, architects and their clients face as the work of
professionals is automated: who or what is responsible for the implications of
decisions made by machines, and can they be sufficiently understood to
ever

place the public's welfare in their care?

Machine Learning

THE DUTY OF CARE AND ITS RISKS

The United Kingdom and United States share a common law tradition
establishing the competency standard for architects (emphasis added):

>> The Architect/Consultant will exercise the reasonable skill, care

and diligence to be expected of an Architect/Consultant experienced
in theprovision of such services for projects of a similar size, nature
and complexity to the Project. (RIBA Standard Professional Services
Contract 2018: Architectural Services) 2

The Architect shall perform its services consistent with the

professional skill and ordinarily provided by architects

care

practicing in the same or similar locality under the same or similar

circumstances. The Architect shall perform its services as expeditiously
as is consistent with such professional skill and care and the orderly

progress of the Project. (AIA Document B101-2017: Standard Form of

Agreement Between Owner and Architect) 3 <<

In each case, the measure of the architect's performance is how it might

compare to that of other competent practitioners in the same circumstances.
This standard is not delineated by statute in any other form, but rather
determined in a court of law after testimony by expert witnesses who posit
correct professional behaviour and application of appropriate precedent.

As examined in Chapter 1.3 the relationship between architects and the

,

public is essentially an exchange of trust: architects are empowered to

make important decisions about the public's health, safety and welfare,
and as such enjoy special privileges of this professional status; presumably
influence, autonomy and compensation. In exchange, they must take personal
responsibility for their actions and cannot delegate that responsibility in the
manner of corporations, where the company, rather than the individual, is
culpable for bad decisions.

Public policy dictates that a design professional be involved in a building

project precisely to assure that the public's welfare is protected in the
built environment and, at least for now, there is a responsible person held
accountable. The requirements differ slightly between the US and the UK: in
the latter, a 'principal designer' must be specified for any project involving
4
more than one construction contractor whereas in most US jurisdictions, a

licensed architect must design any building for human habitation larger than
approximately 275 sqm. 5
 2.2 Laws, Policy and Risk

In performing her duties, the architect assumes two types of risk:

1. Business risk: the possibility that the obligations of service will require
more resources than available in the contract, particularly fees.
2. Professional liability: the possibility that an error in judgement will result
in an assertion of professional negligence in violation of the duty of care.
The business risks of machine intelligence in architecture are more existential,
and these are addressed in a later section in this
chapter. Of more direct,
practical consequence are questions of professional responsibility, duty of
care and the implications of machines making complex decisions either in

support of (augmentation) or in lieu or (automation) human architects.

TECHNOLOGY AND TECHNICAL RESPONSIBILITY

Despite admittedly slow pace, broad swathes of the design and

its
construction supply chain are being digitised today with an array of
procedural, automation and data management tools. In architecture, BIM
is the most prominent of such instruments, but this software falls largely
6

into the category of automating, rather than autonomous, instrumentation.

While certain aspects of its functionality are entirely autonomous (such as the
generation of schedules and views), BIM relies on minute interactions with a
designer to design data, and the attribution of responsibility for
instantiate
the results of this interaction are unambiguous: the architect using the tool is
responsible for its output. This is consistent with the concept of responsible
control established by both the ARB and NCARB. 7 8

Even in circumstances where the responsible designer relies on technology

for substantive portions of technical analysis, there can be no assumption
of delegated authority to software or its producers. A vivid example
of this relationship is in the allied discipline of structural engineering,
where software like Tekla Structures has been used for years for routine
calculations of loads and generation of structural details. Repeated use
(and professional validation) of this technology has led engineers to rely
upon it to produce calculations upon which buildings literally stand, but
the engineer of record is still personally liable for that work, irrespective of
computer output.
Representational tools like BIM, or analytical tools like structural engineering
software, have two important distinctions from their potential AI-enabled
successors:

1. A given set of inputs to such software reliably produces a set of results,

creating predictability and reliability in the relationship between the two.
2. If necessary, the creators of the
underlying software code can explain,
with precision, the foundationallogic whereby a given input produced a
given output, and potential anomalies can be, at least in theory, repeated
and, if necessary, diagnosed and corrected.

A future machine learning-based tool, however, has neither of these

reassuring characteristics. Given that its underlying code evolves constantly
as it is exposed to more data, the algorithm may decide to make different

choices at different times, based on the same set of constraints, parameters

or inputs. And since deep learning systems in particular are notoriously

obscure -

they construct their own internal logics

as -

it will be impossible to
determine why a given decision has been made.

Of course, as an architect becomes

more experienced, she may also make

different decisions about identical set of circumstances and, as a result,

an

generate different results. The difference, however, is that if she makes an

error, we know who to hold responsible. This is hardly the case with software.

THE RESPONSIBILITY OF ALGORITHMS

In fact, the polar opposite is the case. Before you are allowed to access any

piece of commercial software, you are required to acknowledge agreement

with that software's 'End User License Agreement', or EULA. The EULA
explains, in turgid terms, exactly what you can and cannot do and expect
- -

9
from the software you have licensed.

My former employer, Autodesk, has millions of users worldwide, including a

lot of architects, engineers and contractors. At the risk of either inspiring you
to turn the page in boredom or abandon the chapter entirely, I quote at length
below two relevant but nonetheless redacted passages from the standard
Autodesk EULA, with emphasis (bolding) added. While painful to read, they are
illuminating:
 5.2 Disclaimer. EXCEPT FOR THE EXPRESS LIMITED WARRANTY PROVIDED
IN SECTION 5.1 (LIMITED WARRANTY), AND TO THE MAXIMUM EXTENT
PERMITTED BY APPLICABLE LAW, AUTODESK AND ITS SUPPLIERS MAKE,
AND LICENSEE RECEIVES, NO WARRANTIES, REPRESENTATIONS, OR
CONDITIONS OF ANY KIND, EXPRESS OR IMPLIED (INCLUDING, WITHOUT

LIMITATION, ANY IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS

FOR A PARTICULAR PURPOSE, OR NONINFRINGEMENT, OR WARRANTIES
OTHERWISE IMPLIED BY STATUTE OR FROM A COURSE OF DEALING
OR USAGE OF TRADE) WITH RESPECT TO ANY AUTODESK MATERIALS,
RELATIONSHIP PROGRAMS, OR SERVICES (PURSUANT TO A RELATIONSHIP
PROGRAM OR OTHERWISE)... AUTODESK DOES NOT WARRANT:... THAT
THE OPERATION OR OUTPUT OF THE LICENSED MATERIALS OR SERVICES
WILL BE UNINTERRUPTED, ERROR-FREE, SECURE, ACCURATE, RELIABLE,
OR COMPLETE, WHETHER OR NOT UNDER A RELATIONSHIP PROGRAM OR
SUPPORT BY AUTODESK OR ANY THIRD PARTY...

6.1Functionality Limitations. The Licensed Materials and Services ...

are commercial professional tools intended to be used by trained

professionals only. Particularly in the case of commercial professional
use, the Licensed Materials and Services are not a substitute for
Licensee's professional judgment or independent testing. The Licensed
Materials and Services are intended only to assist Licensee with its design,
analysis, simulation, estimation, testing and/or other activities and are

not a substitute for Licensee's own independent design, analysis,

simulation, estimation, testing, and/or other activities, including those
with respect to product stress, safety and utility. 10

To make absolutely sure the user understands the importance of the

Disclaimer, it is printed IN ALL CAPITAL LETTERS, whereas the section on
Functionality Limitations is in mixed case. If it is not abundantly clear here,
there are two important concepts to which an end user agrees when using
this software. First, you use the software entirely at your own risk, and the
company assumes no responsibility whatsoever for its fitness for purpose,
accuracy or other outputs. This is often hard for licensees to swallow, given
the cost of a software subscription.

Second, as if that is not

enough, the EULA specifies that the software is to
be used by trained professionals who will, allegedly, understand its purpose
and functionality. It goes further to explain that use of this tool by such a
professional is no substitute for professional judgement itself. You may have
some powerful digital instruments at your disposal, but the duty of care
still obtains -

and you are still on the hook. Please make sure tokeep your
subscription current.

To be fair to my former employers (and their vendor brethren), there is

some logic in this contractual deal. While responsible software companies
11
extensively test their products before release, it is simply impossible to
anticipate every combination of user interactions and design conditions the
software may need to accommodate during its use. In fact, during my time
negotiating contracts for an architectural practice, I used similar 'suitability for
use' language when clients demanded digital versions of our drawings, as I
had no idea what they might do with that data and how they might try to hold
my firm responsible for its uncontrolled use. I also required them to rely on
the paper versions of same. There is no practical way to predict or manage the
resulting liability of unrestricted data in the wild.

The challenge, of course, is that the complexity of data interactions that

include user inputs and software outputs is a magnitude of complexity higher
when the software is constantly evolving as a machine learning algorithm.
History suggests that software vendors will, in response, move further from
12
responsibility with their next generation tools.

FAILURES OF EXECUTION

Failures in the building industry are common, ranging from the more typical
broken schedules and blown budgets to infrequent but calamitous disasters
like the Grenfell Tower fire in west London. Somewhere in between these
extremes lies the responsibility of buildings to be technically, environmentally,
socially andcontextually appropriate. Surely technology can play an important
part in helping building professionals and especially principal designers
-

such as architects do a better job with such outcomes. For the purposes of
-

this particular exploration, the question remains about the extent to which
the responsibilities of managing the resulting risks of project execution are
increased or diminished by Al-driven tools.

Those risks are sketched in Figure 2.2.1 based , on Figure 1.6.5 , that examines
the fundamental risks of failure when Al is enlisted to assist the architect in
each of her four fundamental roles during project execution:
 2.2.1:
ROLES AMD
PROFESSIONAL
RISK

Consider these risks in the context of the terrible fire at Grenfell Tower in
2017. In that disaster, a small
appliance fire on the fourth floor of a residential
high-rise spread uncontrolled through the building envelope, and 72 people
perished. Components of that envelope had been replaced during a 2015
refurbishment conducted by the building owners and managers and without
direct involvement of the original principal designers. In fact, the requirement
that every project even involve a principal designer was implemented in 2015,
too late to be relevant during the refurbishment project. The Grenfell disaster
was a result of a confluence of technical decisions and errors made by a

combination of players from the client, design and construction/building supply

industries. It may have been the complex interaction of these players that will
ultimately be found to be responsible for the inexcusable deaths at Grenfell,
13
although at the time of writing all those players deny any responsibility.
What is clear from this evaluation is that all the responsibilities of
building
-

andparticularlythe architect -

in execution are correlated to

complex
and ambiguous obligations, tasks, dependencies and outcomes that will
be difficult to delegate to machines, and that in order to assure these
important obligations are fulfilled, humans must remain in charge, EULAs
notwithstanding.

POLICY-MAKING

Building well is undeniably a strong public interest and assuring that

buildings are designed and constructed well is a necessary component of
public policy. The building enterprise continues to increase in complexity,
as clearly demonstrated by phenomena as disparate as the climate crisis (at

a global scale) and the Grenfell disaster (at a project scale). Understanding,

managing and optimising the complex characteristics and interactions of

design decisions, construction strategies, building performance, material
characteristics and even market conditions is a task well suited to big data
and AI/ML. Proper responsibility for managing the application of machine
intelligence in the building enterprise could benefit the public if the
relationship between the two is correctly mediated.

Two policy initiatives are suggested by this logic. Given that each of the
responsibilities described in Figure 2.2.1 might be easily characterised as
a 'wicked problem', it makes little sense for the resulting obligations to be

delegated strictly to machines, which will play an important but not exclusive
-

14
-

part in solving them. Thus, despite some suggestions to the contrary,

professional licensing requirements and statutes should be strengthened so
that educated, experienced architects and engineers can remain at the centre
of projects and accept responsibility accordingly.

Publicpolicy with regard to the development, deployment and efficacy of

thetechnologies upon which we are increasingly reliant must catch up, as
much as possible, with the accelerating pace of machine intelligence. Just as
governments establish, control and enforce regulations about medical devices
and medications, air traffic and aircraft safety, and other aspects of public
health, safety and welfare, it must support the creation of standards and
certification protocols for systems and algorithms upon which the building
industry will rely, including the data trusts that might be repositories for
related information. Much like Underwriter Laboratories in the US is legally
tasked with certifying the
safety of electrical devices, similar structures should
be established for digitally empowered design and building.

BUSINESS RISK AND BEYOND

A final note on the business risks of machine intelligence in the design

professions, which have been alluded to elsewhere in this text, particularly
Chapter 1.5 .

In the short term, certain firms will likely establish a viable but short-lived
competitive advantage by early adoption of Al that will differentiate their
services by capabilities or efficiency. As more firms follow their lead, this

advantage will disappear.

Over the long term, however, architects are likely to face the same questions
ofdisruption and replacement by cognitive automation as other knowledge
15
workers, although the argument above suggests that the timeline of our
destruction may be attenuated. Our demise could be largely eliminated,
however, by using the capabilities of Al technologies to increase the value of
our services and by implication, of the built environment itself and to make
- -

society more dependent on architects and the machines that assist them, in
that order.
PROFESSIONALS
FOR
DEMAND
THE
3
2.

>> IF MACHINES CAN REPLACE KNOWLEDGE

WORKERS, DOES THE BUILDING INDUSTRY,

OR SOCIETY WRIT LARGE, REALLY NEED
ARCHITECTS? ONE THESIS SUGGESTS THAT
MUCH OF OUR WORK AS ARCHITECTS -
WHO
PURPORTEDLY DESIGN VERY FEW OF THE
WORLD’S BUILDINGS ANYWAY -
COULD BE
EASILY AUTOMATED. PERHAPS ARCHITECTURAL
EXPERTISE COULD BE MORE WIDELY
DISTRIBUTED VIA INTELLIGENT MACHINES
RATHER THAN BY A WIDER REACH AND AGENCY
OF ARCHITECTS? ALTERNATIVELY, THOSE
MACHINES AND THEIR INCREASINGLY POTENT
SUCCESSORS COULD BE SEEN AS TOOLS THAT
EMPOWER ARCHITECTS TO TRULY IMPROVE THE
BUILT ENVIRONMENT. <<

DOI: 10.4324/9781003297192-11
When the UK Government's Health and Safety Executive (HSE) enacted
the requirement that building projects require a principal designer, it
acknowledged that built assets manifest in three distinct phases: planning,
execution and use (followed, presumably, by eventual demolition). Before
2015, when the HSE CDM (Construction, Design and Management) regulations
went into effect, designers were considered desirable, but not necessary,

participants in the delivery process. This is in stark contrast to US law,

where any habitable building of significant size must be designed by a
licensed architect, 1 and is likely an indication of the political strength of the
construction industry in dictating the terms of building delivery.

HSE makes the case for the necessity for architects (as one option) crisply in
their regulation, indicating that principal designers must (emphasis added):

» Plan, manage, monitor and coordinate health and safety in the

pre-construction phase. In doing so they must take account of
relevant information {such as an existing health and safety file)
that might affect design work carried out both before and after the
construction phase has started.
»
Help and advise the client in
bringing together pre-construction
information, and provide the information designers and contractors
need to carry out their duties.
» Work with any other designers on the project to eliminate
foreseeable health and safety risks to anyone affected by the work
and, where that is not possible, take steps to reduce or control
those risks.
» Ensure that everyone involved in the pre-construction phase
communicates and cooperates, coordinating their work wherever
required.
» Liaise with the principal contractor, keeping them informed of any
risks that need to be controlled during the construction phase. 2

Missing, of course, from this otherwise nifty summary of the need for
designers is anything about the quality of the resulting artefact, including
its suitability for relationship to context, expressive nature, or even
use,
environmental social appropriateness. These are results that clients
or

who hire architects clearly desire, even though they could meet the CDM
requirements with any party willing to assume the role of principal designer.
Even so, consider whether an intelligent machine in the foreseeable future
 Machine Learning

might 'plan, manage and coordinate health and safety', 'help and advise
the client', 'eliminate foreseeable risks' or 'ensure everyone communicates
and cooperates'. If these things were even remotely possible, I suspect
construction managers, who perform many of the same tasks during their
phase of the work, will join architects at the unemployment office. However, is
even considering such a future a good idea?

THE PLAN OF WORK

As an architect who practises in the United States, I have always admired

the clarity, flexibility and elegance of the RIBA Plan of Work. It evolves over

time, changes to reflect delivery and computational realities, and presents

clients and collaborators with very clear definition of both the arc of a
a

project's lifecycle and the architect's possibilities to participate in a project

entirety. This
in its structure contrasts starkly with its American counterpart,
defined by the American Institute of Architects (AIA) as 'Basic Services' in
several stolid and largely unhelpfully described phases like 'Schematic Design',
which have remained largely unchanged for decades despite the fluid nature
3
of practice, particularly in the digital age.

A comparison of the two structures, derived from Figure 1.5.2 can be seen ,

in Figure 2.3.1 in which I have also diagrammed the subtle but important
,

implications of the Plan of Work: it can be abstracted to understand the work

of architects in four 'super-stages':

1. Project Definition, comprised of the stages of work necessary to create

the overall
approach for the project
2.3.1: 2. Technical Development, where the approach is refined as an engineered
SUPER-STAGES
and buildable idea, technical insight integrated and detailed information
OF THE
ARCHITECT’S
in preparation for construction created
SCOPE OF
SERVICE
 2.3 THE DEMAND FOR PROFESSIONALS

3. Execution, where the asset is constructed by the contractor

4
4. Use, when the building is occupied.

In contrast with the 'bottom up' analysis of services we looked at in Chapter 1.5 ,
let us consider the implications of machine intelligence on these super-stages
and how the work of the architect might be either augmented or replaced.

PROJECTIVE RESPONSIBILITIES

At the heart of the architect's value in creating the built environment is what
I will call, for purposes of this discussion, her 'projective responsibilities' to

generate and instantiate ideas about the future state of the building she is
designing. To do this job, her conceptual skills must range from broad-scale
predictions about the implications of her building in the city, to the minute
choices of finishes in the interior; this is a very broad remit, particularly when
each of these decisions should support an integrated vision of the project.

Design, as apart from construction, is essentially an Enlightenment era idea

about how humans should make things, and was defined for architecture
by Leon Battista Alberti around 1450, when he wrote that buildings should
be 'conceived in the mind, made up of lines and angles, and perfected in
the learned mind and imagination' 5 and then executed without deviation by
6
builders, whose job was to convert the
design projection into a built reality.
And while it has been argued that architecture has long relied on structured,
formalised systems ranging from Vitruvian Orders to off-the-shelf storefront
window systems 7 that make for easy pickings by Al, the range of imaginative
obligations demanded of today's architects, particularly in the project
definition
super-phase of the work, defy systematic outputs by algorithms.
There is
simply too much ambiguity, need for judgement and trade-offs, and
demand to solve wicked problems at a variety of scales.
Today's AI/ML systems see patterns, particularly those that humans cannot
divine, by virtue of the vast ocean of data available to those systems, but they
are only projective to the extent they have specific computational templates

to follow, like the rules of chess or Go. Their ability to project the future
state of, say, a building design, is a function of past experience (as defined

by data generated by other projects) and whatever rule set they have been
programmed to follow. Completely missing in today's systems is the ability
to reason counter-factually or to understand causality (why something did
or did not happen) versus correlation (something might happen because,

statistically, it has happened under the same circumstances before). Ironically,

this argument is made best by computer scientist Judea Pearl, who invented
the statistical theory called Bayesian networks, upon which today's correlation-reliant
machine learning neural network systems are based. 8

Pearl is convinced that truly intelligent machines are not possible until they
can reason causally and climb what he calls The Ladder of Causation'
(see Figure 2.3.2 ), which has three rungs;

Association (where understanding is a function of observing data);

Intervention (where actions are possible based on projecting the
implications of the future), and Counterfactuals {where understanding
leads to ideas based not just on actions but counter-factual assertions
about the future). Today's systems are firmly planted on the lowest rung.

The ladder is a sophisticated the theories of Roger Schank,

successor to

mentioned earlier. Schank believed that the true test of any intelligent
machine was the ability to reason inferentially -

a different take on Pearl's

assertions of Intervention and Counterfactual reasoning. He suggested that
the logic of inference was at the heart of human language and cognition. If a
machine could draw a conclusion by 'understanding' the logic of implications,
it could be said to be reasoning like a human. Unfortunately, this thesis was
strongly hobbled by the crude machines programmed in the 1970s and the
we

need to explicitly code all the resulting cognitive logic. Schank's thesis faded
with the 'Al Winter' of the 1980s.

Pearl's ladder is not simply a software roadmap for 21 st-century Al systems,

but rather yet another assertion about the fundamental nature of human
intelligence, one that differentiates humans from other species and accounts
for our accelerated progress:
 2.3.2:
PEARL’S
'LADDER OF
CAUSATION’

» What humans have that other species lacked was a mental

representation of their environment a representation that they could
-

manipulate at will to
imagine hypothetical environments for planning
and learning (they have) the ability to create and store a mental
...

representation of their environment, interrogate the representation,

distort it by mental acts of imagination, and finally ask the 'What if?'
kinds of questions.9«
Pearl speculates that the most important ideas in history were the result
of 'wildmodelling strategies' (like the Earth astride a giant turtle as early
astrophysics) and that algorithms that merely fit data to scenarios could never
generate such concepts. 'Imagining hypothetical environments', or, in more
10

prosaic terms, 'Project Definition', is the central value of a good architect and
demands third-rung talents that are unlikely to be achieved by machines
anytime soon. In fact, should machines reach the second rung, we might
achieve tools that help architects speculate on 'what if, the more modern
versions of today's analysis software, that could be hugely helpful to the
human architects occupying the top of Pearl's ladder.

TECHNICAL COMPRESSION

If the value of projectivity firmly roots human architects in the responsibilities

of project definition, our future involvement in technical definition is less
clear. Once design strategies have been defined choosing and spatially
-

coordinating systems, generating coherent and coordinated documents,

analysing performance and cost, and organising and transmitting information
from designers to builders are formalised protocols that are more suited
-

to automation. As digital project histories become available as data sources,

empiricist systems or even those reaching Pearl's cognitive capabilities of the

'intervention' rung, may assume more responsibility for technical integration.

Digitally automated construction and fabrication systems can consume data

systematically generated by their Al counterparts on the design side. All of
which is to suggest that the technical development super-stage is much more
subject to Al replacement than its predecessor in project definition.

However, there is an important caveat here. Health and safety considerations

are at the forefront of HSE's requirement for a principal designer, and while
Al tools are likely to get much smarter in guiding projects towards safer
outcomes, there is little benefit in allocating that responsibility exclusively
to algorithms, which are far from being capable of making decisions and
therefore taking the corresponding obligations.

As philosopher Daniel C. Dennett suggests, computers are not conscious

entities, cannot suffer consequences of failure and do not assume human

obligations. He asks if it is possible to 'Give me the specs for a robot that could
sign a binding contract not as a surrogate for some human owner but on its
-

11
own ...
morally responsible agent.' If there is
as a any lesson from Grenfell,
it is that the complexity of the building enterprise -

in all dimensions of
 construction, supply chain management, certification and regulation,
design,
and human behaviour is beyond the ken of machines. As Dennett further
-

posits: 'We don't need conscious agents. There is a surfeit of natural conscious
agents, enough to handle whatever tasks should be reserved for such special
and privileged entities. We need intelligent tools.' 12

HEALTH, SAFETY, WELFARE AND THE FUTURE

In summer 2020, academic administrators (like me) were facing the

daunting task of
planning the upcoming school year in the face of a global
pandemic. Under the rubric of 'everything looks like a nail when all you
have is a hammer', our team at the School of Architecture began looking
for anarchitectural strategy to address an epidemiological crisis. We knew
that if we could translate the spatial demands of public health parameters
(established by our colleagues in the School of Public Health) we had a chance
of maintaining some semblance of the education our students so reasonably
expect. After a summer of careful planning, we produced an operating plan -

50 pages of architectural and engineering analysis that allowed us to open

-

our building and give our students access to our facilities while working in

studio. Every classroom was evaluated and rearranged so those faculty who
could teach in person might do so. Air systems were evaluated and adjusted,
and a safe occupancy schedule established and enforced. The school year
ended without almost no positive cases among our students.

Architects in the 21 st century face an array of similar challenges that must

draw the profession away from its obsession with making beautiful objects
for thewealthy. At the top of this list is climate change, followed closely by
housing inequity, structural racism in the built environment, especially cities,
even questions of modern slavery in the building supply chain. 13

In the aggregate, these challenges comprise a new definition of the public's

health, safety and welfare, an idea that catalyses the need for a principal
designer in the UK and is the basis of
professional licensure in the US. Facing
these questions as a set of spatial challenges demands the essentially human
capabilities of Russell's intuition, insight and intuition that are unlikely to be
provided by machines soon, and certainly not in time to attack these problems
with the 'wild modelling strategies' they will demand. As Dennett so wisely
14
posits, Al producers should be 'making tools, not colleagues'. Architects are
needed as never before, empowered by those tools.
2.3.3:
YALE
ARCHITECTURE
STUDIO SPACE
PLANNING
(COURTESY
APICELLA
+ BUNTON
ARCHITECTS)
TRAINING
AND
CERTIFICATION
EDUCATION,
4
2.

>>
TECHNOLOGIES, BY THEMSELVES, RARELY
BEND THE ARC OF ARCHITECTURAL EDUCATION
OR CERTIFICATION. NEW TOOLS CHANGE THE
NATURE OF THE ARCHITECTURAL PROCESS,
ALBEIT VERY SLOWLY, AND THE INTELLECTUAL
INFRASTRUCTURE OF THE PROFESSION -

SCHOOLS, PROFESSIONAL ASSOCIATIONS

AND LICENSURE BOARDS -
ARE EVEN SLOWER
TO RESPOND. THE ADVENT OF INTELLIGENT
MACHINES AND THE DEMANDS ON 21ST-CENTURY
DESIGN WILL REQUIRE THESE INSTITUTIONS
TO RETHINK HOW ARCHITECTS ARE TRAINED,
THE CRITERIA BY WHICH THEY ARE CERTIFIED
AND THE INTELLECTUAL INFRASTRUCTURE OF
THE PROFESSION ITSELF. <<

DOI: 10.4324/9781003297192-12
I have argued to this point that while Al systems are likely to both augment
and impinge onthe work of architects, they are unlikely to replace us as
designers, a capability that will require the development of artificial general
intelligence (AGI). The data scientist Herbert Roitblat correlates AGI with just
such an ability to attack the 'wicked' problem of design, suggesting that 'To
have a truly general intelligence, computers will need the capability to define
1
and structure their own problems,' which is an excellent way to characterise
the value of a talented designer.

There are, of course, many problems facing architects in the day-to-day

business of design that are well structured and suited to emergent, empiricist
machine learning systems. This will be increasingly true as architecture and
the scientific disciplines necessary to attack issues of climate change, material
performance or even the socio-economic dynamics of building become more
entwined. Although data-dependent analytical tools that can help architects
with these issues are a far cry from the representational tools like CAD or
BIM, which have been treated largely as instruments of expression in both
the academy and the office, their emergence suggests that the two poles
of intellectual infrastructure of architecture academic and professional
-

institutions -

must plan for the resulting implications.

LEARNING OBJECTIVES

We will touch first on the well-trod and contested ground between those who
prepare architects for practice in the academy and those who establish the
criteria for accredited curricula and, eventually, professional certification that
leads to licensure. The uneasy truce between providers of architectural talent,
the certifiers of competence and the consumers of that talent in daily practice
is underpinned by a basic tension: what does it mean to educate a competent

practitioner? To crudely summarise the positions of the contestants,

educators argue that architecture is best understood as a form of culture,
and it is difficult, time-consuming and expensive to train students in the
design skills necessary to achieve that end, so there is little time to do much
else. Licensure certifiers define competency in terms of the legal demands
of the public's health, safety and welfare, and demand technical competency
first.
They are bolstered by professional associations that protect the brand
and potency of architects and steer toward capabilities with marketplace
relevance. Meanwhile, the practices just want folks who can function the day
they first sit down behind their assigned computer.
 Machine Learning

The means by which these competing aspirations place demands on the

architect include:

» curriculum and accreditation constraints (in the academy)

»
testing and experience and continuing professional education (by the
licensing authorities), and
»
qualifications for membership and more continuing education (by
professional associations).

While the standards and structures differ slightly, these arrangements are

largely the same in the UK and US, and compiled in Figure 2.4.1 .

2.4.1:
INTELLECTUAL
DEMANDS
ON THE
ARCHITECT
 2.4 Education, Certification and Training

These competing constituencies, with our architect in the middle, take

divergent positions about what constitutes competence and use different
instruments to enforce them. The emergence of new technologies, at least
to date, has changed these positions little, if it all, since those technologies
have largely been technical means (software) to representational ends
(drawings, models). However, once computers begin to augment (or replace)
competence, the conditions on this pitch will need to change.

TERMS OF TOOLS

The advent of BIM demonstrates this intransigence and holds potential

lessons for the upcoming era of Al. As that technology began widespread
adoption around 2010, at a point where the software was sufficiently mature,
machines capable and when propellants like the UK BIM Mandate came into
focus, there was widespread conversation about how a new means of data-
rich representation might empower architects and the building industry writ
large. Early academic enthusiasm, however, soon faded, and despite massive
investments in software and hardware, design pedagogy remained largely
2
unchanged. While BIM is begrudgingly taught in most schools, it is done so
as a necessary evil to prepare students for practice, and widespread research
on the possible implications of BIM for design pedagogy are somewhat
3
unexplored territory.

I would argue that the disinterest in fully engaging design curricula

BIM in
is indicative of the larger inclination of educators to see technologies only
as tools or instruments, although there are specialised, post-graduate BIM

technical degrees for those so inclined.4 Since BIM joined a crowded field
of'representational' instruments (used directly to depict a design), and that
software is deployed largely in the service of form- and image-making, this
conclusion is understandable. At my institution we are careful to say that we
teach principles and theory, not tools, and there are no parts of the curriculum
5
(save one) where learning tools can result in credit toward the degree.

This arm's-length relationship, however, will not serve either students or the
overall professional well in the long term, and the advent of machines that
can do knowledge work is best faced now by educators and other leaders

of our profession. As intelligent machines move from efficient depiction and

data management (CAD, BIM) to analysis, insight and evaluation (AI/ML), the
academy must face two parallel obligations:
1. Instantiating a different source of design insight into design pedagogy.
2. Providing the foundational research around the data sources and uses

thatnewly intelligent instruments require.

These two objectives are self-reinforcing; by laying out the proper terms of
intelligent digital competence in the enterprise of design, the academy can set
the direction for their use in the marketplace.

it is important to draw the distinction here between the development of

algorithms that design things themselves and those that provide a supporting
role. Subscribing to the earlier argument that we need 'tools, not colleagues', 6

some of the most interesting research today in the architecture/ML nexus looks

at what we can learn from algorithmic generation of building plans or room

configurations, like that of Stanislas Chaillou of Harvard Graduate School of

Design (see Figure 2.4.2 ).

This work is important in that it may yield insights into building organisation,
or even optimisation. It allows designers to see problems in a different light,

but it does not solve those same problems. Stuart Russell suggests that'... Al
research has focused on systems that are better at making decisions, but that is
not the same as making better decisions'. 7 Thus, this research is not likely to be
of the most immediate use in a world when structured, scientific and technical

interrogation and evaluation of design work will be increasingly demanded.

Finally, graduates often introduce new technologies and methods

recent
to practice, wheretechnology hesitancy is a function of conservative work
processes, long project schedules, costs of implementation and low profit
margins. If Al strategies for architecture can be incubated in the academy,
perhaps they can be carried into regular practice by the most technologically
adept new employees.

PROFESSIONAL PREROGATIVES, CERTIFICATIONS AND DEMANDS

On leaving the academy, our future architect enters the domain of the
profession and its disparate masters: registration authorities and professional
associations and accreditors, both with roles in certifying competence, and
therefore in defining the knowledge and skills that a professional architect
must possess. However, where technology once relieved architects of the

obligation for such prosaic obligations as consistent hand-lettering and line

weights on drawings (that were not of much concern to certifiers), or even
2.4.2:
AI-BASED
PLAN
GENERATION
 properly coordinating callouts on CAD drawings, its role in both dispensing
and generating knowledge that supports the design process should draw
more careful attention.

In the US, the National Council of Architectural Registration Boards is the

organisation, comprised of licensure authorities of 54 US jurisdictions, that

sets certification standards for architecturalregistration. part of their
As

protocols for establishing testing and experience standards, as well as

influencing the accreditors of architectural schools, they prepare a 'Practice
Analysis of Architecture' to determine what skills practising architects and
educators believe are necessary for recent graduates and licensees to master,

whether in school or otherwise. Even in 2012, the year the last analysis was

completed (see Figure 2.4.3 ), there was strong agreement that technological
skill was necessary to be deemed competent, if only towards various
representational ends (like drawings).
2.4.3:
NCARB 2012
PRACTICE While we wait for the 2020 analysis, it is safe to presume that the 2032
ANALYSIS OF
analysis will include knowledge and skills of Al applications, and the overall
KNOWLEDGE/
SKILLS summary of competencies will reflect the idea that certain functions of today's
RELATED TO architects, particularly those related to building science, will be performed by
TECHNOLOGY 8 machines and managed and integrated into project process by architects.
The lessons of the UK National Level 2 BIM Standard, which stipulates both
the information outputs and performance levels of the design process when
powered by BIM, may be instructive. 9 It was created by an industry consortium
and eventually evolved from a UK-only template (PAS 1192) to an international
standard (ISO 19650). While some practitioners in the UK may avoid working in
BIM, any government-funded project requires it for the large number of industry
projects they fund, and it is a matter of time before BIM techniques and data
strategies will be instantiated into the standard of professional care, expected of
competent practitioners.

The Level 2 standard was based larger industry strategy established by the
on a

Cabinet Office as part of a national economic agenda to improve building and UK

construction competitiveness globally, decrease climate impacts of construction
10
and make government building more efficient. The resulting technology
requirements were designed to provide process and outputs towards those ends,
and acknowledged the importance of information strategies:

» Information management using building information modelling

can enable dramatic improvement in delivery and performance
efficiencies by catalysing increasingly innovative ways of working
across the built environment. /Is an information-based industry, this

approach is helping to support better strategic decisions, improved

predictability through better risk management and when coupled
-

with a soft-landings methodology (Annex B of this guidance) can lead

-

to certainty of operational outcomes and improved learning. 11

«

The particular implications for the instruments of professional certifiers -

licensing criteria, testing, experience and continuing professional develop

-

should be derived both organically from emerging practice standards and

duties of care, and strategically from cross-industry efforts like Level 2 BIM
that set technological objectives and standards of use from agreed goals.

This suggests that the most important role of certifiers in establishing the
use of machine intelligence should not stem from determining or driving the

particular and unique requirements of AI-supported architectural practice but

rather in concert with larger industry collaborators who can prioritise the most
important objectives and use of these new systems. This approach will be not
just desirable but necessary, given that, unlike BIM tools that generate data
by virtue of its use, AI/ML systems require large, well-curated data sets for
training and optimisation, and those data are most useful when contributed
across the entire delivery chain. The training obligations demanded by
certifiers do not serve only as requirements for professionals, but for the
machine learning systems as well.

Professional associations, like RIBA or AIA, have more vague ideas about
certification for membership, as such considerations are primarily designed
to assure that members have credibility with the
marketplace. Continuing
professional development requirements ask only that member architects
be regularly exposed to a broad spectrum of technical and professional
concerns, stipulating general categories (such as 'Health, Safety and Welfare',
for example, in the US) and numbers of hours of attendance. Over the next
decade, as architectural clients rely on AI-based processes in their business
or government operations, it is possible that those clients will ask the same of

their architects, and professional education and CPD certification are sure to
follow the desires of the customer base.

PRESSING PROBLEMS OF INSIGHT

The marketplace is currently pressing the building industry about carbon and
climatechange, and that challenge gives us a good opportunity to speculate
on how the academic and professional platforms of architecture might

respond in the time of machine intelligence. Imagine the following scenario:

In 2032 the Ministry ofJustice issues a request for proposals for a new

headquarters building in Westminster. The project is to adhere to the recent

update of ISO 19650, the so called'Intelligent Level 3 BIM' mandate, and reflect
best practices in responsible environmental design, including certifications that the
project will be net zero, generate at least 500 construction jobs and be free of any
evidence of modern slavery in labour or material practices.

The Construction Industry Council, with participation of RIBA, ARB and a

consortium including Cambridge, Liverpool, Manchester and the
of universities
Architectural Association, has certified a complement of artificial intelligence
platforms for the evaluation of embodied carbon and the labour supply chain.
Those systems were built, based on research in the consortium, by several
companies in the M4 Corridor, the so-called 'England's Silicon Valley', and trained
with the National Building Data Trust, created and curated by the Infrastructure
and Projects Authority in 2026 with data provided by the UK's global design and
construction industry.
The RFP (Request for Proposal) stipulates that submitting architects must be RIBA
certified in machine data analytics for Level 3 outcomes, and able to deploy AI
platforms to evaluate submitted schemes for compliance to climate, economic
and labour performance. The team must also include at least two architects
with registrations in the recently approved SPLC (Speciality Professional Licence
Certification) created by ARB in climate change modelling and supply chain
management evaluation. Several graduate candidates from Leeds and Liverpool
have sat for, and received, these special registrations.

The SPLC program has been created in concert with a new definition of Principal
Designer established in 2029 by the Health and Safety Executive that includes
environmental and labour equity in the responsibilities for that designation. Many
firms have been experimenting with platforms, smartTALLY (see Figure 2.4.4)
two AI
and buiidFRDM (see Figure 2.4.5), that assess Level 3 BIM schemes for embodied
carbon and forced labour, and collect information about design decisions and
strategies that are contributed to the National Building Data Trust. The HSE has
further stipulated that projects for human habitation larger than 300 sqm must
have an assigned Principal Designer who is a licensed architect, causing some
consternation amongst the country's construction/design managers none of -

whom have been certified in data-driven design methods.

This admittedly rosy scenario presupposes that our profession organise itself
in ways as yet unseen to accomplish three ends:

1. a strategic focus on key social challenges

2. the intelligent deployment of technology in service of that focus

3. the integration of academic, professional and technical resources used in

concert toward those ends.

In these circumstances, Al technology is not just an available tool, but more

importantly catalyst of change in standards, relationships and processes, and an

a

opportunity to synchronise education, certification and subsequent professional

training in technology suitable for 21 st-century design and building.
2.4.4:
TALLY
CARBON BIM
ASSESSMENT
TOOL THAT
EVALUATES
EMBODIED
CARBON IN
BUILDING
MATERIALS
FROM A
DIGITAL
CONCEPTUAL
DESIGN IN
REVIT

Option 1: Corrugated Shingle Cladding

Option 2: Translucent Panel Cladding (selected)

Results Per Life Cycle Stage, Itemised by CSI Division
2.4.5:
FRDM, AN
AI-BASED
TOOL FOR
FINDING FORCED
LABOUR IN THE
MANUFACTURING
SUPPLY CHAIN
RESULTS
03

»JUST AS ACCELERATING STORAGE AND PROCESSING

CAPABILITIES OF TODAY’S COMPUTERS HAVE VIVIFIED
MACHINE INTELLIGENCE, THOSE CAPABILITIES IN TURN
BRING THE POTENTIAL OF EXPONENTIALLY MORE PRECISE
AND INSIGHTFUL PROTOCOLS TO DESIGN, SUGGESTING THAT
COMPUTERS (AND, ONE HOPES, THEIR HUMAN MASTERS) MAY
BECOME MUCH MORE SKILLED AT NOT JUST PROJECTING THE
FUTURE STATE OF A BUILDING, BUT IN PREDICTING ITS
BEHAVIOUR. «

DOI: 10.4324/9781003297192-13
OBJECTIVES
THE
1
3.
DESIGN
OF

» THE INEVITABLE ADDITION OF ARTIFICIAL

INTELLIGENCE TOOLS AND ‘BIG DATA’ IN
THE BUILDING INDUSTRY WILL CHALLENGE
TOMORROW’S ARCHITECTS TO RE-EXAMINE
THE CLASSICAL DEFINITION OF THEIR
OBLIGATION TO DEFINE THE ‘DESIGN INTENT’
OF A PROJECT. AT THIS STAGE, ARCHITECTS
LEAVE MANY OF THE IMPLICATIONS OF THEIR
IDEAS -
FOR CONSTRUCTION AND EVENTUALLY
BUILDING OPERATIONS -
TO BE RESOLVED
AND REFINED BY OTHERS. AI TOOLS ARE
LIKELY TO ACCELERATE THE USE OF DATA-
DRIVEN EVIDENCE TO AUGMENT THE PRECISION
OF DESIGN, GIVE ARCHITECTS GREATER
UNDERSTANDING OF PROJECT PERFORMANCE
ACROSS THE MANY DIMENSIONS OF PROJECT
DELIVERY AND RAISE EXPECTATIONS OF THE
RESULTS. THE PROFESSION CAN DECIDE TO
SEE THESE OPPORTUNITIES AS AN ADVANTAGE
OR A THREAT TO ITS AGENCY. «

DOI: 10.4324/9781003297192-14
When Alberti defined the architect's design of a building as 'conceived in the

mind, made up of lines and angles, and perfected in the learned mind and
imagination', he centred those responsibilities on the abstract projection of
1

a future state, which, according to Carpo, was to be translated from idea to

concrete resultby virtue of the architect's 'sound advice and clear drawings'. 2
Carpo, channelling Alberti, further explains:

»
Designers first need drawings and models to explore, nurture and
develop the idea of the building (those) models should also be used
...

to consult experts and seek their advice; as revisions, corrections and

new versions accumulate, the design changed over time; the whole

project must be examined and re-examined... The final and definitive

version is attained only when each part has been so thoroughly
examined that any further addition, subtraction or change could only
3
be for the worse. «

In Carpo's interpretation we find three important architectural strategies:

1. The use of abstraction, in the form of models and drawings, to

memorialise ideas.
2. The incorporation of outside expertise in completing the design so it is

suitable for construction.

3. Designing in successive iterations to refine the project until it could not
4
possibly be adjusted for the better.

Today's architects use much the same approach, bolstered by various digital
armaments, but what happens when those tools become agents of design?

THE AI-ENABLED DESIGNER

Let us assume that an architect by, say, the year 2030 has a complement
of Al-enabled tools at her disposal, along with significant advancements
in the resolution, precision and flexibility of modelling platforms that one

hopes would be the logical successors of today's BIM. While there may be a
conceptual breakthrough, some time in the future, in what I have described
as 'cognitive' Al platforms that can reason inferentially about the complex

interactions that comprise a building, let us assume that by 2030 we only are
at the point of useful architectural versions of 'undeniably single-mindedly
successful' platforms like today's language and game-playing software, for
5
Machine Learning

6
example GPT-3 or AlphaGo. Such systems would be tightly tied to design
modelling/representational platforms and their data, and receive training
from other information sources like engineering systems, real-world data
collection about context from LIDAR or GIS, construction management sources

that describe process and results from contractors, and building operations
data from existing projects controlled by sensor-driven building management
and control systems. These systems are likely to be semi-autonomous, cloud-
based agents that operate in the background of the architect's process,
appearing when the architect demands some piece of insight or analysis.

While it is
impossible to accurately predict what will comprise this new set of
Al-empowered tools, Table 3.1.1 summarises a few speculative suggestions
designed to sketch the potential future of autonomous, Al-based tools.

Notably absent from this list, save perhaps the

last item, are systems tasked
with generating entire design solutions (at any scale) for a project. A central
thesis of this book is that such systems will not be useful until far in the
future -

if at all. They are unlikely to provide useful insights and present an

unnecessary existential threat to architects. The world is already populated

with many not-quite or barely competent buildings; the creation of a design
generator capable of even simple buildings is likely to have unintended
andunpleasant consequences for the profession. And with so many other
opportunities, energies are best focused elsewhere augmenting the design -

process to improve the and results of and 7

performance design building.

THREE NEW DIMENSIONS OF DESIGN

These Al examples suggest that Alberti's components of thedesign process

-

representation, iteration and

expertise -

will be transformed as intelligent

systems that augment (but do not replace) the central role of the designer.
First, as has already been seen in our now data-rich world, the extensive
availability of information in digital form, combined with the predictive and
analytical power of Al systems, will make the role of evidence in supporting
design decisions much more apparent. While, as I have argued previously,
the credibility of design decisions stemmed primarily from the (presumably)
sound judgement and intuition of an experienced architect, those 8

judgements will need to be substantiated, at least in part, by evidence and

analysis to back them up. The built environment has traditionally disgorged a
collection of ambiguous, heterogenous data sets, but the ability of Al systems
to divine and understand patterns within it gives architects the opportunity to
 3.1 The Objectives of Design

AI-EASED TOOL CAPABILITY TRAINING DATA

ZONING AND PLANNING ANALYSER EVALUATES CONFORMANCE TO PLANNING CODE, RECORD OF

SPECIFIC PLANNING AND ZONING VARIANCES, EXISTING BUILDINGS
CONSTRAINTS OF THE PROJECT IN SIMILAR JURISDICTIONS

BUILDING CODE EVALUATOR CHECKS EMERGING SCHEME FOR STANDARD CODES AND LOCAL
CONFORMANCE TO BUILDING/LIFE IMPLEMENTATION, EXAMPLES OF
SAFETY PERFORMANCE CONFORMING CONFIGURATIONS
FROM OTHER PROJECTS

SPATIAL COORDINATOR CHECKS 3D COORDINATION OF CLASH ANALYSIS OF CURRENT

BUILDING ELEMENTS WITH DESIGN SCHEME AND EXAMPLES
AN UNDERSTANDING OF WHEN FROM OTHER PROJECTS OF
CONFLICTS ARE PROBLEMATIC APPROPRIATE AND PROBLEMATIC
INTERFERENCES IN PLENUMS AND
OTHER SPACES

CARBON IMPACT CALCULATOR COMPUTES GENERATED AND ENERGY AND CARBON

EMBODIED CARBON IN THE DESIGN CALCULATIONS AND STANDARDS,
RECORDS FROM OTHER PROJECTS,
SUPPLY CHAIN AND MATERIAL
DATABASES

MEANS AND METHODS EVALUATOR EXAMINES THE CONSTRUCTION PROCEDURE

CONSTRUCTABILITY/SEQUENCING INFORMATION FROM BUILDERS AND
AND PROCEDURES NECESSARY TO PRODUCT SYSTEM MANUFACTURERS,
CONSTRUCT A GIVEN BUILDING CONSTRUCTION SIMULATION ATS.
ELEMENT TO TEST ITS VIABILITY
DURING DESIGN

PRODUCT RECOMMENDATIONS AND EXAMINES CHARACTERISTICS AND PAST PROJECT RECORDS,

SPECIFICATION GENERATOR RECOMMENDS POSSIBLE MATERIALS MATERIAL AND SPECIFICATION
AND PRODUCTS AND COORDINATES DATABASES, PRODUCT
THE NECESSARY SPECIFICATIONS MANUFACTURING INFORMATION,
SUPPLY CHAIN AVAILABILITY
DATA

COST MONITOR AT THE PROPER LEVEL OF CONSTRUCTION COST DATA FROM

RESOLUTION9 PROJECTS THE COST PROVIDERS AND PAST PROJECTS;
OF CONSTRUCTION AND OPERATION BUILDING OPERATIONS DATA FROM
EXISTING PROJECTS

SUPPLY CHAIN AVAILABILITY WORKING WITH THE PRODUCT MANUFACTURING AND SUPPLY
PROBE RECOMMENDER, EVALUATES THE STREAM SHIPPING AND
SUPPLY CHAIN CONSTRAINTS OF A MANIFESTS, CERTIFICATION
PRODUCT SELECTION, INCLUDING DATA, LABOUR STANDARDS INPUTS
AVAILABILITY, COST AND FORCED FROM LOCAL CONDITIONS, MODELS
LABOUR ISSUES. OF PAST PROJECTS

SKETCH PROBLEM SOLVER BASED ON PARAMETERS SET BY EXAMPLES OF SIMILAR

THE DESIGNER, GENERATES CIRCUMSTANCES IN DESIGN
ALTERNATIVES FOR A GIVEN MODELS OR EXISTING BUILDINGS
DESIGN PROBLEM AND PROVIDES
MEASURES OF PERFORMANCE AND
SUITABLE FIT FOR PURPOSE

3.1.1:
FUTURE
AI-BASED
TOOLS
generate and leverage just such evidence. And since many of today's clients
rely on Al data systems to run their enterprises, architects will be expected
to do the same to substantiate the decisions that form the design.

Alberti has asserted that the architect should produce designs that are
perfectly ready for a builder to enact physically. At the end of the design
process, Carpo interprets that:

» This is when all revisions and the

final blueprint (as we would have
said until recently, both literally and figuratively) is handed over to the
builders. Thenceforth, no more changes may occur. The designer is no
longer allowed to change his mind, and builders are not expected to
have opinions on design matters. They must build the building as is -

as it was designed and notated. 10 «

The master builder of Brunelleschi's ilk, a central repository of all things design
and construction and the maker of every decision, gives way to the architect,
generator of complete, immutable and clearly depicted ideas.

If only. There is a profound mismatch between Enlightenment aspiration and

the realities of modern construction, where design documents are a frequent
source of contention and the architect is relegated as a subconsultant of
a contractor in the name of better control, while the builder has not just
opinions about design, but control over it. However, the advent of Al gives
architects an interesting opportunity to close this gap and realise design ideas
with great fidelity, if not greater control, of the design-build relationship.

If architects will benefit from Al systems focused on specific tasks, it is equally

likely that our colleagues in construction will see similar progress, with Al-
driven systems automating aspects of construction in the field through
robotics and industrialised methods of digitally driven mechanisms. An Al that
plays brilliant chess likely be repurposed to control a robot that installs
can

curtain walls, for example. In doing so, the procedural knowledge of building
that architects are oft accused of lacking will have been instantiated digitally
and will be accessible to them as an evaluative/performative measure of
the efficacy of their design. 11 While sharing the 'mind and imagination' with
an Al, the architect can perfect the design with substantial new, accessible

understanding of how it can be built.

Finally, deploying evidence in combination with the predictive powers of Al
systems will make performance a profound objective of the modern design
process and its deliverables. Architects today operate in a mode that might
be characterised as 'implicit' performance: our work process is organised and
calibrated to produce physical artefacts (drawings) that lead to an object (a
building) that is hoped to achieve certain ends once complete, and those goals
are rarely explicitly defined as measurable objectives or outcomes. Implicit in

the design process are expectations that those goals will be achieved, costs
will be met, documents completed on schedule, materials properly specified
and codes conformed. At the same time, the resulting buildings consume
resources and produce carbon, require regular maintenance and staffing,

organise the circulation of people and materials, and most importantly

create platforms for their owner's objectives: students learning, patients

healing, goods selling. They contribute (or detract from) the environment,
economic health and social fabric of their locales. As Al systems learn from
the data derived from the built environment, and to the extent that these
characteristics model in predictive Al systems, architects get the ability to
'explicitly' design projects towards improved ends, demonstrating a priori, by
virtue of the resulting simulations, that such outcomes are the result of the

design itself. And while the earliest opportunities may be of a more limited
technical nature (as suggested in Table 3.1.1 ), more sophisticated systems will
model and evaluate larger, more complex contexts.

This power of prediction is perhaps the most important implication of Al

for the design process. It seems likely that Al technology, and the building
industry data necessary to train it, will be in great supply by 2030. Prediction,
according to a recent analysis of Al implementation in business, 'takes
information you have, often called "data", and uses it to generate information
you don't have'. 'Information you don't have' might be the watchword of
12

today's building industry.

BREAKING FROM ALBERTI’S DESIGN PROTOCOLS

Alberti'sdesign protocols of representation, iteration and instantiation of

expertise have not been dramatically transfigured by the first few waves of
technology, including even BIM. Representation is still centred on creation
of drawings, and while much of the resulting information is now digital, the
means by which it is generated by an iterative refinement informed by
-

outside expert consultants is today exactly as asserted c. 1450. The third

-

wave of digital tools, those driven by intelligent computation, will follow, but

dramatically break from these traditions.

3.1.2:
AI-SUPPORTED
STRATEGY
FOR GENERATING
OFFICE
LAYOUTS
Al-enabled representation will stretch the definition of 'model', which once
referred to a scaled physical artefact (maquette), a mathematical simulation
of geometry (CAD) or a parametric, meta-data-infused simulation of building
components in three dimensions (BIM). An Al evaluation of a design will be a
predictive model that will substantiate design decisions that are represented
by more traditional means, but at the same time it will expand the range
of the architect's instruments of service. Those Al models will be required
for both validation and as contributions to larger data repositories that can
further train other Als.

An architect I once worked with compared the current iteration process of

design to a circular staircase. Looking straight down on it, you appear to be
going around in circles, but with each cycle you rise slowly toward the goal.
In today's design environment, climbing those process steps is accelerated

by the inherent flexibility and accessibility of digital models and enhanced by

emergent strategies called generative design, where computerised scripts
generate alternatives by varying specific characteristics of a scheme. Properly
constrained -

so as to not lead to the 'systematic generation of useless

alternatives' as César Pelli once described the misuse of CAD Al-enabled
-

generative design will set its own constraints. It will be informed

by the logic
of previously approved schemes as a set and
simultaneously provide
training
evaluation of its own results. Design exploration will still demand choices by
humans to make decisions that solve 'wicked' problems, but the process will
be much more intelligent and systematic.

Generative strategies that connect representation, iteration and instantiation

may be the most important implication of Al-supported design processes of

the future. Today's modelling and analysis tools can only episodically optimise
limited parameters of a design challenge adjusting the dimensions of a solar
-

shade to limit exposure and thereby reduce the size of a cooling system, for
example -

but this is hardly a strategy for the complete design of the building
enclosure. Over time, Al systems will be able to manage multiple variables while
evaluating design representations created by the architects and engineers,
instantiating expertise while simultaneously recommending alternative
solutions that meet design objectives. As designers select solutions, these
systems will come to learn which combined strategies are best, and thereby
improve their performance -

and that of their architect masters.

Consider the case of an architect coordinating the design of a mechanical/

electrical room (MER) in her project. That room, never big enough to satisfy
the engineers who fill it with complex equipment and connections, must be
integrated in 3D with the balance of the project. She must assure the systems
all fit, there is sufficient room for servicing the equipment, and that none of
the architecture, structure, lighting and fire suppression interferes with the
locations and pathways of the systems. An intelligent Al, trained on many
similar rooms and the components that often fill them, can do more than
just check the MER layout for clashes (a common feature of today's BIM) but

identify potential operational problems, recommend potential configurations,

even strategise howto sequence and install the systems. Rapid simulation

and evaluation of these issues will speed the process and make a successful
solution more likely, improving over time as the Al systems 'learn' what is
best. Our architect can use the additional time made available to resolve the
proportions of the facade.

Early Al efforts in the 1980s

purported to create 'expert systems' that would
memorialise knowledge and insight of humans in computer code. Neither the
theory nor the technology were up to the task. By contrast, in our Al-enabled
design future, sources of expertise that are today provided almost exclusively
by human consultants will be greatly expanded by the analytical insights that
computers can provide. While conceptual decisions that are more strategic
in nature are best dispensed by human experts who can evaluate systems

approaches and large-scale choices, specific outputs can be provided for

specific tasks with particular inputs that result from well-understood rules. In
the examples above, evaluating the particulars of life safety code compliance
(and, with it, the arrangement of fire sprinklers or rated corridors) might have
been provided by consultant with that expertise. Our designer of the future
will apply an Al overlay to her design to yield much the same results, faster,
and allowing for further iterations and resolution of the scheme in real time.
Her decisions will be catalogued by the assisting Al and guide successive work.

NEW OBJECTIVES, NEW OUTPUTS

At the heart of the potential future changes in the objectives of the design
process wrought by Al lie the implications of data and its use. Today's
architects use digital tools to create data, translate it into various forms
like drawings or specifications, and dispense it as evidence of their design
decisions. When machines can consume, create and deploy data to assist
in those decisions, the models they create extend to and are entwined with
the descriptors of the
design itself. The
resulting capabilities can make the
results of design more precise, transparent and predictable. The balance of
this section will explore strategies for enabling these capabilities in ways that
further empower architects accordingly.
3.1.3:
MECHANICAL
INFRA-
STRUCTURE
MODELLED IN
BIM FOR A
MODERN
HOSPITAL
CURATING
AND
CONSUMING
CREATING,
2
3.
DATA

>> DIGITISATION OF THE BUILDING INDUSTRY

MAKES DATA MORE PORTABLE, TRANSMITTABLE
AND, TO SOME DEGREE, FUNGIBLE. AI
SYSTEMS WILL REQUIRE LARGE SWATHES OF
SUCH DATA, FIRST FOR TRAINING AND THEN
TO PERFORM. A GIVEN PROJECT GENERATES
DATA IN A WIDE VARIETY OF FORMATS,
SCALES AND LEVELS OF RESOLUTION BY
DISPARATE PLAYERS WITH A VARIETY OF
MOTIVATIONS TO SHARE IT, OR NOT.
AS AI CHANGES THE DEMAND FOR AND
CONSUMPTION OF DIGITAL INFORMATION BY

ARCHITECTS, HOW DOES THAT CHANGE THEIR

RESPONSIBILITIES AND PROCESS? <<

DOI: 10.4324/9781003297192-15
In 1994, my former employers at Autodesk organised about a dozen
companies across the AECO industry in an effort to address a growing
concern. The company's increasingly ubiquitous CAD platform, AutoCAD©, was

becoming the data standard for the building industry with their proprietary
file format, .DWG. At the same time, the company was building a global
ecosystem of third-party developers to create additional functionality on top
of the AutoCAD© platform, and other software companies were looking to
consume AutoCAD© DWGs in their ownsystems. The file format itself was
understandably defended zealously by Autodesk, as much of their intellectual

property was contained in each file.

The 12 companies, including architects, mechanical equipment manufacturers,

engineers and at least one real estate developer, called their consortium
the 'International Alliance for
Interoperability', and opened it to all members
by 1995. Their mission was to achieve the ability to seamlessly move data
1

between applications without translation or the need to duplicate DWG

functionality. Rebranded BuildingSMART in 2005, the global consortium writes
and distributes a data standard called Industry Foundation Classes (IFCs),
an attempt to create a common denominator exchange that can transport

relevant data between any software that has been written to generate or
consume it. BuildingSMART's efforts have turned from CAD data (DWG) to BIM
past years, creating standard data exchanges and libraries designed
2
in to
make BIM more open.

Achieving interoperability standards in the building industry, even given

BuildingSMART's admirable global efforts, is a daunting task. By the late 1990s,
DWG had become the standard of data exchange in design and construction

powered by two divergent realities: the industry was turning to digital tools,
primarily AutoCAD©, but even more importantly, the vector of information
exchange for architects and engineers was still largely drawings, rather
than more robust data. This made the transmission of information via IFC
relatively simple, using common definitions of geometry, lightly dusted with
meta-data about that geometry. Given the explosion of digital tools today, the
problem, however, is much more complex, and accepted non-proprietary data
standards for the AECO industry have not been established. 3

MORE SOFTWARE, MORE DATA, LESS COOPERATION

The move to BIM was but one part of the inevitable digitisation of AECO writ
large. Building things is an information-rich enterprise, and architects among
others had to wait until machines and networks were sufficiently powerful to
Machine Learning

handle the required data. Cloud computing, high-speed interconnectivity

and capable mobile devices all have spurred a veritable explosion in digital
tools, formats, processes and even hardware. Standards for building
4

information, analogue or otherwise, vary widely from country to country.

This makes the mission of BuildingSMART even more challenging, despite
their new branding as enablers of the 'full benefits from digital ways of
5
working in the built asset industry'.

However, what became apparent as architects and their collaborations

began using larger collections of digital tools was that data exchange and
relationships were not just a function of technical standards. Each constituent
of the building process has its own contract and risk models, tools, data
expectations, representational schema, content and business expectations,
and is digitising at its own pace and in its own terms.

In 2004, the US National Institute of Standards and Technology published a

report purporting to identify an annual cost of $15.8 billion lost to software

interoperability in the capital projects industry. The report gained significant
notoriety, and certainly brought attention to the process inefficiencies of
non-interoperability among software used for building. It failed, however, to
address the underlying structural questions in the industry – discontinuity
in business models, disaggregation in the supply chain, conflicting risk

management strategies and lack of optimisation incentives in project delivery

– that plague construction, all of which are strong disincentives to work
6

together with data.

Those same challenges of cooperation and integration can be seen more

broadly in the structure and use of data in the building industry, and especially
for architects. Being careful not to create too much responsibility or risk for
construction, while managing limited fees with which to produce information,
design data is held closely when released at all. Other learned' professions
carefully generate, curate and consume knowledge about their disciplines:
databases of case law, medical research about treatments and outcomes or

pharmaceutical efficacy. An attorney researching the legal precedents for her

client can explore a complete, cross-referenced database of every relevant

legal decision in the history of jurisprudence (see Figure 1.3.3 ). Architects have
access to no such central data. The move towards interoperability was a plea

to play nicely together, but not to share any toys.

3.2 Creating, Consuming and Curating Data

3.2.1:
AN IMAGINED
FUTURE OF
BIM
INTEROPERABILITY,
IN 2013 7
3.2.2:
PRECONDITIONS OF INTEROPERABLE DATA
COMPONENTS
OF A
The advent of machine learning-based Al systems demands that our industry
POTENTIAL
not just share toys but builds a new sandbox in which to play with them. This
DATA TRUST
FOR THE is the first and most important precondition of moving towards and taking
BUILDING
complete advantage of the power of Al for architects and other players in
INDUSTRY
the building enterprise. The ability to leverage the potential of Al lies in the
profession working closely with industry partners who might also benefit, and
sharing data to do so in responsible ways.

The problem, of course, is the other external factors, not the least being
underlying motivations (or lack thereof) to share data. We will address some
of the structural risk and reward questions in Chapter 3.5 but for purposes of
,

this discussion the issue is diagrammed in Figure 3.2.2 which represents four
,

hypothetical architectural projects that are otherwise unrelated. While there

is some motivation to allow the project data generated within your office
to roam more freely in the domain of the project, there is no structure nor
incentive to organise or share it beyond that limited use.
At the centre of the diagram is a proposal to address this issue, which must
be solved in order for Al to have any real chance of adoption or use in
architecture or the allied building disciplines: a cross-industry data trust that
would be the steward of a global building industry information resources.

The concept of data trusts has evolved in the last several years to address
questions of information coherence, privacy and fiduciary responsibility
in circumstances where individuals contribute their personal data that is
then used, for commercial purposes, by third parties. A data trust is an
independent, third party who collects, manages, anonymises and provides
access to such a large-scale collection of data:

»
Typical use casesdata sharing are fraud detection in financial
for
services, getting greater speed and visibility across supply chains,
improving product development and customer experience, and
combining genetics, insurance data, and patient data to develop new
digital health solutions and insights. Indeed, the research has shown
that 66% of companies across all industries are willing to share data.
Nevertheless, sharing sensitive company data, particularly personal
customer data, is subject to strict regulatory oversight and prone to

significant financial and reputational risks. «

8

While it is
unlikely that 66% of architects, or contractors for that matter,
would be willing to share data today, the benefits of access to a central global
repository of project data, properly anaesthetised for attribution, would
be too great to pass up, as both a useful reference tool and the necessary
information infrastructure to begin Al in earnest. And of course, the challenges
presented by European data sharing standards must be overcome.

Such a data trust would,

by necessity, need to be cross-disciplinary and
include information from designers, builders, subcontractors, product
manufacturers and suppliers, and operating building owners; the entire
supply chain that builds assets. For architects, there is marginal utility in
an 'architecture only’ data set, as it is likely to be sparse and inconsistently

curated. And if my experience of years as a technology vendor is any

indication, most architects are highly sceptical– and unwilling to pay for–
new, disruptive technologies, especially in comparison with their colleagues
in the building supply chain – engineers, contractors, subcontractors. The
architecture business is small, relatively unprofitable and generally unwilling
to invest in disruptive technologies, creating a classic chicken-and-egg
dilemma in establishing the necessary foundations of artificial intelligence.
 Thus cross-industry cooperation is necessary for both data assembly and Al
platforms. With the exception of the decision by Autodesk to invest, initially,
$133 million in Revit and then develop that platform first for architects, there
has been scant history in the technology of major investment in new software
or platforms for architects, per se. So while there might be short-term comfort

in the idea that perhaps it is too expensive to invest in Al that would replace
architects as individual contributors, there is clearly benefit in data sources and
Al platforms suited for the entire industry and it is more likely that investment
will be made by vendors to address a broader market of customers.

BUILDING AND USING TRUSTABLE DATA

So let us assume that some combination of industry inspiration, government

support and academic research has yielded, in our imaginary Al year of 2030,
3.2.3: a global building industry data trust where firms are paid to contribute data,
PROJECT DATA
FLOWS AMD
THE DATA
TRUST
and in exchange have access to that data and the Als that have been trained
to use it. Let us further assume that there is enough useful data in the trust
that it
begins to enjoy widespread adoption and grows with each significant
project designed or built, worldwide. Designers, builders, their supply chain
and building owners all contribute data that results from their work.

This new relationship is described in Figure 3.2.3 , an elaboration of

Figure 1.4.2 .

The data sources created by a project team, including data from modelling,

analysis and other artefacts of architectural design, would be contributed to

the data trust. Assuming that its use has become widespread, we might expect
some standardisation of software, but I suspect the continued explosion

of tools that will result from the widespread digitisation of the industry will
make this challenging. It is more likely that Al itself may provide the means to
standardise and conform project data from architects and others before it is
contributed to the common cause.

Training an Al to recognise a representation of, say, a window in a BIM

model, purchase order or shop drawing (if such a thing exists in the future)
across those data sets is
question of pattern matching– something
a

machine learning systems do well. One imagines an Al-based 'transaction

layer' shown in the diagram, which would collect, translate and standardise
project data into consistent relationships and formats, and would be an
excellent opportunity for supervised learning, combined with BIM and other
evolving model typologies, for next-generation Al platforms. And if such a
capability could evolve, it would also improve the project-based common

data environments used by individual architects, creating integrated

representations of a project before that data streamed, at the appropriate
point, into the data trust itself.

An architect would therefore access and consume information, in this

construct, in three ways:

1. At the level of the individual project, as a result of the development of the

design.
2. From the data trust, for both reference and to deploy Al-assisted tools
(like, for example, the cost estimating and analysis tools described earlier).
3. From outside data sources that can inform the development of the
design, like economic models of the project context, weather data or

information about the availability of site utilities.

DESIGN CREDIBILITY

When structural engineers began to rely on software for routine calculations,

thecredibility of those results relied not so much on the regulation of or
promises by the technology vendors but rather on that the engineer herself
was responsible for the output of those systems and any errors that might
occur as a result of their use. Just as BIM has now become a tool that, under

the duty of care, an architect may be expected to use on a project,

Al-produced results will become part and parcel of the architect's
professional judgement.

However, at least in today's Al systems, the complexity of the data structures

that comprise neural networks are too great for humans to really understand,
and those systems are trained with enormous data sets and measured by the

validity of the outputs, not the

specific computations that produced them.
Kate Crawford, who writes on the challenges of Al implementation, describes
this challenge well: 'In the case of Al, there is no singular black box to open, no
secret to expose, but a multitude of interlaced systems of power. Complete
9
transparency, then, is an impossible goal.'

Opacity will make it impossible, in my view, for architects or others to rely

on these systems without some sort of third-party validation of their results.

Should the building industry, with architects as important contributors, decide

to build a global data trust to drive Al, a component of that trust would include
entities who would extensively test and certify the results of these systems
before releasing them into the wild. The future leaders of BuildingSMART have
a much bigger enterprise on their hands.

Beyond the proximate concerns of professional efficacy and certification,

much work is currently underway to understand and evaluate the social
and ethical dimensions of Al for decision-making. There are two important
dimensions of this work. First, Al systems are trained from data that is the
result of 'real world' inputs. Current natural language systems such as GPT-3
or facial recognition systems build their networks from scraping data, text
or photographs from the internet, and as such that data has, inherently,
the structural biases of its contributors. Princeton computer scientists who
research this idea call it 'veridical' bias, 10 and suggest not only that is it is
endemic in the world's data structures, but also a potential dashboard to
understand social bias itself. 11
A second dimension of Al only now beginning to be understood is its
implications on the environment. In the famous paper that resulted
now

in her being fired from Google, computer scientist Timnit Gebru argues
that the inherent environmental implications of building Al systems
are underappreciated and accrue to the detriment of underprivileged
communities who do not benefit from them. Calling naturallanguage Als like
GPT-3, upon which Google increasingly relies, 'stochastic parrots', Gebru and
her co-authors recommend that Al systems be designed to acknowledge the
enormous contribution to atmospheric carbon they contribute by virtue of
their intensive training, and rigorously curated to design out the inherent bias
of available training data. Given that, for example, much of the content in the
proposed data trust would likely be sourced from Western projects initially,
carefully tending this data to be free of its obvious neo-liberal dynamics will
12
be critically important. Crawford further argues that the damage to the
environment of extractive mining necessary to build the enormous compute
infrastructure of Al is an externality that should be reflected in its cost and
13
development.

THE FUTURE OF DATA

With the advent of artificial intelligence systems, the building industry

should be strongly motivated to share data beyond the solving of the near-term
problem of the inefficiency that results from incompatible formats.
Generating, consuming and properly curating digital design, construction
and building operations data will allow these systems to be properly
trained, and then unleash real power of computation for design. Doing so
is an enormous opportunity for architects and their industry collaborators
and comes with numerous pitfalls. However, perhaps the benefits of next
generation technologies will finally motivate them to address and solve
the larger questions of collaboration and integration, as well as ethical and
environmental responsibility, that creating such a data source demands.
AUTOMATED
TASKS,
3
3.

»AI
WILL EXTEND THE RANGE OF AUTOMATED
PROCESSES AVAILABLE TO ARCHITECTS AND
CREATE OPPORTUNITIES FOR OTHER ASPECTS
OF OUR WORK TO BE FULLY AUTONOMOUS,
OPERATING IN PARALLEL WITH HUMAN
COUNTERPARTS. SUCH CHANGES ARE LIKELY
TO FIRST OCCUR IN THE TECHNICAL
OBLIGATIONS OF ARCHITECTS, AND MORE
SPECIFICALLY IN THE RELATIONSHIP BETWEEN
DESIGN AND CONSTRUCTION, WHERE AI-ENABLED,
AUTONOMOUS ANALYSIS CAN IMPROVE
THE EFFECTIVENESS AND CREDIBILITY OF
DESIGN AND PAVE THE WAY TO OTHER AI
CAPABILITIES. «

DOI: 10.4324/9781003297192-16
 3.3.1:
DOXEL’S
CONSTRUCTION
SITE INSPECTION
ROBOT, THEIR
FIRST ITERATION
OF SCANNING/
AI-BASED FIELD
VERIFICATION 1

In 2018, BuildTech 2 start-up, Doxel, announced its first product, an AI-enabled

robot that, rolling around like a tiny tank that can traverse rough terrain and
even climb stairs, inspects a construction site using scanning and computer

vision. It compares the results to BIM-based design information to determine

construction progress on site. With the tagline'Artificial Intelligence for
Construction Productivity– Software that inspects quality and tracks progress
so you can react in minutes, not months', 3 this system purports to evaluate the

completeness, precision and installed value of work in place, automatically.

Meanwhile, in the
analogue world of humans, standard services contracts
for architectsstipulate the architect's responsibility for construction
administration (emphasis added):

»
Carry out visual site inspections, as stated in item F of the Contract
Details, to review the general progress and quality of the works
as they relate to the architectural design and issue site inspection
reports to the Client. (RIBA Standard Professional Services Contract 2020)
4

The Architect shall visit the site ... to become generally familiar with
progress and quality of the portion of the Work completed, and to
determine, in general, if the Work observed is being performed in
a manner indicating that the Work, when fully completed, will be

in accordance with the Contract Documents. However, the Architect

shall not be required to make exhaustive or continuous on-site inspects to
check the quality or quantity of the Work. (AIA B101-2017) «
5
 Machine Learning

While here in America we clearly have a penchant for more turgid contractual
definitions– a function of years of construction foibles and resulting litigation
– the overlap is clear. Doxel's deep learning-enabled system is designed to
either dramatically augment, or eliminate altogether, the need for human
inspection of construction progress. It is a perfect example of the potential
of AI in the design-to-construction process continuum: a Doxel robot uses
computer vision and machine learning from other projects and related BIM,
looking for something very specific ('is that column installed in the right
place?'), and creates analytical results more quickly, cheaply and accurately
than a person walking the construction site twice a day– an obligation
specifically excluded from the AIA's definition of construction observation.

MISPLACED ANXIETIES

Doxel's value proposition – understanding and managing construction faster

and more accurately fits several trends that AI in building industry is likely to
-

follow. Early investments in AI systems are:

» predominantly in the construction space (where there is more money

spent, and to be had)

» focused on well-defined problems with technical inputs and outputs,
and
»
operates in the intersection of originating digital design data and
construction execution (start with BIM, add computer vision, then
measure progress).
3.3.2:
DOXEL
ANALYSIS OF
CONSTRUCTION
COMPLETENESS6
 3.3 Tasks, Automated

The technical emphasis of this system, and many to follow, was anticipated
early in the last
technological transition to BIM by the architect Patrick
MacLeamy, who created what is now known as the 'MacLeamy Curve', as seen

inFigure 3.3.3 .

Originally an argument for the efficacy of BIM for architects, MacLeamy posited
that since the greatest value of the architect's work was early in the design
process – where important decisions have the best chance of positively affecting
results without disrupting progress
–

the bulk of our work process and value

should shift to the earlier phases in the design-to-construction schedule.
Further, he predicted that BIM would automate much of the production of
technical documents and other information needed by contractors, making
work in that portion of the scope of service far less valuable.

AIsystems like Doxel's, which autonomously perform technical tasks that once
required humans, is a logical extension of this same argument. However, while
construction progress evaluation will clearly benefit from additional, digitally
enabled help, a construction site is technically, geographically and, to some
extent, in a way that our little tank is unlikely to be able
politically complicated
to fully understand. So human
our architect, continuing to act on behalf of
the client to protect her interests during construction, will continue to have

3.3.3:
THE MACLEAMY
CURVE
a role even here (despite the likely reduction of
human-required production
tasks during the creation of working drawings, argued in Chapter 1.5 ). In
as

any case, since the technical aspects of design and construction– including
document production and coordination, technical evaluation like code
compliance or coordination, and managing and evaluating the information
flow from the construction process are the most likely candidates for AI-driven
-

task automation, I remain convinced that anxiety by architects about

being replaced as designers by autonomous AI is at best misplaced, and this
argument was anticipated by MacLeamy in 2003.

RECONSIDERING RELATIONSHIPS

InChapter 1.6 I defined

, the architect in terms of her relationships with four
key constituents:

1. As a protector of the public's health, safety and welfare.

2. As an agent expert to the client.
3. As a guide and translator of design intent to the builder.

4. As a manager/leader/integrator to the balance of the design team.

Each of these roles requires the innate human ability to understand context,
manage relationships and make trade-offs and judgements; these are
tasks that precisely the opposite capabilities of AIs, and particularly those
are

whose neural networks require training in large, well-curated data sets. The
multi-valent responsibilities of the architect, writ large as in this definition, are
therefore unlikely to be replaced wholesale by computers, hence the argument,
consistent with Susskind, about task replacement: limited tasks, perhaps, but

complete supplanting by autonomous computerised agents, certainly not; and

where tasks are AI-involved, they will be of a technical nature.

Of these four archetypal relationships which anchor the architect firmly

in project delivery models, such technical task automation will alter most
significantly the connection and obligations between the architect and the
contractor/building supply chain. That relationship is the most transactional
of the lot, characterised largely by exchanges of information, with the levels
of precision and completeness often contested. AI systems will characterise,
catalogue and eventually measure the quality of these interactions and
connections in an effort to make those exchanges more effective. As the
construction process is further automated by AI-assisted devices and systems,
the demand for specific design information that is suited as input to those
systems will increase, further tying architectural design to construction

production in the name of efficiency. And as the demands for more intelligent
supply chain decisions and management increase in an era of climate change,
reduction of toxicity or even attempts to reduce forced labour, the architect's
specification of materials will tie even more closely with an understanding of
the supply itself. Contractors will therefore come to rely on architects for such
7
intelligent decisions. This is an inversion of the relationship of technology to
design that emerged with the blob-makers of the 1990s, when the architect's
digital shape-making tools made possible ever more elaborate forms that
were left to engineers and builders to actualise.

The UK tendency to novate the architect's contract to the contractor after

design, whilesimultaneously confusing and terrifying to this American
architect, is consistent with the conclusion that design and construction can
be operationally bound more tightly in contract, procedure and technology. As
construction processes are more reliant on AI-based automation, the demand
forlogical relationships between design information and decision-making will
naturally gravitate toward the design-construction interface.

PROCEDURAL PRODUCTION AND ITS POSSIBILITIES

So, an extension of this line of reasoning that includes MacLeamy's thesis,

combined with the task analysis of the jobs of architects (see Figure 1.5.3 for
the complete version, and Figure 3.3.4 for production work), suggests that the
most likely opportunities for AI-based augmentation or automation of tasks
is deep in the production phases of design, when a project is translated from

design intent to the information readied for the contractor.

The automation of selected tasks of working drawing production is the best

example to date of the implications of BIM in design. Working drawings

today are more precise, better coordinated, more accurate and useful to
contractors than their CAD predecessors, and BIM is a good platform, as both
a representational schema and a training ground, for AI to continue this trend.

As many of the tasks of the working drawing phase are procedural in nature,
it should be relatively straightforward to train AIs to perform them. Other
objectives of the hand-off between designer and builder will still demand
the judgement and coordination of the architect, who will be supported by
machine intelligence in areas where large data sets, pattern matching and
complex calculations and predictive algorithms could be of most use.
3.3.4: TASK
ANALYSIS
OF THE
TECHNICAL/
PRODUCTION
STAGES
OF THE
ARCHITECT’S
WORK
To that end, consider the problem of construction cost projection, and
particularly the heated negotiations oft held between designers and builders
about conformance to target costs. The dimensions of this complex dance of
data, judgement and computation are sketched in Figure 3.3.5 .

Just as BIM was anticipated to accelerate aspects of cost estimating by

automating the tasks of quantity take-off and calculation, the various
8

analytical tasks that converge to generate cost estimates, based on careful

analysis combined with professional judgement, can generate greater insight
and precision in cost projections – if AI were available in the following ways:

3.3.5:
ANALYTICAL
ELEMENTS OF
CONSTRUCTION
COST
Quantity surveying: Design documents define design intent, and builders
must interpolate precise system characteristics, unspecified materials and

implied construction elements in order to price them. By training on data sets

that map construction documents to detailed previous take-offs, an AI could
create these mappings from working drawings and create more accurate and

historically informed bills of materials. The resulting data collections would be

the basis of further training of the systems.

Market conditions: Macro-economic conditions introduce extreme volatility

in construction pricing. Global or local circumstances change the availability
of labour or the cost of materials more quickly than capital plans for projects
can respond. This volatility is demonstrated by the difference between
material purchase prices from manufacturers versus their cost as stockpiled
in warehouses by builders in the US, as can be seen in Figure 3.3.6 At the time.

of writing, material prices have jumped as much as 40% (timber) since January
2021, as the US economy recovers from the pandemic; a similar pattern was
seen in the recovery from the 2008 crisis. AI-based analysis and predictions,

based on previous economic models and pricing profiles, could elucidate the
potential implications for pricing, bidding and market conditions that affect
cost models.

Cost of labour and materials: The costs of both labour and materials are
estimated traditionally from historical information, but the actual pricing of
projects generated by builders is a combination of historical data, in particular
competitive advantages or disadvantages based on the builder's capabilities,
as well as larger market forces such as the availability of skilled workers and

other competitive pressures. A given builder or trade could train an AI from a

combination of past projects to create a more reliable and accurate projection
of such costs and map thoseprojected costs to the designer's digital models to
provide continuous cost modelling during the course of a project.

Supply chain conditions: The availability, price, performance and

suitability of building materials is subject to many conditions in the
delivery supply chain, from original sourcing to fabrication to delivery.
Carbon implications of both embedded carbon and transportation costs,
toxicity and labour equity (modern slavery) all affect the possibilities that
a given material can be supplied to a project properly. AI systems could

be trained on industrial material flow models, shipping manifests, bills

of materials for sub-systems, even labour assessments, to factor these
questions into cost projections of built assets.
 Construction schedule: As most builders model the construction schedule
for projects in order to plan and manage the build, such schedules could be
training set for AI systems that would establish patterns and relationships
between building typologies, construction locations, specific combinations of
subcontractors and fabricators, and construction cost. The insights provided to
builders could be provided, a priori, to design teams to understand whether a
particular design was difficult or more expense to construct. Such information
would be particularly useful in the concept stage, but it would allow a valid
performance evaluation during technical production as well.

Build strategy: As contractors use increasingly sophisticated systems to model

the sequence and methods of construction, these '5D' data sets instantiate
libraries of construction logic at a macro-scale. With sufficiently large training
sets they could be the basis for AI-generated construction planning. If the
resulting projective modellers, which would become familiar with a wide variety
of construction approaches and building types, could function as 'constructability
evaluators' via AI-enabled simulators and were available to architects, the
Albertian gap between intent and execution could be further closed.

Theforegoing example, a speculation through the lens of a critical yet

suboptimised dependency between design and construction, is meant to
reinforce the idea that AI systems will appear first, and be most useful and
efficacious,in the translation of design into technical performance, and

particularly with regard to construction. Given the well-trod conclusion

that the construction enterprise is unproductive and unpredictable, such
improvements would be both welcome and embraced by architects, builders
and their clients alike. 9

CONSTRUCTING AUTOMATION

The painting robot we examined in Chapter 1.4 signals important changes for
designers, not the least of which is the likelihood that their projects will be
festooned with more precisely applied colours and textures. As Negroponte
suggested in his early explorations of the automation of processes by
machines, a first step towards incorporating digital technology is to use
it to replicate an analogue process, and surely this is where our painting
robot will begin. However, Negroponte further speculated that once these
processes are fully integrated and understood, their capabilities will expand
10
far beyond what their originators could have anticipated. Our painting robot
will improve its technique by repeatedly painting surfaces and sharing the
lessons of its success and failures with other robots doing the same on other
3.3.6: projects. Further insights will be supplied by data coming from the construction
MATERIAL
supervision robots made by Doxel and their inevitable competitors. Eventually,
PRICE
the robotic painter will have AI-automated, autonomous colleagues installing
VOLATILITY
IN THE US, and assembling other aspects of the construction project, and the procedures
11
2006-21 and protocols they generate could combine into an accessible, evolving source
of construction insight that could truly modernise building.

For architects, this is a profound implication. It is not the automation of their

tasks, so much as those of builders, that could close the divide between design
intent and construction execution, a divide that Alberti defined six centuries ago
and that has since characterised, or plagued, our industry. The instantiation of
construction logic makes it available to the robots that enact it on site, but also
to the designers who are configuring the eventual results of that robotic work.
Where today an architect projects the eventual state of her design through a
building information model, that design could be informed by a BIM connected
to a construction simulator that might answer questions from as small as 'does
that fit' to as large as 'can that be built?' In that sense, the task automation
potential of AI enlarges, rather than diminishes, the potential of design.
3.3.7:
FOSTER +

PARTNERS
USING A BOSTON
DYNAMICS
ROBOT FOR
CONSTRUCTION
PROGRESS
ANALYSIS 11
 DESIGN
OF
LABOUR
4
3.

>> THE LABOUR OF KNOWLEDGE WORKERS

LIKE ARCHITECTS, AS WITH MANY OF THEIR
PROFESSIONAL COUNTERPARTS, WILL LOOK
DIFFERENT WHEN MACHINES CAN PERFORM
ARCHITECTURAL TASKS. THE TYPES OF LABOUR
AND LABOURERS, THE STANDARDS UNDER
WHICH THEY OPERATE AND THEIR
INTERACTIONS WITH PRACTICES AND PROJECTS
WILL DEVELOP AS DIGITAL PLATFORMS
EVOLVE FROM INSTRUMENTS TO COGNITIVE
COLLABORATORS. <<

DOI: 10.4324/9781003297192-17
My first full-time job in an architect's office was in 1979, the pre-CAD era, when
our work wastaped large
to drafting tables and prepared with plastic lead on

giant sheets of mylar. The small office in Charlotte, North Carolina was known
as a solid, if stolid, practice that did complete working drawings that resulted

in routine buildings which did not leak or miss theirbudgets. I was assigned to
help prepare construction documents for a bland shopping mall in Tennessee
that still haunts my dreams.

There were about 20 or so of the

drafting room which I hesitate to call a
us in -

studio, as there was so little design to speak of going on at least two or three
-

of whom were architectural draughtsmen (no women). These were older guys,
some with architectural degrees, none with licenses, who drew and lettered

beautifully, knew a lot of about how to put a building together, could never be in
front of a client and were constant sources of knowledge to the younger, better
educated but far less experienced architects-in-waiting like myself. In that era,
many offices of any size had folk like this, whose main job was to draw, leaving
all other architectural responsibilities to others. Late in my 15-month stint in this
firm, the office manager began researching a new idea called 'computer-aided
drafting', which the draughtsmen dismissed entirely as a gimmick.

TECHNOLOGY AND DESIGN LABOUR FROM DRAFTING

TABLE TO BIM

Two decades later, these sorts of wise but unregistered drafters were largely
missing from practices, unable to make the transition to CAD. Their jobs were
replaced by young, digitally enabled CAD operators who were unafraid of the
computer and had the skills to use it to draw. Almost no one in an architecture
firm was trained or hired at this point as a drafter; young designers in training
filled these roles. Just as today's secretaries no longer do much typing but are
more general support staff, architectural jobs were no longer differentiated by

3.4.1:
THE DRAFTING
ROOM AT
SKIDMORE,
OWINGS &
MERRILL’S
OFFICE,
CHICAGO,
1958
Machine Learning

production tasks (drafting) but more around roles and responsibilities (design,
construction administration, specifications).

As firms are completely reliant on technology of all sorts these days, there
are specialists in network management or even coordination of BIM data, but
such roles are limited and certainly do not contribute to billable work. And even
if yesterday's 'CAD monkeys' are today's 'BIM monkeys', BIM work is not routine
drafting, as using a BIM tool requires a strong understanding of how a building
goes together and how to properly represent it. But even today, while larger
firms are hiring specialists in data management or, in some cases, software
development, practice is largely devoid of technological specialists.

The advent of machine learning tools will reverse this trend. It will create
demand for different sorts of architectural workers. As I have argued (see
Chapter 2.4 ) that artificial
general intelligence (AGI) is far in the future -

particularly AGI that can step into the multi-faceted role of a proper architect
-

Al systems in architecture will be specific to tasks, technical in nature and will

support the broader enterprise of design. Those systems will require specialised
understanding of inputs, outputs, data demands and relationships of the Al
system to the broader infrastructure of design information. These are skills that
architects trained as generalists are unlikely to understand, nor, frankly, have
much engagement with: the outputs of such systems will be of great interest;
the process by which they are generated, not so much. While it would be nice
to simply ask the 2030 version of Alexa, 'How much carbon is embodied in my

project?' the route to that answer is likely significantly more complex.

TECHNOLOGY, SUPPLY AND DEMAND

I have argued up to this point that Al systems are likely to automate the
more routine aspects of technical drawing production, data and document

management, and information control. Just as today's modelling platforms

3D
that generatehigh-resolution renderings have
largely put professional
(analogue and) digital renderers out of business, those jobs in offices will
be lost.

However, there will be new jobs. Architecture will need experts who can
manage these systems in production, particularly in relation to affiliated Albased
processes that will translate knowledge and insight from construction
back to design. Complex technical analysis and building performance
evaluation will be a necessary element of design generation. The dilemma
 3.4 Labour of Design

for practices will be mapping demand for the specialists to handle such work
and the available supply of talent. If an Al-based system is helping optimise
the carbon footprint or forced-labour-free supply chain procurement of your
building in the phase of conceptual design, that work is episodic at best, and
probably punctuated with interactions with other intelligent systems. The data
trust proposed in Chapter 3.2 will provide ubiquitous data, but using it well
will be challenge.

In a recent survey of architectural and engineering professionals in Europe

and the Middle East, 70% of firms reported that big data, data science and
machine learning were important emerging technologies. 1 Far fewer are
actually doing anything about it. Those firms trying to realise actual work from
Al are competing with limited success, I suspect with the global demand
- -

for data scientists and machine learning experts, many of whom earn starting
salaries well above those of their staff architects in UK practices. 2 While
expertise in Al will grow in the next decade, matching supply and demand will
be a challenge.

As Daniel Susskind argues in A World Without Work: Technology, Automation and

How We Should Respond (the successor to The Future
of the Professions: How
Technology Will Transform the Work Of Human Experts), machine intelligence
will, over time, take over responsibility for tasks rather than complete jobs.
However, the aggregate task elimination in architecture will eliminate jobs,
if for the simple reason that production-related activities such as creating
working drawings or managing construction administration data comprise, at
least in the USA, as much as 35% of a typical fee for an architect's services. 3

Automating this work, at least under current business models, means fewer
workers. The so-called 'canonical model' described by Susskind -

where jobs
destroyed by innovation are replaced by the jobs required to create the
new technologies has given way to a new thesis, the so-called 'Autor-Levy-Murnane
-

(ALM) Hypothesis', which declares that the routine tasks of work will
be eliminated by computers, as those tasks are the easiest to teach machine
learning systems to replicate. Skilled workers, like my early drafting colleagues,
4
are eventually replaced, resulting in fewer jobs that do not reappear.

Or at least fewer jobs will reappear than disappear, given the current demand
for architects. However, there is another scenario, where Al technology
empowers architects to the extent that demand for professionals even those
-

doing different jobs, like their counterparts in the early days of CAD will be -

much higher. The history of BIM in the UK suggests that this might be the
case, as demonstrated in Figure 3.4.2 .
3.4.2: Despite the inherent efficiencies introduced by BIM and its adoption driven by
DEMAND FOR
the Level 2 requirements, which in theory should have decreased architectural
UK ARCHITECTS
AND RELATED positions, and even despite a drop in designed construction value in 2020,
CONSTRUCTION the number of architectural positions in the UK has steadily increased for a
VOLUME decade. Perhaps better work begets more employment.
DESIGNED 5

TALENT, SUPPLY AND DEMAND

Practice in the era of Al will therefore entailarbitraging the value of expert labour,
machine production and data. The profession will need to find a way to access
this talent across many scales of business, particularly since most firms, world-wide,
are relatively small, averaging less than ten staff each. While larger firms,
6

with more human and financial resources, may get to Al capabilities first, for real
change to be possible the wider profession must be able to access Al assets.

An intelligent design of the industry data trust described in Chapter 3.2 would
entail a central marketplace for such talent and the infrastructure to make
it accessible. The technologies of the so-called gig economy, which match
demand with capacity, are the template for such a system. Like ride- and
apartment-sharing services today, these systems use artificial intelligence
themselves to align unused resources with those who might need them. Given
that Al enactment in design practice will be task-oriented, perhaps the data
trust might manage such a platform as part of the value proposition it puts
forth to the global design industry.
 3.4.3:
BUSINESS
MODEL OF
AIRBNB,
MATCHING
HOUSING
CAPACITY WITH
TRAVELLER
DEMAND

More than a year of digitally enabled

knowledge work by architects during the
Covid-19 pandemic demonstrated that design work can be conducted with a
relatively low loss of efficiency, with staff working remotely and distributed
around the world. Al expertise could be similarly dispensed, making it more
likely that this talent would be used fairly and effectively.

ETHICAL LABOUR AND AI

The architecture does not enjoy a sterling reputation for labour

profession
equity, and the advent of machine intelligence should not be seen as a
strategy to extend that poor record. Salaries relative to other learned
professionals are low, 7 steady jobs as
uncertain as the economy and even
the most well-known offices have well-earned reputations for labour abuse. 8
In the US, most architectural workers are exempt from labour laws that

require employers to limit hours and provide minimum pay for overtime,
characterising them as 'professional workers'. Doing rote CAD work, not so

much; wrangling an Al, absolutely true.

Like any disruptive technology, Al offers opportunities and threats to the

knowledge work of architects. As has been argued by my Yale colleague,
Peggy Deamer, many architectural workers, particularly those who are
unlicensed or less experienced, can be precarious workers whose work lives
are destabilised by long hours, low pay, competitions and the generally poor

management of human resources that many architects learn in the design

studios of school (where time is considered an unlimited resource). 9 Al, like
many of the automation technologies of the past, can be deployed in the
interest of ruthless efficiency, with little consideration for the welfare of
workers; ask anyone working today in an Amazon warehouse, Systems that 10

match workers with jobs have been demonstrated to instantiate the biases of
the data set (like résumés of employed workers), denying qualified applicants
opportunities generated by machine-based hiring systems. Workers in the
extractive industries that drive modern computation are often exploited and

unlikely to see any benefit from the computers they supply.

In her ill-fated paper for Google, Timnit Gebru makes the case that Al
development must occur in parallel with an understanding of its broad
implications. She suggests:

>> Work synthetic human behavior is a bright line in ethical Al

on

development, where downstream effects need to be understood and

modeled in order to block foreseeable harm to society and different
social groups. Thus what is also needed is scholarship on the benefits,
harms, and risks of mimicking humans and thoughtful design of target
tasks grounded in use cases sufficiently concrete to allow collaborative
design with affected communities. 11 <<

Modelling downstream effects is what an Al-enabled future of design might

look like. In architecture's case, the affected communities including both -

our clients and our workers, and the profession working in concert with
-

providers and the academy, would best heed her advice and begin plotting
the route to equitable Al today.

The ALM hypothesis (as discussed in Chapter 1.5 ) proposes that highly skilled
workers -

like the ones that will create, develop and manage Al will remain in -

demand once machines carry out knowledge work. Very low-skilled but highly
localised jobs, like those in personal services or dining, will also remain. In
Susskind's 'massacre of the Dilberts', many jobs in between will be eliminated.
Most architectural jobs in this scenario are probably safe, but the society for
which we design buildings is likely to be dramatically affected. Like other issues
of social equity, it is best to add this question to the list that the profession must
address to responsibly design the built environment of the future.
PROPOSITIONS
BUSINESS
VALUE
5
3.
AND
MODELS

» TECHNOLOGICAL CHANGE IN ARCHITECTURE

HAS CHANGED THE PROCESSES, BUT NOT THE
VALUE, OF THE ARCHITECT’S SERVICES,
WHILE THE BUSINESS MODELS OF PRACTICE,
WHICH ORIGINATED CENTURIES AGO,
HAVE REMAINED STUBBORNLY IN PLACE.
AS THIS NEXT WAVE OF NEW TOOLS AND
CAPABILITIES WROUGHT BY INTELLIGENT
MACHINE AUTOMATION BECOMES APPARENT, THE
PROFESSION HAS ITS BEST CHANCE TO BREAK
THIS CYCLE BY REVISITING AND REVISING
ITS VALUE RELATIONSHIP TO THE BROADER
BUILDING INDUSTRY. «

DOI: 10.4324/9781003297192-18
Machine Learning

In his superb history of American architectural practice, Assembling the

Architect: The History and Theory of Professional Practice, George Barnett
Johnston explores the psyche of early 20th-century American architects
through the eyes of one Tom Thumtack, the fictional alter ego of architect
Frederick Squires, who wrote an illustrated volume about practice called
Architec-tonics: The Tales of Tom Thumtack, Architect. Tom expounds on the
anxieties of fees, as quoted by Johnson:

» We're paid on a percentage of the cost, but the capable architect
is the one who keeps down the cost. Therefore, by doing his best he
reduces his compensation... The client wants to keep the cost down, and
his architect must help him in this, but the less the cost of a particular
job, the less the compensation and the less likely to be the beauty of its
1
execution from which the architect obtains his reputation. «

Tom was explaining the seeming illogic of a system of compensation for

architects that had, within it, two deep contradictions. First, when the
architect's fee is based on a percentage of construction cost, the harder the
architect works to bring the project into cost conformance, the less she is
paid. Second, in the cases so common today when said fee is converted into
a lump sum, the client has transferred the financial risk of the fee over to the
architect, who perversely is now incentivised to work less, rather than more, to
service that client, and thereby preserve some remainder of the fee as profit.

The idea that an architect should be paid in some proportion to the cost of
construction seems to have originated centuries ago in Europe. Johnston
explores this trajectory and quotes Benjamin Latrobe, the so-called 'first
architect' of the United States, European-trained, British immigrant to America:

England the established custom of

» It is in France, Germany &
Architect (in England, confirmed by many decisions of the Courts) to

charge for their works, 1., a commission of 5 prCent on the whole

amount of the expence incurred in executing their design, -2., a certain
sum for fair drawings, if furnished, according to their difficulty,

number, or beauty; -3., if the work be at a distance from the usual

residence of the Architect,-all traveling expences, & a certain sum pr
day for loss of time. «
2

The origins of today'spercentage-of-construction-costs fees, adjusted by

project complexity, plus reimbursable expenses of travel, are evident here and
remain in place on both sides of the Atlantic today.3
 3.5 Value Propositions and Business Models

COMMODITY RESISTANCE

It is hard to imagine another modern enterprise, even one so reluctant to

really modernise like architecture, whose business models are essentially
unchanged from their 18th-century precedents. Yet architecture, like much
of the construction industry, remains tied to a fundamental value strategy of
lowest first cost, where services are bid and purchased in a way not dissimilar
to steel, sheetrock or carpeting: maximum pressure on competitive price, with
far less attention paid to the value delivered, particularly over the cycle of a
project's lifespan.

Neither of two immediate implications of AI for practice are particularly

our seemingly intractable business model.
sanguine for architects, given
More productivity may provide short-term gains to the early adaptors, but
such competitive advantages are short lived when eventually available to all
competitors in the market and quickly fade. Should AIs replace jobs in the
4

architect's office, commoditised fees will fall in proportion, or worse. 5 Given

this inevitable economic logic, finding new value propositions and business
models will not just be nice benefits that accrue from these new technologies,
but rather an existential necessity for the profession itself. If neither CAD nor

BIM inspired such change, will AI?

3.5.1:
KENT
ROCKWELL’S
'FEES: A
REDUCTIO
ABSURDUM’ FROM
ARCHITECTURE
AND BUILDING,
46, 1914 6
There is no question that those technologies improved the processes and,
in some ways, the results of architectural services. CAD made drafting more
accurate and efficient, while allowing architects to depict projects that were

technically and formally more complex. BIM has allowed all members of
the delivery team to generate, organise, integrate and exchange design
information at much higher levels of resolution and transparency. It also,
in some minimal way, begin to bridge the information gulf between design
and construction; builders who saw the value of 3D data began to request
it to assist their work. Other digital technologies have improved information

exchange and client-facing images of projects (think renderings or even virtual

reality models). Yet despite these improvements, the centuries-old methods
for computing architectural computation remains largely intact, suggesting
that these improved services have not translated into business terms, nor
profit. The MacLeamy Curve, as described in Chapter 3.3 suggests that
,

the real value of design work lies early in the delivery process, despite the
relatively small degree of effort entailed there compared to production and
delivery stages. Perhaps AI will begin the value shift.

A willingness to examine innovative business strategies for new services,

organisational strategies and products can translate the threat of
even new

AI into an opportunity to improve both our performance as professionals

and our business results, if we apply the same sort of creative thinking oft
reserved for the design studio to this problem.

EXPLORING NEW VALUE

For the last several years we have offered a course at Yale called 'Exploring
New Value Propositions of Design Practice', where our students are asked
to interrogate the business models of architecture with the assumption that
better jobs can be done
designing them. Each semester, teams of students
create what they believe to be the most provocative new models for practice,
in response to the question: 'Where can the value of architectural services be
best translated into a business model?' The projects are required to conform
tojust two conditions: the proposal must be based on something that a
competent architect is capable of providing, and the compensation strategy
must completely abandon any vestige of commoditised pricing, so no fixed or

hourly fees allowed to make money.

By definition, this second constraint eliminates any option to extrapolate
traditional practice, which is something of a pedagogical conceit intended to
push the students far from the comfort zone with which they are familiar by
dint of theirprofessional experience and the businesses largely run by their
professors. imply, however, that there are no approaches for
It does not

transfiguring traditional practice models towards new technologically inspired

value, an idea we will examine later in this chapter.

After few years of teaching this class, several consistent strategic themes
a

emerged from the students' research, which range across the opportunities of
services, organisation and products. These ideas were congruent in that they
7

could be mapped in relation to the connections between architects, clients

and builders, as described graphically in Figure 3.5.2 and included:
,

» verticalisation strategies, where the architect assumes her role plus at

least one other traditional delivery role
»
supporting strategies, where the architect uses special skills, talents
(and often technology) to provide a service to some part of the supply
chain (including other architects)
»
spanning strategies, where a business derives value by creating an
important connection between two sectors of the delivery model.

Should AI diminish the demand for architects, these approaches are a

sketchy roadmap for other opportunities. However, they also suggest that
thedisruptive power of autonomous computing through intelligent machines
might create new leverage and power for architects in the overall process of
making buildings. For example:

A robust, AI-enabled design schema aimed at leveraging prefabrication and

industrialised construction, based on procedural knowledge of manufacturing
and assembly draws architects close enough to the means and methods of
constructions so that they could
perform sub-contracting or even construction
management duties. There have been attempts to leverage BIM in similar
ways as current precedent, tightly binding design strategy with construction,
like SHoP Architects' early provision of digital fabrication data directly to the
exterior enclosure fabricator in their Porter House project in New York.
3.5.2
DELIVERY
RELATIONSHIPS
AND
STRATEGIES OF
ALTERNATIVE
ARCHITECTURAL
PRACTICE
(FROM YALE
COURSE 2230B/
EXPLORING
NEW VALUE
PROPOSITIONS
OF DESIGN
PRACTICE) AND
A FEW EXAMPLE
BUSINESSES
SUGGESTED BY
THE STUDENTS
CLASSIFIED BY
THIS TYPOLOGY
A firm making a significant investment in an AI-based decision-making system 3.5.3:
SHOP
that supports a specific technical design objective recoups the investment and
ARCHITECTS,
makes a profit by offering the resulting expertise to other firms with similar PORTER HOUSE

challenges, or to clients as a validation service for other designs. An architect with ADDITION,
ELEVATION
deep expertise in a given discipline, say healthcare design, may have made
a significant investment to develop AI-based analysis systems to evaluate or

generate solutions in that building type. A healthcare practice, for example,

might have transferred its deep knowledge of operating theatre layout into
an AI that has been trained (with good and bad examples) from the global

data trust. While their competitive advantage is in deploying the analytical

results into the larger context of an overall solution for a hospital, the AI
platform which is very 'knowledgeable' about operating theatres could
- -

be made available to other firms doing work (perhaps in another geographic

location where competition is not an issue) or as a service to current hospital
clients to evaluate their existing facilities. This could be a profitable business
in and of itself, as well as a valid business development strategy to create new

opportunities. As mentioned in Chapter 2.4 the current precedent today is

,

Philadelphia architects Kieran Timberlake, who shares its expertise in energy

assessment through its Tally© carbon assessment tool.

An architect with a long-term relationship with an institutional client like a

-

university or retail operator has designed numerous buildings for them. Mapping
-

design model data with data streams from building control operations systems,
3.5.4: she has collected a large enough data base of this client's building base that,
TALLY©
combined with data from the trust, has allowed the firm to create an AI-based
CARBON
ASSESSMENT optimisation tool that can apply to future designs as well as tuning the operation
TOOL of current assets. The
resulting contract, extending in blocks of five years
post-occupancy for every building in the portfolio, requires the architect to
manage and evaluate the data streams from projects, evaluate operational
optimisation and make recommendations to the client. In addition to an
annual service contract fee, the architect is also paid a small percentage of the
operational savings in energy, maintenance and staffing resulting from these
services. Today, architects
EskewDumezRipple reserve 2% of fees to provide
post-occupancy clients, discovering strategies for improving future
services to

projects and building credibility for new projects as well.

8

TRANSFORMING BASIC SERVICES

These short vignettes are meant to stimulate thought about what alternative
value propositions for architects, underpinned by the new capabilities of AI, might
look like in the future. More important, however, are considerations of how, if at
all, the basic services of architects at the core of our value directly designing the
-
3.5.5:
UNCERTAINTY
built environment -

may evolve when we share the job with intelligent machines. FACTORS IN
It is naïve to believe that architects can immunise themselves completely from
PROJECTS,
the pressures of productivity improvements and knowledge work replacement ACCORDING TO
A SURVEY OF
that these systems will inevitably bring, and further that a profession that has
US DELIVERY
operated with essentially the same business model since before the invention of PARTICIPANTS 9
electricity can, within a generation, turn to fundamentally new busines strategies.
We are stuck with what we have, but can we fix it?

Facing these realities, which include continued competitive pressure to deliver

our work as a commodity, there are two approaches that can reform, rather
than completely replace, current fee-for-service approaches: reliability and
results. Project uncertainty the concern that project process will result in
-

unexpected and unfortunate results is a significant concern for architects,

-

clients and builders alike.

A 2014 study in quantified these worries, identifying the top

the United States
causes for project
instability as perceived by these constituents. While there is
general agreement about the list of challenges, there is significant disagreement
about their relative importance, as can be seen from the data in Figure 3.5.5 .
 Of this list, two of the factors are the direct responsibility of the architect
(design errors, omissions) and several more include the significant
involvement of the architect in some way (owner changes, accelerated
schedule and construction coordination). The resulting projects fail to
perform, in someway, as seen in Figure 3.5.6 .

The anxieties in Figure 3.5.5 portend failures of results in Figure 3.5.6 .

Emerging AI strategies in construction give us a clue where this might be

going. My former colleagues at Autodesk are developing a system that applies
machine learning to construction administration data to identify potential
problems on a job site either in production or with specific contractors
- -

in
advance. A dashboard is shown in Figure 3.5.7 .

Another provocative start-up, SmartVid.io, makes a tool that uses a combination

of machine learning and computer vision to scan activity on a construction site
and identity potential safety violations, as seen in Figure 3.5.8 .

Note that neither of these companies is using AI to replace the work of human
construction coordinators, risk managers or safety leaders, but rather augment
their capabilities and allow them to significantly improve their performance.
There is no reason why such an approach, applied to architectural design
3.5.6: process, could not go right to the heart of client and contractor uncertainty
WHEN in design process, be it checking for properly coordinated construction
PROJECTS
documents, cost prediction, lifecycle modelling of materials for durability/price
MEET
EXPECTATIONS, trade-offs, and even missing information from the documents that comprise the
OR NOT 10 contract for construction.
 3.5.7:
PROBLEM
DASHBOARD
FROM BIM
360 IQ

Even generative design, where algorithms generate alternative design solutions,

is not likely to be really useful until the resulting schemes can be sifted by
intelligent evaluative systems driven by AI. Those systems can measure,
mathematically, the performance of the resulting generated schemes, but
choosing, refining and implementing those decisions will remain far beyond
the reach of their capabilities, and human architects will always make the final
determinations of what is best.

NUMERACY, CREDIBILITY VALUE

Z. Smith, the Director of Sustainability and Performance at EskewDumasRipple,

believes that key
one to success is for architects to reach what he calls 'numeracy':
'Illiteracy is about
language, innumeracy is about numbers. We don't like
numbers. It's not what people thought they were getting into when they went
into architecture. But you have to do it if you want to make a good building.' 11

Architects largely measure the success of projects by an intangible sense of 'good'

and 'bad' design, and size up the competition accordingly. As the respective chief
technologists of three of the world's largest firms said to me once during an early
BIM conversation: 'We don't compete with each other with technology. We beat
each other on design.'

The definition of good design must move, at least in part, into the numerative,

performance-based aspects of cost, quality and schedule not to mention-

environmental and social impact and the proper design and implementation
-

of AI systems is the key to this change. Numerate and talented designers and
the results they can create will have the dual benefits of increasing the credibility
of architects and firmly anchoring our value in the building supply chain. And
perhaps building an AI-assisted numeracy in setting performance objectives and -

getting paid for them will achieve the alignment of value and business models
-

that Latrobe bemoaned almost three centuries ago.

3.5.8:
SMARTVID.
10 SAFETY
ANALYSIS
CONCLUSION

>> THE PHILOSOPHER ARTHUR I. MILLER

DEFINES CREATIVITY AS ‘THE PRODUCTION
OF NEW KNOWLEDGE FROM ALREADY EXISTING
KNOWLEDGE AND ... ACCOMPLISHED BY PROBLEM
1
SOLVING’. IN THAT SINGLE SENTENCE, HE
CAPTURES THE ESSENCE OF THE CHALLENGE FOR
ARCHITECTS AS ARTIFICIAL INTELLIGENCE
TECHNOLOGIES MATURE. <<

DOI: 10.4324/9781003297192-19
Our profession is, at its core, a creative enterprise that is valued for our ability
to both define and solve problems in unique, appropriate and beautiful ways.
Using Al should extend, rather than exterminate, that obligation to our clients
and the public at large.

While it may take some time until it reaches the far corners of the design
and construction professions, eventually the pattern analysis, autonomous
processing and data evaluation capabilities of AI/ML will appear in the
architectural landscape. We can declare, a priori, that the technology
represents an existential threat to architects, or we can use our problem-solving
skills and design a strategy that determines their ultimate destiny and
use. And in doing so, we can increase the influence and credibility of architects
with clients, mend decades of broken relationships with our construction
collaborators and maybe even break the chain of our commoditised value
propositions and stunted fee structures.

Any strategy for guiding the development and use of Al systems in

architecture should serve two goals, to improve the quality of the built
environment and to enhance the relevance of the human architects who
are best suited to make those improvements through design. Given that
the development of increasingly capable modes of automation are inevitable, I
propose that the profession embrace five strategies to guide its future:

1. Explicitly guide the definition and creation of technologies that will frame
future practice. Given that the next generation of technology may well define
the future of architectural practice, the profession must establish means to
declare its needs and direction in a way that does not defer to the business

whims of software providers, whose motivations will ultimately prioritise

profits and shareholders. 2 Architects have spoken with individual voices as

customers rather than in a united fashion as a collective of important users.

They should organise, collaborating with regulators, clients, designers and
builders, to declare an industry technology strategy that prioritises the most
important data and AI/ML capabilities and then demand the industry provide
them. Contrary to common wisdom, software providers actually prefer such
an approach, which saves them the time and effort of extensive and usually

incomplete requirements research. The national BIM initiative in the UK,

begun in 2010 and now a template for global technology adoption, is an
excellent example of how this can be done at scale. 3
Maschinelles Lernen

2. Expand the remit of design to include explicit performance. Building

performance is increasingly a necessary component of competent
design. Starting with life safety at the beginning of the 20th century, and
energy performance at the beginning of the 21st century, the range of
performance parameters that architects must address will continue to
expand. Al systems, driven by data, can empower architects to integrate a
broader set of these parameters into their design processes, connecting the
generation of solutions to performance models of, for example, occupancy,
economics, epidemiological implications, embedded carbon and even
embodied labour. These considerations do not supplant the importance
4

of design in its traditional sense, but rather expand it, while simultaneously
enlarging the effectiveness of building and the credibility of architects.

3. Create the data infrastructure that can serve as platforms for design.
Today's designers have created tens of thousands of digital models,
mostly through BIM, of projects that represent an enormous resource
for Al-generated insight. Contractors are doing the same with drone
scans, computer-vision analysis of construction and digital construction
management tools. Building control systems are generating huge lakes
of digital information about systems performance. The potential of these
resources is wholly unrealised without a strategy to organise and access

them, particularly in the industry's contentious and risk-averse delivery

models. The collaborative organisations described above could guide
software strategy and create policy and platforms for the collection,
organisation, access and use of this data, ostensibly through a global
building data trust managed by a third-party fiduciary and accessible to all.

4. Change the relationship between

design, construction and asset
operation. The emergence of BIM as a broadly understood concept
around 2004 coincided with the development of new models of integrated
project delivery, based on the assumption that readily available,
transparent of project information would accelerate cooperation
sources

between owners, designers and builders, the lack of which is a well-known

5
pathology of the building industry. More than 15 years later, that
promise remains largely unfulfilled. Digital data created by the various
players may be transferable, but it is often incompatible and rarely
more

shared. Al platforms, which could develop, manage and integrate the

6

data relationships between these various representations and process,

can be a catalyst for allowing architects to cross the traditional boundaries
that separate project definition, design, construction and asset operation
 Fazit

and making knowledge reciprocally available across the design intent-construction

execution divide. The tools may make the opportunity, but
practitioners must want to embrace it.

5. Shift the value propositions of design. The commodification of the

architect's services is a primary inhibitor of innovation and value creation
by the profession.
7
Artificial intelligence tools strategically deployed in
the service of performance-enhanced design solutions could be the
catalyst for changing the fundamental business propositions of practice,
converting the value of the architect's services from deliverables and
fixed fees to outcome-based delivery models and related services. AI/ML
systems could radically accelerate the capability of today's algorithmically
driven software tools to predict the future state of project performance,
generating best value by virtue of simulation. As soon as reliable Al-based
tools – as a result of implementation of the previous four strategies–
become widely available, architects could embrace the largest challenges
of architecture and society, and finally escape the tyranny of commodified
fees, limitedresources and public scepticism about the value of the

buildings they design.

The architect Eyal Weizman leads a London-based research team called

Forensic Architecture, which is also the term that he uses to describe his

'investigative practice' that 'regards the common elements of our built

environment – buildings, details, cities and landscapes, as well as their
representations in media and as data– as entry points from which to
interrogate contemporary processes and with which to make claims for
the future'. 8 His team uses architecturalapproach methods supported by
sophisticated data collection and analysis to examine political questions
of government jurisdiction and use of force. The work is not traditional
architecture in that it deconstructs information sources and reassembles
them in ways that might not have been otherwise understood, using digital
tools and analysis. Forensic Architecture's work – which is understandably
highly controversial, especially to the governments he exposes– is
a template

for a strategy for Al in architecture as a whole. The data

fragments that
Forensic Architecture assembles for analysis are the basis for speculation
about future states, which is a fundamental responsibility of an architect,
especially those designing the built environment. To be able to make such
claims about how design actions are manifest, as Weizman declares, architects
need a critical stance about the collection, use and deployment of information
about the world.
An obsessive reliance on data without such a critical view of its results
could lead to disastrous results, particularly in a field of endeavour such
as architectural design that purports to protect the public welfare. In early

research on neural networks applied to medical treatments at a time when

the internal logic of such systems was legible, ML platforms that yielded life-
and-death decisions – even when such systems were generally reliable – led
hospitals to wholly reject the use of Al. They became suspicious when the
systems reported that asthma patients were less, rather than more, likely
to suffer severe consequences from catching pneumonia, a clearly counter-intuitive
result. Apparently, those patients are referred immediately to critical
care early in the disease, and as such have far better outcomes. The neural

network connected the inputs and the outputs, with no understanding

whatsoever of what happened in between. Systems upon which such
decisions made must be both transparent and, in the
are case of professional
judgement like architecture, validated externally.
9

If the necessary data can be collected, and the systems validated, architecture
can expand its remit dramatically. In a recent article addressing the fraught
between architecture and incarceration, Garrett Jacobs and
relationship
Deanna Van Buren, leaders of the non-profit design alliance, Designing Justice
+
Designing Spaces, argue that architects must apply their skills to 'end the
racism that is embedded in the built environment'. To create prototypes for
their 'Alternatives to Incarceration Plan' for Los
Angeles, they have
implemented
adesign process deeply dependent on complex, interrelated data sources,
explaining that 'We are partnering with data visualization, mapping and
research organizations to understand how various systems– such as health
care, first response, pre-arrest diversion, housing, post-incarceration re-entry,
and more – interact at a district scale.' 10 In a pre-digital era, even collecting this
data would have been impossible, much less evaluating it or deploying it as the
basis for new design. At this juncture, when empiricist systems are coming to
the fore, but cognition is largely missing, human designers play an irreplaceable
role to direct data and marshal it. Only then will strategies for computation and
design combined accomplish completely new results that humans create to
improve our condition. This is the optimal outcome of the highest and best uses
of artificial intelligence in architecture.
 BIBLIOGRAPHY

Agrawal A. Gans J. and Goldfarb A. Prediction Machines:

, , , BIM Industry Working Group BIM Management for Value, .

The Simple Economics of Artificial Intelligence Harvard , Cost & Carbon Improvement: A Report for The Government
Business Review Press Boston, MA 2018 , , . Construction Client Group UK Government Department of ,

Alberti L.B. On the Art of Building in Ten Books MIT Press

, , ,
Business, Innovation and Skills London https://www.hse.- . ,

Cambridge, MA 1988 American Institute of Architects. 'AIA , . gov.uk/construction/areyou/principal-designer.htm 2021 ,

Document B101-2017: Standard Form of Agreement Between (accessed 6 June 2021).

Owner and Architect, 2017 American Institute of Architects, , Brockman J. Possible Minds: Twenty-five Ways , of Looking at
Washington, DC: AIA, 2017. Al Penguin Press New York 2019
, , , .

–
The Architect's Handbook of Professional Practice: 15th Brynjolfsson E. and McAfee A. The Second , ,

Edition Wiley Hoboken, NJ 2014

. , , . Machine Age: Work, Progress, and Prosperity in a Time of
-
The Business of Architecture 2020 American Institute of .
Brilliant Technologies W.W. Norton & Company New York , , ,

Architects Washington, DC 2020 . .

2014 .
,

'

Anderson S. Problem-Solving and Problem-Worrying',

,
'
Brynjolfsson E. and T. Mitchell What Can Machine Learning , .

Lectures given at Architectural Association London March Do? Workforce Implications', Science Vol. 358 w, https://doi.- ,
, ,

1966 and ASCA Cranbrook, 5 June 1966. org/10.1126/science.aap8062 2017 (accessed 27 December ,

2020).
Architects Council of Europe with Nacey Research The .

Architectural Profession in Europe 2020 2021 , .

Brynjolfsson E. and A.McAfee 'Will Humans Go the Way of , .

Horses? Labor in the Second Machine Age', Foreign Affairs Vol. ,

Architects Registration Board ARB Criteria at Part 3 Prescription .

94 issue 4 2015 pp 8 14 -
.
, , ,

of Qualifications Architects Registration Board 2012 . , .

Carpo M. The Alphabet and the Algorithm MIT Press

, , ,
-
The Architects Code: Standards of Professional Conduct and
Cambridge, MA 2011 , .

Practice 2017 , .

'
–
The Digital Turn in Architecture 1992-2012 , Wiley , Chichester ,

Asadi E. Li B. and Chen I. Pictobot: A Cooperative Painting Robot

, , ,
2013 .

for Interior Finishing of Industrial Developments', IEEE Robotics &

-
The Second Digital Turn: Design beyond Intelligence MIT Press
Automation Magazine Vol. 25 issue June 2018 2018, p 13 , , , .
, ,

Cambridge, MA 2017 , .

Association of General Contractors of America Construction .

'
Carr N. G IT Doesn't Matter', Harvard Business Review https://-
Inflation Report Q1 2021 AGC America Washington, DC 2021 , , , .
, , ,

' hbr.org/2003/05/it-doesnt-matter 2003 (accessed 15 June ,

Autor D. Why are there Still So Many Jobs? The History

,
2021).
and Future of Workplace Automation ',Journal of Economics
Christian B. The Alignment Problem: Machine Learning and
Perspectives Vol. 29 issue 3 2015 pp 3 30 , , , ,
-
.
,

Human Values W.W. Norton & Company New York, NY 2020 .

, , ,
'
Baker K. How Many Architects Does Our Economy
, '
Corcoran H. Architecture Schools Adapt to an Uncertain
Need?', ARCHITECT 5 January 2018 , .
,

Future ', Architectural Record May 2020 2020, p 1 , , .

Baker K. et. al. Compensation Report 2019 American Institute

, ,

Crawford K. Atlas of Al: Power, Politics, and the Planetary Costs

of Architects Washington, DC 2019 , , .
,

of Artificial Intelligence Yale University Press NH 2021 , , , .

Bernstein H. M. Managing Uncertainty and Expectations in

,
'
Davis D. Architects Versus Autodesk', Architect Magazine 27
Building Design and Construction McGraw Hill Construction ,
, ,

Analytics New York 2014 , , .

August 2020 .

Deamer P. and Bernstein P. Building (in)the Future:

Bernstein P. Architecture, Design, Data: Practice Competency in
,
, ,

the Era of Computation Birkhauser Basel 2018 , , , .

Recasting Labor in ArchitecturePrinceton Architectural ,

'
Press New York 2009 . , .

–
Canonical Models of Architecture ', LOG: Model Behavior Vol. ,
Deamer P. The Architect as Worker: ,

50 issue 2021 pp 239 48

, ,
-
.
Immaterial Labor, The Creative Class, and the Politics of Design ,

Bernstein P. and Deamer Peggy BIM In Education Yale

, , . , Bloomsbury Academic London and New York 2015 , , .

University School of Architecture NH 2011 , , .

Deltek Clarity with Longitude Deltek Clarity Architecture and .

'
Bernstein P. Intention to Artifact' in Digital Workflows
, Engineering Industry Study: Trends and Benchmarks in EMEA and
in Architecture: Design–Assembly–Industry ed. Scott Marble , ,
APAC Deltek 2021, , .

Birkhäuser Basel 2012 , , .

Dickson B. DeepMind scientists: Reinforcement Learning is
,

'
–
The Distractions of Disruptions: Technical Supply in an Era Enough for General Al. BDTechTalks.com https://bdtechtalks.- .

of Social Demand ', Architectural Design (AD) Vol. 90 issue 2 , , , com/2021/06/07/deepmind-artificial-intelligence-reward--

2020 pp 11 , . maximization/ 2021 (accessed 9 June 2021). ,

'
–
Innovative Tools in the Movement Toward Slave-Free –

Why Al Can't Solve Unknown Problems. BDTechTalks.com https://- ,

Building', in Sharon Price Harriet Harriss Phillip Bernstein , , , bdtechtalks.com/2021/03/29/ai-algorithms-representations--

Susan Jones Christ Sharples and Michael Green (eds), Grace ,
herbert-roitblat/ 2021 (accessed 8 June 2021). ,

Farms/Pratt Institute Seminar Series 2020

Domingos P. The Master Algorithm: How The Quest For The
, .
,

Ultimate Learning Machine Will Remake Our World Basic Books, ,

a member of the Perseus Books Group New York 2015 , , .

Egan J. Rethinking Construction: The Report of the Construction
,
-

Reinventing Construction: A Route to Higher Productivity https://-

Task Forcehttp://constructingexcellence.org.uk/wp-content/- www.mckinsey.com/industries/capital-projects-and--
uploads/2014/10/rethinking_construction_report.pdf 1998 , infrastructure/our-insights/reinventing-construction-through--
(accessed 10 April 2021). a-productivity-revolution 2017 (accessed 16 June 2021). ,

'
Folkens B. Chihuahua or Muffin? Cloudsight https://-
, , , Jacobs G. and Van Buren Deanna Decarceration
, , .

blog.cloudsight.ai/chihuahua-or-muffin-1bdf02ec1680 , and Designing to Divest. Architectural Record

https://www.- ,

2017 (accessed 29 December 2020). architecturalrecord.com/articles/14828-commentary--

Ford J. and Galliford T. Information Management According
, ,
decarceration-and-designing-to-divest 5 October 2020 ,

to BS EN ISO 19650-Guidance Part C: Facilitating the Common (accessed 6 October 2020).

Data Environment (Workflow and Technical Solutions) UK BIM , Jha S. Can You Sue an Algorithm for Malpractice? It
,

Alliance London 2020 , , .

Depends, STAT https://www.statnews.eom/2020/03/09/can-- ,

Gallaher M. et. al. Cost Analysis of Inadequate Interoperability in

,
you-sue-artificial-intelligence-algorithm-for-malpractice/ 2020 ,

the U.S. Capital Facilities Industry US Department of Commerce ,

(accessed 25 October 2021).
Technology Administration 2004 , .
Johnston G. B. Assembling The Architect: The History and
,

Gebru T. Bender E. McMillan-Major A. and Mitchell M. On

, , , , , ,
'
Theory of Professional Practice Bloomsbury Visual , Arts ,

London 2020 .
the Dangers of Stochastic Parrots: Can Language Models Be ,

'
Too Big?'FaccT'21 Virtual Canada 2021 , , , . Leach N. Design in The Age of Artificial Intelligence ', Landscape
,

Grace Farms Foundation Architectural Frontiers Vol. 6 issue 2 2018 pp 8 19 -

Design for Freedom Grace Farms , , , , .

. ,

' '
Foundation 2021 , .
–

Digital Buildings? in 1st International

Do Robots Dream of
Gutman R. Cuff D. Wriedt,J. and Bell B. Architecture from the
, , , , ,
Conference on Computational Design and Robotic Fabrication
Outside In: Selected Essays Princeton Architectural Press New , ,
(CDRF 2019) ed. Philip F. Yuan et al. Springer Beijing, , , ,

York 2010 .
China 2020 , .
,

Hao K. The Coming War on the Hidden Algorithms that Trap Marcus G. and Davis E. Rebooting AI: Building Artificial
, ,
,

We Can Trust Pantheon Books New York 2019

Intelligence
People in Poverty, MIT Technology Review , , , .
,

https://www.technologyreview.com/2020/12/04/1013068/- Marcus G. The Next Decade in AI: Four Steps Towards Robust
,

algorithms-create-a-poverty-trap-lawyers-fight-back/ , Artificial Intelligence Unpublished manuscript, https://arxiv.org/- .

2020 (accessed 1 April 2021). abs/2002.06177 2020 (accessed 20 December 2020). ,

'
–
These Weird, Unsettling Photos Show That AI Is Getting –Artificial Intelligence Is Stuck. Here's How To Move It
Smarter, MIT Technology Review https://www.technology-- , Forward ', The New York Times 30 July 2017 , .

review.com/2020/09/25/1008921/ai-allen-institute-gen-- McNulty N. Introduction to Bloom's Revised Taxonomy https://-

, ,

erates-images-from-captions/ 25 September 2020 ,

www.niallmcnulty.com/2019/12/introduction-to-blooms--
(accessed 4 March 2021).
taxonomy/ 2019 (accessed 7 March 2021). ,

–Inside The Fight to Reclaim AI from Big Tech's Control, MIT Metz C. One Genius' Lonely Crusade to Teach a Computer
,

Technology Review https://www.technologyreview.- ,

Common Sense, Wired https://www.wired.com/2016/03/doug-- ,

com/2021/06/14/1026148/ai-big-tech-timnit-gebru-paper-- lenat-artificial-intelligence-common-sense-engine/ 24 March ,

ethics/ 14 June 2021 (accessed 20 March 2021).

,
2016 (accessed 27 December 2020).
Heaven W. AI Is Learning How To Create Itself, MIT
,
Miller A. I. The Artist in The Machine: The World f AI Powered
,

Technology Review https://www.technologyreview.- ,

Creativity The MIT Press Cambridge, MA,2019 , , .

com/2021/05/27/1025453/artificial-intelligence-learning--
create-itself-agi/ 27 May 2021 (accessed 6 June 2021). ,
Mokyr J. Vickers C. and Ziebarth N. L. 'The History of
, , , ,

Technological Anxiety and the Future of Economic Growth: Is

Higgin G.Jessop
, W. N. and the Tavistock Institute of Human
, ,
This Time Different? Journal of Economic Perspectives Vol. 29 , , ,

Relations London Communications In The Building Industry: The ,

issue 3 2015 pp 31 50 -
.
, ,

Report of a Pilot Study Tavistock Publications London 1965 , , , .

'
Moretti E. and ProQuest The New Geography ofJobs
, . ,

Hosey L. Going Beyond the Punchlist: Why Architects Should

,
Houghton Mifflin Harcourt Boston 2012 , , .

Embrace Post-Occupancy Evaluations', Metropolis https://- ,

Mumford S. and Anjum R. Counterfactuals and Causal

www.metropolismag.com/architecture/architecture-post-- , ,

occupancy-evaluations/ 2019 (accessed 15 June 2021). ,

Inference: Methods and Principles for Social Research (Analytical
Methods for Social Research) Oxford University Press New , ,

Illich I. Disabling Professions M. Boyars London and Salem,

, , ,
York 2014 .
,

NH 1977
, .

NBS 10th Annual BIM Report 2020 NBS London 2020

. , , , .

McKinsey Company An Executive's Guide to AI https://www.- . ,

National Council of Architectural Registration Boards Architect .

mckinsey.com/business-functions/mckinsey-analytics/-
our-insights/an-executives-guide-to-ai 2020 (accessed 12 ,
Registration Examination® (ARE®) 5.0 Handbook NCARB , ,

December 2020). Washington. DC 2020 , .

– The Next Normal in Construction: How Disruption is Reshaping Negroponte N. The Architecture Machine MIT Press , , ,

Cambridge, MA 1970 .
the World's Largest Ecosystem 2020 , .
,
 '
Parman J. Is Architectural Licensing Necessary?' Common Susskind D. A World Without Work: Technology, Automation,
,

Edge https://commonedge.org/is-architectural-licensing--
,
And How We Should Respond Metropolitan Books/Henry Holt ,

necessary/ 2020 (accessed 31 October 2020). , & Company New York 2020 , , .

Pearl J. and Mackenzie D. The Book of Why: The New Science of

, , Susskind R. and Susskind D. The Future of the Professions: How
, ,

Cause and Effect Basic Books New York 2018 , , , .

Technology Will Transform the Work of Human Experts Oxford ,

Pelli C. Observations for Young Architects Monacelli Press New

, , ,
University Press Oxford 2015 , , .

York 1999 , . The Infrastructure and Projects Authority Government ,

Construction Strategy 2016-20 Cabinet Office London 2016 .

Peña W. and Parshall S. Problem Seeking: An Architectural
, ,
, , ,

Programming Primer Wiley New York 2001 , , ,

.
Thurairajah N. and Goucher D. Advantages and Challenges of
, ,

Picon A.What About Humans? Artificial Intelligence in

,
Using BIM: a Cost Consultant's Perspective in 49th Annual Associates
Schools of Construction Conference ed. California Polytechnical
Architecture in 1st International Conference on Computational ,

State University San Luis Obispo, California 2013

Design and Robotic Fabrication (CDRF 2019) ed. Philip F. Yuan et , , .
,

al. Springer Beijing, China 2020

, , , . Till J. Architecture Depends MIT Press Cambridge, MA 2009
, , , , .

Rabenek A.The Place , of Architecture in the New Economy ,

Tombesi P. On the Cultural Separation of Design Labor Building
,

Routledge Abingdon, , Oxon 2016 , .

(In) the Future: Recasting Labor in Architecture eds P. Deamer and ,

Rittel H. and Webber M. Dilemmas in a General Theory of P. Bernstein Princeton Architectural Press New York 2010, , , .
, ,

'

Planning, Policy Sciences Vol. 4 issue 2 1973 pp 155 69 , , , ,

-
.
Turing A. Computing Machinery and intelligence ', Mind Vol.
, ,

Rowe P. A Priori Knowledge and Heuristic Reasoning In

' 59 issue 236 October 1950 pp 433 60
, , ,
-
.
,

Architectural Design ',Journal of Architectural Education Vol. 36 , , UK Building Information Alliance Information Management .

issue 1 1982 pp 18 23 , ,
-
.
According to BS EN ISO 19650 Guidance Part 1: Concepts –
.

Royal Institute of British Architects RIBA Plan .

of Work 2020 buiildingSMART London 2019 UK Health and Safety , , .

Overview RIBA,London 2020 .

Executive. Are You a Principal Designer? https://www.hse.-
, ,

' gov.uk/construction/areyou/principal-designer.htm ,
–
Standard Professional Services Contract 2020: Architectural 2016 (accessed 6 June 2021).
Services', RIBA,London 2020 p. 48 , , .

'
–

Principal Designers: Roles and Responsibilities https://www.- ,

Ruhaak A. How Data Trusts Can Protect Privacy', MIT

,
hse.gov.uk/construction/cdm/2015/principal-designers.htm ,

Technology Review https://www.technologyreview.- ,

2015 (accessed 7 June 2021).
com/2021/02/24/1017801/data-trust-cybersecurity-big-tech-- D. Meet the People Who Train the Robots (to
privacy/ 2021 (accessed 3 May 2021). ,
Wakabayashi ,

Do Their Own Jobs) https://www.nytimes.com/2017/04/28/- ,

Russell S.J. Human Compatible: Artificial Intelligence and the

,
technology/meet-the-people-who-train-the-robots-to-do--
Problem of Control Viking New York 2019
their-own-jobs.html 2017 (accessed 11 April 2020).
, , , .
,

Scheer D. The Death of Drawing: Architecture in the Age of

,
Weizman E. Forensic Architecture: Violence at the Threshold
,

Simulation Routledge London and New York 2014

of Detectability Zone Books-The MIT Press Brooklyn, NY
, , , .
, ,

Schön D.A. The Reflective Practitioner: How Professionals Think

,
and Cambridge, MA 2017 , .

in Action Basic Books New York 1983 Witt A. Shadow Plays: Models, Drawings, Cognitions, LOG: Model
, , , .
,

Simonite T. What Really Happened When Google Ousted

' Behavior Vol. 50 issue 2021 pp 29 38
, , ,
-
.
,

Timnit Gebru ', Wired https://www.wired.com/story/google-- ,

Yuan P.F. Xie M. Leach N. Yao J. and Wang X Architectural
, , , , , , , ,

timnit-gebru-ai-what-really-happened/ 2021 (accessed 10 , Intelligence Selected Papers from the 1st International Conference
June 2021). on Computational Design and Robotic Fabrication (CDRF

Slavid R. RIBA Contracts CPD Day Review', Architecture.com

,
'

,
2019) https://yale.idm.oclc.org/login?URL=https://doi.-
,

https://www.architecture.com/knowledge-and-resources/- org/10.1007/978-981-15-6568-7 2019 (accessed 1 May 2020). ,

knowledge-landing-page/riba-contracts-cpd-day-review , Yudkowsky N. and Bostrum N. , ,

2019 (accessed 6 June 2021). The Ethics of Artificial Intelligence Cambridge Handbook ,

A. Ganzoni D. and Carpo M. The Working Drawing: of Artificial Intelligence, Cambridge University Press
Spiro , , , ,
,

The Architect's Tool Park Books Zurich 2013 , , , .

Cambridge 2011 , .

F. and Kent R. Architec-tonics: The Tales of Tom Zarkadakis G. "Data Trusts" Could Be the Key to Better
Squires , ,
,

AI' Harvard Business Review https://hbr.org/2020/11/data-trusts-

Thumtack, Architect The William T. Comstock Company New , ,
, ,

York 1914 , .
could-be-the-key-to-better-ai 2020 (accessed 11 June 2021). ,

,
'
Staff T. A Wave of Venture Capital Is Pouring Into Zweig Group 2019 Fee & Billing Survey Zweig Group . .

Construction Tech Sector', The Real Deal: Real Estate News ,

Consulting Fayetteville, AR 2019. , ,

https://therealdeal.com/national/2019/07/03/a-wave-of--
venture-capital-is-pouring-into-construction-tech-sector/ ,

2019 (accessed 29 December 2020).

Steenson M. Architectural Intelligence: How Designers and ,

Architects Created the Digital Landscape The MIT Press , ,

Cambridge, MA 2017 , .
 REFERENCES

INTRODUCTION Intelligence 'Writing Architecture',

, MIT Design, Data: Practice Competency in the
Press Cambridge, MA 2017
, , . Era of Computation Birkhäuser Basel , , ,

1 For reference, the MacBook Pro 2018 p 23

2 See © RMN-Grand Palais / Art , .

laptop I am writing on at this moment Resource NY 14 Ajay Agrawal Joshua Gans and
has 16GB of internal memory, or , . ,

Avi Goldfarb Prediction Machines:

62,500 times more as that ill-fated 3 This argument is made in detail in ,

PDP-11. It also has yet to catch fire. The Simple Economics of Artificial
Carpo, The Alphabet and the Algorithm .

2 https://en.wikipedia.org/wiki/PDP-- 4 Architecte
Intelligence Harvard Business Review
,

plan < https://cdli.-

au ,
Press Boston, MA 2018
, , .

11 (accessed 5 July 2021). ox.ac.uk/wiki/doku.php?id=architecte_- 15 Spiro et al op. cit p 280

au_plan > (accessed 11 April 2021). ., ., .

3 David R. Scheer The Death of ,

1.2
Drawing: Architecture in the Age of 5 Leon Battista Alberti On the Art ,

Simulation Routledge London and , , of Building in Ten Books MIT Press , ,

1 This quote from Walter Gropius
New York 2014 , .
Cambridge, MA 1988 xxiii , , .
can be found in a superb history and
4 Richard E. Susskind and Daniel 6 Annette Spiro David Ganzoni and , analysis of Nicholas Negroponte's
Susskind The Future of the Professions:
, Mario Carpo The Working Drawing:
, early work on architectural
How Technology Will Transform the The Architect's Tool Park Books Zurich , , , technology: Molly Wright Steenson ,

Work of Human Experts Oxford , 2013 p 279

, . Architectural Intelligence: How Designers
University Press Oxford 1 st edn, , ,
7 See Nicholas Negroponte The ,
and Architects Created the Digital
2015 xiv p 1 , , .
Architecture Machine MIT Press Landscape MIT Press Cambridge, MA
, , ,
, ,

5 This 2017 p 18
particular text was created by a Cambridge, MA 1970 , .
, .

system called InferKit which you can ,

8 One distinct advantage of paper-based 2 Ibid ., p 170 .

try yourself at https://appinferkit.com/- information exchange is that 3 Noam Chomsky at MIT, Terry
generate (first accessed 21 April 2021). it requires no special software nor Winograd at Stanford University and
6 Personal email exchange with Dr data standards, beyond graphic Roger Schank at Yale University (with
Mark Greaves of Pacific Northåwest convention, to transmit or translate it whom I studied in the late 1970s)
National Laboratory 27 November , between originator and consumer. each put forth competing theories of
2020 .
9 The aerospace, automotive and cognition from which various early AI
7 Erik Brynjolfsson and Andrew efforts developed.
manufacturing design disciplines had
Mcafee The Second Machine Age: Work,
, deployed modelling tools decades 4 Gary Marcus and Ernest Davis ,

Progress, and Prosperity in a Time of before architects. Similarly, the Rebooting AI: Building Artificial
Brilliant Technologies W.W. Norton & , form-making and rendering tools of movies Intelligence We Can Trust Pantheon ,

Company New York 1st edn, 2014 , , . and games (such as Autodesk's Maya) Books New York, 2019 p 41
, , .

were appropriated by architects once '

8 Neal Leach has done interesting 5 See AI Winter ', < https://-
work on this question, including the they ran on sufficiently affordable en.wikipedia.org/wiki/AI_winter >
article about computation, design and machines.
(accessed 29 December 2020).
dreaming: Neil Leach, 'Design in the 10 See hypar.io (accessed 11 April 6 This technology is known as 'neural
Age of Artificial Intelligence' Landsc. , 2021). networks' and operates, in part, on a
Archit. Front. 6/ 2 2018 pp 8 19 -

11 For McKinsey & form of statistical correlation called

.
, ,

'
,
example, '
see

9 Peter G. Rowe A Priori Knowledge , Company An executive's guide

, Bayesian probability.
and Heuristic Reasoning in to AI ', < https://www.mckinsey.-
7 Joseph Paul Cohen Paul Bertin and ,

Architectural Design ',Journal of com/business-functions/- Vincent Frappier Chester: A Web

'

Architectural Education 36 / 1 1982 pp , , , mckinsey-analytics/our-insights/- Delivered Locally Computed Chest

18 23
-

an-executives-guide-to-ai?cid=other-- X-Ray Disease Prediction System ',

.

10 Daniel Susskind A World Without ,

eml-alt-mip-mck&hdpid=d36c6b61-- ArXiv abs/1901.11210 (2019).
,

Work: Technology, Automation and 313b-431c-b3d4-141bc805e7e28&hctk- '

8 See Alphago Zero Cheat Sheet',
How We Should Respond Metropolitan y=116253808&hlkid=36ab108977a849a
,

2ba50687fd2e467de>, 2020 (accessed

<
https://medium.com/applied-data--
Books/Henry Holt & Company New ,

20 November 2020). science/alphago-zero-explained--

York 1 st edn, 2020 .
,

12 Steve O'Plear Spacemaker, AI

' in-one-diagram-365f5abf67e0 >
1.1 ,
(accessed 11 April 2021).
software for urban development, is
1 Among the varied histories of 9 Afrequently cited example of AI
acquired by Autodesk for $240M ',
is the teaching of ML systems to
technology in architecture, those by <
https://techcrunch.com/2020/11/17/-
Mario Carpo are perhaps the most spacemaker-ai-software-for-urban-- recognise pictures of cats. Rather than
prescient, particularly Mario Carpo , development-is-acquired-by-- directly program the computer to
The Alphabet and the Algorithm 'Writing autodesk-for-240m/ >, 2020 (accessed understand what a cat looks like, the
,

Architecture', MIT Press Cambridge, , 24 November 2020). systems were trained to study millions
MA 2011 xi and Mario Carpo The of pictures of cats (readily available
, , , ,
13 Phillip G. Bernstein Architecture, ,
on the internet). Eventually, the
Second Digital Turn: Design Beyond
 '

computer could learn to recognise an 16 Peter G. Rowe A Priori , 1.3

image of a cat. This process, through Knowledge and Heuristic Reasoning
various technological improvements, in Architectural Design Journal of
' 1 Richard E. Susskind and Daniel
has been extended to other large Architectural Education 36 no. 1 1982 .
Susskind The Future of the Professions:
,
, , ,

data sets, like all the text ever posted How Technology Will Transform the
17 There are several examples of
online. It is not always successful, Work of Human Experts Oxford ,

the early uses of empiricist systems,

however; see Brad Folkens, University Press , Oxford 2015 p 22 , , .

primarily in construction Software .

'Chihuahua or Muffin ', Cloudsight ,

maker Autodesk has applied 2 See Ivan lllich Disabling Professions , ,

2017 < https://blog.cloudsight.ai/-

,
ML techniques to construction Ideas in Progress, M. Boyars London ,

chihuahua-or-muffin-1bdf02ec1680 > administration data to anticipate and Salem, NH, 1977 .

10 See Pedro Domingos The Master , process failures (such as delays or 3 Donald A. Schön The Reflective ,

Algorithm: How the Quest for the change orders); BuildTech start-up Practitioner: How Professionals Think in
Ultimate Learning Machine Will Remake Smartvid.io uses computer vision Action Basic Books New York 1983
, , , ,

Our World Basic Books New York

, , , and ML to discover job site safety p 5 .

2015 concerns See < www.smartvid.io >

.
.
4 Horst W.J. Rittel and Melvin M.
11 Gary Marcus The ' Next ,
(accessed 30 December 2020). Webber Dilemmas in a General
,
'

'
Decade in AI: Four Steps Towards 18 Cade Metz One Genius' Lonely , Theory of Planning', Policy Sciences 4 ,

Robust Artificial Intelligence', 2020 ,

Crusade to Teach a Computer no. 2 1973 , .

p 51 < https://arxiv.org/ftp/arxiv/-
, Common Sense ', Wired 24 March ,
5 For statistics on the generally
papers/2002/2002.06177.pdf> 2016 .
dismal state of architectural
(accessed 19 December 2020). 19 Marcus op. cit , . ,
p3 .
graduates see Kermit Baker et ,

'
12 For an explanation of a theory 20 Ibid al. Compensation Report 2019 ',
p 27
,
., .

of computational causality (and American Institute of Architects,

21 Personal conversation with Mark
a compelling argument against a Greaves 18 December 2020 .
Washington, DC 2019 ). ,

statistical view of knowledge), see ,

6 Of note for this particular

'

Judea Pearl and Dana Mackenzie The , 22 See Niall McNulty Introduction to ,
comparison is the exchange
Book of Why: The New Science of Cause Bloom's Revised Taxonomy < https://-
programme that has existed between
,

and Effect Basic Books New York

, , , www.niallmcnulty.com/2019/12/- Yale and the University of Cambridge
2018 .
introduction-to-blooms-taxonomy/> for many years, where one graduate
13 This is a convention I always use, (accessed 15 December 2020). from each programme receives a
in my writing and in the classroom, 23 After the now infamous Move 37 full scholarship to attend the other
as a way of acknowledging that the when AlphaGo defeated world Go school. Despite the professionally
expectation is traditionally 'his'. champion Lee Sedol, Demis Hassabis, certified credentials granted by each,
'
14 See Cyc ', < https://en.wikipedia.- the CEO of Google's AI research the graduates from each reciprocal

org/wiki/Cyc > (accessed 31 December project DeepMind, said: 'It doesn't program cannot use their degree to
2020). play like a human and it doesn't play qualify for licensure in the opposite
'
like a program. It plays in a third, country.
15 See Stanford Anderson Problem-Solving ,
almost alien way.' Cade Metz, 'In '

and Problem-Worrying', 7 See Master of Architecture I

Two Moves, AlphaGo and Lee Sedol Professional Degree Program ',
Lectures given at the Architectural Redefined the Future ', WIRED (2016 ).
Association London March 1966
<
https://www.architecture.yale.edu/-
, , <
https://www.wired.com/2016/03/- academics/programs/1-m-arch-i >
and ASCA Cranbrook, 5 June 1966, as
two-moves-alphago-lee-sedol-- (accessed 8 January 2021) for a
referenced by Steenson, Architectural redefined-future/ > (accessed 11 July
detailed list of course requirement for
Intelligence: How Designers and 2021).
Architects Created the Digital Landscape. the M.Arch 1 degree, which qualifies a
An intermediate step between 24 This model mirrors one created successful graduate to sit for the ARE.

problem-solving and problem-worrying, by McKinsey Consulting that used 8 National Council of Architectural
to explain the possibility of AI in
as Anderson defined it, Registrations Boards (NCARB) ,

a business context, where the '

might be William Peña's strategy for Architect Registration Examination®

classifications are
problem-seeking, an approach to (Are®) 5.0 Handbook', Washington,
defining an architectural problem 'Descriptive/Predictive/Prescriptive'. See McKinsey
'
DC : NCARB 2020 , .

& Company An Executive's Guide

before diving headfirst into the design ,

9 Architects Registration Board (ARB)

to AI ', < https://www.mckinsey.- ,

process itself. His definitive text on '

ARB Criteria at Part 3 Prescription
the topic, originally written in 1977, com/business-functions/mckinsey--
of Qualifications ', ed. Architects
was updated in 2001; see William analytics/our-insights/an-executives--
Peña and Steven Parshall Problem guide-to-ai > (accessed 18 December Registration Board 2012 , .

2020 ). 10 For an eloquent exploration of

Seeking: An Architectural Programming some of these differences in opinion,
Primer Wiley New York 2001
, , , . 25 Pedro Domingos , op. cit p xxi , .

see Jeremy Till Architecture Depends , ,

26 Marcus op. cit , . ,

p3 .
MIT Press Cambridge, MA 2009 .
, ,
11 Stuart J. Russell Human , Intelligence: How Designers and 2 Ibid ., p 94 .

Compatible: Artificial Intelligence and Architects Created the Digital Landscape ,

3 While the theoretical construct
the Problem of Control Viking 2019 , , , MIT Press Cambridge, MA 2017
, , .
of parametric, 3D modelling had
p 87 .
3 As quoted by Daniel Susskind in been around for decades before
12 Ibid ., p 20 . Daniel Susskind, A World Without Work: the term 'Building Information
13 See Mario Carpo The Second ,
Technology, Automation and How We Modelling' and viable tools for its
Should Respond Metropolitan Books/Henry use became available on or about
Digital Turn: Design Beyond Intelligence ,
,

Holt & Company New York 2002, very widespread adoption did
'Writing Architecture', MIT Press ,

2020 p 129
, ,

not occur until almost 20 years later.

Cambridge, MA 2017 , .
, .

4 Neil Leach Do Robots Dream of

' The American Institute of Architects
14 These images can be generated ,

Digital Buildings? paper presented at reported in its 2020 Firm Survey that
at the Allen Institute for AI < https://- ,
,

the 1st international Conference on approximately 80% of US firms were

vision-explorer.allenai.org/text_- BIM capable in 2019, an extremely
to_image_generation > (accessed 8 Computational Design and Robotic slow adoption curve for any new
Fabrication (CDRF 2019), Beijing,
January 2021), as described in Karen technology.
China 2020
Hao, 'These weird, unsettling photos , .

showthat AI is getting smarter', MIT 5 Antoine Picon What About

' 4 Daniel Susskind A World Without ,
,

Work: Technology, Automation and

Technology Review, 25 September Humans? Artificial Intelligence
2020. in Architecture ', Architectural How We Should Respond Metropolitan ,

Intelligence: Selected Papers from Books/Henry Holt & Company New ,

15 A favourite example is www.- York 2020 pp 38 39 -

thiscatdoesnotexist.com (accessed the 1st International Conference on , , .

8 January 2021), an AI-driven site Computational Design and Robotic 5 Ibid ., p 73 .

that creates photorealistic images of Fabrication (CDRF 2019) ed. Philip F. ,

6 Niall McNulty Introduction
' to ,

cats that it conjures based on vast

Yuan et al. p 22 , .
Bloom's Revised Taxonomy ', < https://-
collections of pictures of cat faces, 6 Susskind op. cit , . ,
p 43 .
www.niallmcnulty.com/2019/12/-
a resource infinitely available on introduction-to-blooms-taxonomy/ >
7 Richard E. Susskind and Daniel
the internet. The site creates very Susskind The Future of the Professions: (accessed 7 March 2021).
,

realistic feline visages but could not How Technology Will Transform the 7 Susskind op. cit , . ,
p 78 .

answer even a basic question about

Work of Human Experts Oxford ,
8 Stuart J. Russell Human Compatible:
its subjects. ,

University Press Oxford 2015 p 271 , , , .

Artificial Intelligence and the Problem of

16 See Phillip G. Bernstein 8 Andrew Witt Shadow
'
Control Viking New York 2019 p 87
,
, Plays: , , , , .

Architecture, Design, Data: Practice Models, Drawings, Cognitions', LOG: 9 Institute of British Architects
Royal
Competency in the Era of Computation ,
Model Behavior 50 2021 '

(RIBA) RIBA Plan of Work 2020

, , .

Birkhäuser Basel 2018

, , .
,

9 The International Standards Overview', RIBA London 2020 , , .

17 César Pelli, for whom I worked

Organization (ISO) has developed 10 American Institute of Architects
during this time, was particularly data and process workflow standards '
,

B101-2017 Standard Form of

adept at this, and seemed to have for building projects based on
an encyclopaedic memory for Agreement between Owner and
Building Information Modelling Architect', American Institute of
materials, plan configurations, and (BIM) technologies and data. The
other knowledge of past projects at Architects Washington, DC 2017 , , .

standard requires creation of a

the ready during any design review. Common Data Environment, which is 11 Russell op. cit , . ,
p 86 .

While looking at the proposed plan defined at https://www.iso.org/obp/- 1.6

for, say, a new office tower, he would ui/#iso:std:iso:19650:-1:ed-1:v1:en ,

ask everyone: 'Remember the second Section 3.3.15 (accessed 1 February 1 Typical models vary by the role
version of the core that we suggested of the contractor, the relationship
2021).
for the first tower at Canary Wharf? between the designer and the builder,
That would work here. 'Those of 10 This robot is designed to paint the number of contracts between
us who were younger, and far less
an interior wall in concert with a various participants and the client,
human painter, see B. Li I. Chen
experienced, found this ability both '
,
and the timing of ascertaining the
and E. Asadi Pictobot: A Cooperative
extremely useful and incredibly cost of construction. Examples include

daunting. Painting Robot for Interior Finishing of 'design-bid-build,''public-private

Industrial Developments ', IEEE Robotics
1.4 & Automation Magazine 25 June 2018
partnership', and 'integrated project
, ,
delivery'. Fora detailed explanation
1 As it was 1979, every single p 85 .
of such models in the United States,
architect on the floor, save one, was a 1.5 American Institute of Architects
see ,

white male. The Architect's Handbook of Professional

1 Richard E. Susskind and Daniel Practice: 15th Edition Wiley Hoboken,
2 For a superb history and analysis Susskind The Future of the Professions:
, ,

of Negroponte's early work on

,
NJ 2014 Chapter 9.1 'Project Delivery
, ,

How Technology Will Transform the

architectural technology, see Molly Methods', p 508 .

Work of Human Experts Oxford ,

Wright Steenson Architectural ,

University Press Oxford 2015 p 1 , , , .
2 Sir John Egan is generally credited 'Design Assist', where certain 5 Revenue numbers from AIA Firm
with formulating the basis of this aspects of the project, typically the Surveys of 2006 and 2014, overlaid
concern and laying the groundwork mechanical and electrical systems, with employment data from Kermit
'
for modern uses of technology and are designed by the engineer of Baker, How Many Architects Does Our
alternative project delivery models, record, but final traditional working Economy Need?' in ARCHITECT 2018 5 , ,

at least in the UK. See Sir Roger drawings are omitted and the January 2018.
Egan, 'Rethinking Construction: The design itself is documented in detail 6 Economic data on the performance
Report of the Construction Task Force' in shop drawings prepared by of the US architectural industry is
(1998). However, the systematic the subcontractor who is going to
exceedingly difficult to collect, as
exploration of these questions may fabricate and install the system. This the best potential source, the AIA,
be seen to have begun much earlier approach acknowledges that the is constrained, as described below,
by the National Joint Consultative actual, useful value of 'design intent' from collecting and evaluating it.
Committee of Architects, Quantity construction documents for such As such, this chart was derived
Surveyors and Builders in a study systems is minimal, and therefore from a combination of AIA sources,
prepared by the Tavistock Institute of they can be abandoned. specifically the work of Kermit
Human Relations; see Gurth Higgin ,
9 Higgin Jessop
, and Tavistock Baker cited in n. 5, plus additional
William Neil Jessop and the Tavistock Institute of Human Relations London information provided by the AIA
,

Institute of Human Relations London

op. cit. p 43 by email in May 2021, along with
,
, .

Communications in the Building

10 See McKinsey & Company The profitability data generated by the
Industry; the Report of a Pilot Study ,
,

Zweig Group consulting company,

Tavistock Publications London 2nd Next Normal in Construction: How
, ,
which does not have the same
edn, 1965 Disruption is Reshaping the World's
legal nervousness as AIA. Zweig,
.

Largest Ecosystem ', 2020 .

3 Even when designers and builders '

however, tracks both architects and
11 Paolo Tombesi On the Cultural
use digital tools, the exchange of ,
engineers, so the profit data is likely
digital information mostly drawings
– Separation of Design Labor ', Building more volatile than if only architects
-
is still limited by traditional concerns (in) the Future: Recasting Labor in were considered. Finally, the AIA

about precision and liability, resulting Architecture eds. Peggy Deamer

,
collects firm data on a very irregular
in enormous inefficiencies as and Phillip G. Bernstein Princeton ,
basis, so intervening years (indicated
Architectural Press New York 2010
information is recreated along each by the hatch marks on the graph)
, , .

step in the delivery process: design, 12 Andrew Rabeneck The ' Place of ,
are extrapolated from available

fabrication, and construction. That Architecture in the New Economy', data on revenue (indicated by the
inefficiency is leading to modern Industries of Architecture eds. Katie ,
solid blue bars). Sources: personal
attempts to integrate and streamline Lloyd Thomas et al Routledge , , correspondence with Kermit Baker,
the process. For example, the list of Abingdon, Oxon 2016 p 192 , , . 15 May 2021; AIA Firm Surveys 'The
companies that prefabricate large 13 Ibid ., p 192 .
Business of Architecture', 2002 2005 , ,

portions of buildings lengthens 2008 2011 2013 2015 2017 2019 ;

, , , , ,

2.1
each fiscal quarter: Bryden Wood, Zweig Group, '2019 Fee & Billing
'

Blockable, DIRTT, FabCab, and '

1 See Creative Destruction ', < https://- Survey in Fee & Billing Surveys Zweig ,

Mace, to name a few. The author

en.wikipedia.org/wiki/Creative_- Group Consulting Fayetteville, AR ed.
, ,

is a collaborator on a study of such destruction > (accessed 30 April 2021 ). Will Swearingen 2019, p 208
, .

companies to be published in 2022 .

7 I have been engaged in an ongoing
2 Daniel Susskind A World Without ,

4 Higgin, Jessop and the Tavistock conversation with the Chief Economist
Work: Technology, Automation and
Institute of Human Relations London ,
How We Should Respond Metropolitan of the American Institute of Architects,
,

op. cit p 77
. , .

Books/Henry Holt & Company New ,

Kermit Baker, about this question. I
have no doubt that Kermit could easily
5 See 'Design Intent' in the Designing York 2020 p 35
, , .

Buildings Wiki at < https://www.- generate a metric for productivity in

3 At this writing in mid-2021, there the American profession, where he is
designingbuildings.co.uk/wiki/Design_- are several such companies offering the primary researcher responsible
intent > (accessed 11 April 2021).
early planning analysis studies for decades of bi-annual Business of
6 Higgin,Jessop and Tavistock created by AI-driven generative design Architecture surveys. However, owing
Institute of Human Relations London , processes, including Autodesk's to two anti-trust sanctions enforced
op. cit p 40
. , .
Spacemaker AI, Sidewalk Lab's Delve, by the US Justice Department on the
7 For an excellent exploration of this and start-ups Envelope and Plot-Z.
AIA, Kermit must stay far from issues
evolution in the United States, see 4 In the US-based office where I was of compensation, relative levels of
George Barnett Johnston Assembling, practising in 1990, a single PC-based effort and their relationship. This
the Architect: The History and Theory computer with monitor and software likely makes creating any productivity
of Professional Practice Bloomsbury
, exceeded $25,000 per workstation, measurement system nearly
Visual Arts London 2020
, , . more than $50,000 escalated to impossible, and thus these surveys
8 Another variation of delivery 2021. That cost is almost equal to the are silent on this topic.
models current in vogue is called annual salary of the young architect
who operated it.
8 The origins of this argument can 7 UK Architects Registration Board , 2.3
be found, in the pre-AI era, in Phillip The Architects Code: Standards of
G. Bernstein Intention to Artifact',
' 1 Or, in a limited number of US states,
, Professional Conduct and Practice ',
in Digital Workflows in Architecture: 2017 p 14 . by alicensed structural engineer.
,

Design-Assembly-Industry ed. Scott ,

8 National Council of Architectural 2 Health and
'
Safety Executive
Marble Birkhäuser Basel 2012 (HSE) Principal designers: roles and
Registration Boards op. cit p 6
, , , . ,
, . , .

9 For a provocative exploration of 9 Remember that you do not actually

responsibilities', 2015 < https://www.- ,

data that enables AI and its ethical own that software, you are merely
hse.gov.uk/construction/cdm/2015/-
and economic implications in the era principal-designers.htm > (accessed 7
licensing its use for your purposes, June 2021).
of big tech, see Kate Crawford Atlas ,
and those purposes had best conform
of AI: Power, Politics, and the Planetary to the extensive terms and conditions
3 For a detailed explanation of
Costs of Artificial Intelligence Yale ,
of the EULA, which you are very American phases of work, see
University Press New Haven, CT , ,
unlikely to have bothered to read.
American Institute of Architects The ,

2021 .
'
Architect's Handbook of Professional
10 See Autodesk LICENSE AND Practice: 15th Edition Wiley Hoboken,
10 Private conversation between SERVICES AGREEMENT', < https://-
, ,

Mark Greaves of the Pacific Northwest NJ 2014

, .

download.autodesk.com/us/FY17/-
National Laboratory and the author, Suites/LSA/en-us/lsa.html > (accessed 4 It is instructive that there is no
16 April 2021. 6 June 2021). British counterpart to the American
11 See Anouk Ruhaak How Data ,
'

11 During my time at Autodesk, the

phase of 'Procurement', when
contractors compete to be selected
Trusts Can Protect Privacy' in MIT
roadmap for a typical annual release for construction of the project. This
Technology Review 24 February 2021 , .
of Revit was in thirds: one each for
implies that design and construction
2.2 design, coding and testing. activities in the US are possibly more
12 Recent assertions by AI-enabled separated, and early contractor
1 Karen Hao The Coming War on the
,
technology giants such as Facebook involvement less likely. In contrast,
Hidden Algorithms That Trap People and Twitter reinforce this prediction, the lack of clear definition of the
'
in Poverty in MIT Technology Review ,
as each has asserted that they are architect's role in translating the
2020 < https://www.technologyreview.- mere 'platforms' with little or no
,
design into a price in the UK system
com/2020/12/04/1013068/algorithms-- responsibility for the consequences suggests little responsibility by
create-a-poverty-trap-lawyers-fight-- of their use by customers. See, for designers and far more control, earlier
back/ > (accessed 6 June 2021).
example, 'Zuckerberg says Facebook in design, by UK builders.
2 Royal Institute of British Architects not responsible for US Capitol
,
5 Leon Battista Alberti On the Art ,

'Standard Professional Services Riots' at < https://www.ft.com/-

Contract 2020: Architectural Services', content/39a699fc-1730-4a5d-b43e--
of Building in Ten Books MIT Press , ,

634ebe189d79 > (accessed 6 June

Cambridge, MA 1988 , .

RIBA London 2020 p 48

, , , .

6 This interpretation of Alberti has

3 American Institute of Architects 2021).
'
,
been put forward by the historian
Document B101 2017: Standard -
13 According to news reports, Mario Carpo, specifically in Mario
Form of Agreement between Owner 'with the "sole exception" of the
Carpo The Alphabet and the Algorithm
, ,

and Architect', American Institute of Royal Borough of Kensington and

Chelsea – which accepted that the
'Writing Architecture', MIT Press ,

Architects Washington, DC 2017

Cambridge, MA 2011
, , .
, .

4 Health and Safety Executive of HM refurbishment work should not have '

' been signed off all organisations

-
7 See Antoin Picon's What About
Government Are you a principal
,
Humans? Artificial Intelligence in
had denied responsibility in "carefully
designer? ', < https://www.hse.gov.- crafted statements'". See 'Grenfell
'
Architecture in Philip F. Yuan et al. ,

uk/construction/areyou/principal-- Tower fire: Inquiry told firms "deny Architectural Intelligence, Selected
designer.htm > (accessed 6 June 2021). Papers from the 1st International
responsibility'" at < https://www.bbc.-
5 National Council of Architectural com/news/uk-51256738 > (accessed 6 Conference on Computational
'

Registration Boards Legislative ,

June 2021). Design and Robotic Fabrication
Guidelines and Model Law- Model ' (Cdrf 2019), < https://yale.idm.-
14 John J. Parman Is Architectural
Regulations 2016-2017 ', NCARB ,
,
oclc.org/login?URL=https://doi.-
Washington, DC 2016 , .
Licensing Necessary?' in Common Edge ,
org/10.1007/978-981-15-6568-7 >
2020 < https://commonedge.org/-
6
According both RIBA and AIA
to
,
(accessed 6 June 2021).
is-architectural-licensing-necessary/>
sources, a vast majority of architects 8 See Judea Pearl and Dana
(accessed 6 June 2021).
in both countries are using BIM tools Mackenzie The Book of Why: The
,

' 15 See Daniel Susskind A World

today. See NBS 10th Annual Bim ,

Without Work: Technology, Automation

,
New Science of Cause and Effect Basic ,

Report 2020 ', NBS London 2020 p 9 , , , , Books New York 2018
, , .

and How We Should Respond

and American Institute of Architects, ,

9 See Judea Pearl The Limitations of

'

'The Business of Architecture 2020', Metropolitan Books/Henry Holt & ,

'

Company New York 2020 . Opaque Learning Machines in john

American Institute of Architects, , ,

Brockman Possible Minds: Twenty-Five

,

Washington, DC, 2020, p 30.

 Ways of Looking at Ali , Penguin Press , / Master of Science in Computer 5 This is the characterisation of such
New York 2019 p 17 , , . Science or Information Technology. systems by computer scientist Mark
10 Ibid., p 23 . 5 Of the 120 credit hours necessary Greaves. Personal correspondence
' to achieve the equivalent 'Part 2' between the author and Greaves, 7
11 Daniel C. Dennett What Can We
'
,

professional degree in Yale's Master June 2021 .

Do in Brockman p 52 , .

of Architecture/First Professional 6 Those systems use the current

12 Ibid p 51 Degree, nine are attributed to courses technique called reinforcement
., .

13 See Grace Farms Foundation , in visualisation, where theoretical or learning, where the system is
'

Design for Freedom ', Grace Farms historical material is connected to programmed to try alternative
Foundation 2021 < https://www.-
, .
drawing or digital tools. One summer strategies and instantiate successes
designforfreedom.org/wp-content/- course of three credits is dedicated and reject failures; they are the
uploads/2020/10/DesignforFreedom_- directly to BIM, in connection current rage among AI programmers

FullReport_L.pdf> (accessed 7 June with a building construction and in 2021, as described in Ben Dickson,
2021). documentation requirement. 'DeepMind Scientists: Reinforcement
14 Dennett in Brockman p 46 , . 6 Dennett in John Brockman Possible ,
Learning Is Enough for General AI',
Minds: Twenty-Five Ways of Looking at BDTechTalks.com 2021 < https://-
, ,

2.4 bdtechtalks.com/2021/06/07/-
AI Penguin Press New York 2019
, , , ,

1 As quoted in Ben Dickson, Why

'

p 46 .
deepmind-artificial-intelligence--
AI can't solve unknown problems ', reward-maximization/ > (accessed 8
7 Russell in ibid p 32 .

BDTechTalks.com (2021 ), < https://-

.,
June 2021).
bdtechtalks.com/2021/03/29/ai-- 8 NCARB, '2012 Ncarb Practice 7 This is not to say that certain,
algorithms-representations-herbert-- Analysis of Architecture', (National highly constrained building types
roitblat/ > (accessed 8 June 2021). Council of Architectural Registration that operate on strict templates and
Boards, 2013). Note that at the time are relatively simple to create e.g. –

2 For two examples of early but of this writing, NCARB was completing a fast-food restaurant or an Apple
unrequited enthusiasm for BIM, see the 2020 Practice Analysis. Store might not be suitable for AI
–

Phillip Bernstein and Peggy Deamer ,

eds., BIM in Education Yale University ,

9 The technical definition can be generation. But most projects are so
School of Architecture New Haven, found as the UK BIM Standard site- and condition-specific that it is
,

CT,: 2011 ; Phillip G. Bernstein and and ISO 19650 ; see < https://www.- difficult to imagine an AI sophisticated

Peggy Deamer eds. Building (in) the ,

ukbimframework.org/standards-- enough to manage competing design
Future: Recasting Labor in Architecture ,
guidance/> (accessed 8 June 2021). constraints and trade-offs.
Princeton Architectural Press New , 10 UK, The Infrastructure and 8 See Chapter 2.1, The
' Digital
' '
York 2009 , .
Projects Authority Government ,
Transformation of Design in Phillip
3 It is an oversimplification to Construction Strategy 2016-20 ', G. Bernstein Architecture, Design,
,

Cabinet Office London 2016 Data: Practice Competency in the Era of

attribute this failure to the academy , , .

alone. The building industry adopts 11 UK BIM Alliance ,' Information Computation Birkhäuser Basel 2018
, , , .

new technologies and processes at Management According to Bs En Iso 9 Cost modelling in the early stages
a glacial pace, and curricula evolve 19650 Guidance Part 1: Concepts',
–
of a project are parametric (area
even more slowly, even when there buildingSMART London 2019 , , . × cost/unit of area), where later in

may be advocates for exploiting the a project estimates are based on

3.1
opportunities of new tools on the quantitative take-offs (measurements
faculty (which is also rare). Software 1 Leon Battista Alberti On the Art , of installation × cost of materials and
vendors (like my former employer) of Building in Ten Books MIT Press , , labour). It will be important for any
share some of the responsibility Cambridge, MA 1988 , . such system to differentiate these
here as well, having failed to support 2 Mario Carpo The Alphabet and the ,
approaches based on the resolution
academic engagement with BIM while of the design.
Algorithm 'Writing Architecture', MIT
,

simultaneously dramatically slowing Press Cambridge, MA 2011 p 167

, , , . 10 Carpo op. cit. p 21
, , .

the maturation and development of

BIM platforms as they become more 3 Ibid ., p 21 . 11 These accusations are typically in
the form of architects 'not knowing
profitable to sell. See https://letters-- 4 This pursuit of design perfection,
how a building goes together,' or 'how
to-autodesk.com/letter-to-autodesk.- which is of course not possible (nor
one would really build that'.
pdf 'An open letter that reflects
, legally advisable as a warranty of
customer perspectives on Autodesk completeness) is the source of much 12 Ajay Agrawal Joshua Gans and
,

2020' (accessed 11 June 2021). unfortunate behaviour first learned Avi Goldfarb Prediction Machines:
,

4 See in school and then transported to the The Simple Economics of Artificial
Liverpool's Building Information
office studio. Intelligence Harvard Business Review
Modelling and Digital Transformation ,

Press Boston, MA 2018 p 24

(BIM-DT) MSc or the University of , , , .

North Carolina/Charlotte's Dual

Master of Science in Architecture
3.2 9 Kate Crawford Atlas of AI: Power, , Construction Industry', Architosh ,

'
Politics, and the Planetary Costs of <
https://architosh.com/2018/01/-
1 See BuildingSMART', < https://- Yale University
Artificial Intelligence , facebooks-vc-backs-new-doxel-ai--
en.wikipedia.org/wiki/BuildingSMART> Press New Haven, CT 2021 p 12
, , , .
and-computer-vision-to-disrupt--
(accessed 11 June 2021). construction-industry/ > (accessed 16
10 'Veridical'
meaning 'truthful,
2 During my time at Autodesk, I September 2021).
veracious, genuine' < https://www.- ,

served in various capacities with 7 See discussions in Chapter 2.4

merriam-webster.com/dictionary/-
BuildingSMART, primarily as the veridical > (accessed 11 June 2021). about FRDM.org .

company's executive representative.

11 This research is described at 8 Niraj Thurairajah and Dan Goucher ,

3 As part of the UK Level 2 BIM '

length in Brian Christian, The Alignment Advantages and Challenges of Using

standard, the government established Problem: Machine Learning and Human BIM: A Cost Consultant's Perspective
'

a data exchange standard through its

Values W.W. Norton & Company New
, , in 49th Annual Associates Schools of
original data specification PAS 1192-2, York 2020 pp 45 ff.
, , Construction Conference California ,

subsequently superseded by BS EN Polytechnical State University San

ISO 19650. Within the PAS standard 12 Timnit Gebru
Emily Bender , ,
,

Luis Obispo, California 2013

Angelina McMillan-Major and , .

the requirement was to generate data '

in COBIE (Construction Operations Margaret Mitchell On the Dangers , 9 While these conclusions are now
of Stochastic Parrots: Can Language being substantiated by firms outside
Building Information Exchange) '

which is, in essence, a spreadsheet-based Models Be Too Big? in FaccT'21 , the building industry like McKinsey,
extraction of non-geometric Virtual Canada 2021 , , .
many of these concerns were defined
data designed to expose certain 13 Crawford op, cit , . ,
Chapter One by Roger Egan in Sir Roger Egan,
characteristics of the underlying 'Earth', pp 23 ff. 'Rethinking Construction: The Report
of the Construction Task Force',
design. This is a complicated set 3.3
of standards and interactions. A 1998. In the US, statistics about
conformance to budget and schedule
good place to start to understand 1 This robot was Doxel's first iteration
them can be found at https://www.- for such a process. In subsequent range from 30% to 60% of projects
designingbuildings.co.uk/wiki/BIM_- version, the company developed failing to meet these goals.
level_2 (accessed 12 June 2021). a technology called VSLAM (vision-based 10 Nicholas Negroponte The ,

4 This move can come none too soon, simultaneous localisation and Architecture Machine MIT Press , ,

as construction is considered one of mapping) that is much less expensive Cambridge, MA 1970 , .

the least digitised (and productive) and easier for construction managers 11
'
Foster
+ Partners and Boston
to deploy. Email correspondence
industries globally. See McKinsey monitor construction with
Dynamics
Global Institute ,'Reinventing between the author and Kevin
'Spot'the robot dog ', Design Boom ,

Construction: A Route to Higher Ferguson of Doxel, 30 September <

https://www.designboom.com/-
2021 and https://www.doxel.ai/-
Productivity' in https://www.mckinsey.- architecture/foster-partners-boston--
com/industries/capital-projects-and-- post/360-degree-cameras-vslam dynamics-construction-spot-robot--
infrastructure/our-insights/reinventing-- (accessed 30 September 2021). dog-11-11-2020/> (accessed 12 June
construction-through-a-productivity-- 2 'BuildTech' is the term now 2021).
revolution ed. McKinsey & Company,
, generally accepted to refer to 3.4
McKinsey & Company 2017 , .
emergent digital technology in the '

5 https://www.buildingsmart.org/- built asset marketplace. See What 1 Deltek Clarity with Longitude,
'

about/vision/ (accessed 11 June 2021). does BuildTech mean?', < https://www.- Deltek Clarity Architecture and
archdaily.com/924827/what-does-- Engineering Industry Study: Trends
6 Michael P. Gallaher Alan C. ,
buildtech-mean (accessed 11 June and Benchmarks in Emea and Apac in
'

'
O'Connor et al. Cost Analysis of Annual Comprehensive Reports Deltek
,
2021). , ,

Inadequate Interoperability in the '

3 AI Powered 2021
Project Controls',
.

U.S. Capital Facilities Industry', U.S.

<
https://www.doxel.ai/#press> 2 Architects' Council of Europe
Department of Commerce Technology
Administration 2004 , .
(accessed 12 June 2021). with Mirza & Nacey Research The ,

4 Royal Institute of British Architects Architectural Profession in Europe

7 Niraj Thurairajah and Dan '
,

2020 ', 2021

'
Goucher Advantages and Challenges Standard Professional Services .

of Using Bim: A Cost Consultant's Contract 2020: Architectural Services', 3 Construction Documents is typically
RIBA London 2020 p 41 35% of the basic fee, and Construction
Perspective', 49th Annual Associates , , , .

Schools of Construction Conference 5 American Institute of Architects Contract Administration another 20%.
, ,

California Polytechnical State

'
B101-2017 Standard Form of Some large portion of this effort is

University San Luis Obispo, California

, , Agreement between Owner and likely to be automated. The question
2013 . Architect', American Institute of whether that work is replaced with
Architects Washington, DC 2017 something of more value will be
8 George Zarkadakis "'Data Trusts" ,
, , ,

addressed in Chapter 3.5: Value

Could Be the Key to Better Ai' in Sub-paragraph 3.6.2.1, p 9 .

Propositions and Business Models.

Harvard Business Review 2020 https://- , , 6 'Facebook's VC Backs new Doxel—
hbr.org/2020/11/data-trusts-could-- AI and Computer Vision to Disrupt
be-the-key-to-better-ai (accessed 11
June 2021).
4 Daniel Susskind A World without , architectural fees stems from several CONCLUSION
Work: Technology, Automation and lawsuits brought against the former
How We Should Respond Metropolitan 1 Arthur I. Miller The Artist in the
, by the US Department of Justice in ,

the 1980s and 1990s, where standard Machine: The World of AI Powered
Books/Henry Holt & Company New ,

York 2020 pp 37 ff.

, , fee schedules, still available from Creativity MIT Press Cambridge, MA
, ,

RIBA guidelines, were deemed to be 2019 p 5

, .

5 The
' Architectural Profession in
Europe 2020 '. anti-competitive. The AIA has therefore 2 Recently, a collective of UK architects
removed itself entirely from any has expressed strong misgivings about
6 According to the Architects' Council discussions whatsoever of fees. Autodesk's lagging development of the
of Europe, 97% of firms in Europe have
4 Nicolas G. Carr It Doesn't Matter in
' '
flagship BIM platform, Revit. See Daniel
10 or fewer staff. ,
'
Davis Architects Versus Autodesk
'

'
Harvard Business Review 2003 < https://- , ,
,

7 See Construction jobs BOOM: Architect Magazine 27 August 2020

hbr.org/2003/05/it-doesnt-matter
.
> ,

Bricklayers and plasterers earn MORE '

(accessed 11 June 2021). 3 BIM Industry Working Group ,

than architects at < https://www.- '

Bim Management for Value, Cost &
5 Many US architects report that,
express.co.uk/news/uk/930079/UK-- Carbon Improvement: A Report for
jobs-construction-salary-brickiayers-- despite the slow recovery from the
2007-09 crisis, while fee volume has the Government Construction Client
electrician-plumbers-career-university> returned to pre-crisis levels (at least Group', UK Government Department of
(accessed 12 June 2021). Business, Innovation and Skills London
' prior to Covid), fee contractions have ,

8 See Architects who don't pay March 2011

yet to return from pre-crisis levels.
.

"
interns shouldn't be given prestigious 4 Issues of forced labour and modern
commissions says designer who
" 6 Ibid ., p 117 .

7 For a more detailed explanation

slavery were explored in a seminar
revealed Ishigami internships' at
of the course and its outputs, see
taught at Yale in autumn 2020. See
<
https://www.dezeen.com/2019/03/25/- <
https://www.architecture.yale.edu/-
architects-unpaid-internship-- Phil Bernstein, 'The Distractions of
courses/20242-fighting-slavery-in-the--
serpentine-pavilion/ > (accessed 12 Disruptions: Technical Supply in an building-supply-chain > (accessed 11
June 2021). Era of Social Demand' in Architectural
July 2021).
9 Peggy Deamer The Architect as Design (AD) 90 no. 02 2020
, , , .
'
,

'
5 See Phillip G. Bernstein A Way ,

Worker: Immaterial Labor, the Creative 8 Lance Hosey Going Beyond the Forward? Integrated Project Delivery
,
',
Class, and the Politics of Design , Punchlist: Why Architects Should Harvard Design Magazine Spring 2010 , .

Bloomsbury Academic London and ,

Embrace Post-Occupancy Evaluations',
New York 2015 6 A typical example is the alleged
, .
Metropolis 2019 < https://www.-
, ,

uselessness of the architect's building

10 See "'I'm not a robot": Amazon metropolismag.com/architecture/-
information models for construction.
workers condemn unsafe, grueling architecture-post-occupancy--
evaluations/> (accessed 15 June 2021). Those data are created to fulfil the
conditions at warehouse' at architect's requirement to define
'
<
https://www.theguardian.com/- 9 Harvey M. Bernstein Managing ,
design intent, and lack the additional
technology/2020/feb/05/amazon-- Uncertainty and Expectations in '
construction logic and detail that
workers-protest-unsafe-grueling-- Building Design and Construction builders require to complete their work.
conditions-warehouse > (accessed 28 in McGraw Hill Smart Market Reports ,

October 2021) ed. Harvey M. Bernstein McGraw Hill 7 See Phillip G Bernstein Architecture .
,

Construction Analytics New York Design Data: Practice Competency in the

11 Timnit Gebru Emily Bender , ,
, ,

Era of Computation Birkhäuser Basel

2014 p 12 , , ,

Angelina McMillan-Major and , .

'
2018 .

Margaret Mitchell On the Dangers , 10 Ibid ., p 29 .

of Stochastic Parrots: Can Language 8 Eyal Weizman Forensic Architecture: ,

11 Hosey op cit , .
Violence at the Threshold of Detectability
Models Be Too Big?' in FacT'21 Virtual , ,
,

Canada 2021 .
Zone Books; The MIT Press Brooklyn, ,
,

NY and Cambridge, MA, 2017 p 9 , .

3.5
9 This issue (and the example cited) is
1 George Barnett Johnston , explored at length in Brian Christian,
Assembling the Architect: The History The Alignment Problem: Machine Learning
and Theory of Professional Practice , and Human Values 1st edn, W.W. ,

Bloomsbury Visual Arts London 2020 , , , Norton & Company New York 2020 , , ,

p 117 quoting Frederick Squires and

, Chapter 3., pp 82 ff.
Rockwell Kent Architec-Tonics: The ,
10 Garrett Jacobs and Deanna Van
Tales of Tom Thumtack, Architect The ,
'
Buren Designing to Divest', Architectural
,

William T. Comstock Company New ,

Record 5 October 2020 p 37 .
, ,

York 1914
, .

2 Ibid ., p 119 .

3 An important distinction between,

say, AIA and RIBA guidance on
 INDEX

Page numbers in bold indicate figures and tables. business models see value propositions
business risk 85 91 ,
Airbnb 153
Alberti, Leon Battista 4 95 117 118 120 121 , , , , , CAD see computer-aided design (CAD)
algorithmic systems 17 18 18 39 40 -

, ,
-
canonical economic model 73 74 81 151 -

, ,

ALM (Autor-Levy-Murnane) hypothesis 46 47 50 151 154 -

, , , carbon assessment tool 110 111 112 113 161 162 -

,
-
, ,

Amara, Roy 33 34 -

Carpo, Mario 3 4 12 28 117 120

, , , , ,

American Institute of Architects (AIA) 84 104 110 137 , , , certification/registration 24 25 26 104 106 110 108 , , , ,
-

'Basic Services' 49 50 51 52 53 57 94 95 94 95 ,
-
,
-
, ,
-

,
-
Chaillou, Stanislas 106 107 ,

analogue technologies 3 10 11 , , code-checking software 40 41 41 -

analysis and simulations 8 9 10 36 37 77 79 78 , , , , ,

-

, cognitive systems 17 18 18 23 67 91 117 -

, , , , ,

Anderson, Lorin 47 48 , collaboration 8 9 10 10

,
-

Anderson, Stanford 18 see also interoperability of data

Architec-tonics: The Tales of Tom Thumtack, Architect 156 157 , common data environment (CDE) platforms 35 35 39 , ,

Architects Registration Board (ARB) 24 25 26 85 104 , , , , compensation and value 72 81 155 162 160 171 -

,
-

, ,

Architectural Registration Exam (ARE) 24 25 26 , , canonical model 73 74 81 151 -

, ,

architecture curricula 23 26 26 -

, efficiency and effectiveness 74 76 75 76 -

, ,

Aristotle 27 productivity 76 79 78 -

artificial general intelligence (AGI) 19 103 150 , , value of data 79 81 80 -

artificial intelligence/machine learning 7 8 11 12 13 20 -

, , ,
-

computer-aided design (CAD) 4 5 6 9 10 11 12 11 36 74 , , , , ,

-

, , , ,

capabilities of 19 20 19 47 -49 48 57 77 79 -

, , , , ,
-
127 149 150
,
-

cognitive systems 17 18 18 23 67 91 117 -

, , , , , construction automation 9 41 -42 43 120 145 146 , , , ,

-

construction automation 9 41 42 43 120 145 146 ,

-

, , ,
-
construction cost projection 143 145 143 146 147 -

, ,
-

defined 3 construction progress evaluation 55 137 138 137 138 139 ,

-

, , , ,

education and training 105 106 109 110 -

,
-
146 147
,

future AI-based tools 110 111 112 113 117 118 119 -

,
-
,
-

, construction schedule 145

generative design 7 9 12 106 107 122 123 124 125 165 , , , , , ,
-

, , continuing professional development (CPD) 104 104 110 , ,

history of 14 17 14 16 -

, , control and coordination information 38 38 ,

professional knowledge and 27 31 28 29 30 46 47 50 -

, , , ,
-

, cost projections 143 145 143 146 147 -

, ,
-

project delivery and 66 68 66 -

, counterfactual reasoning 96 97 ,

responsibility and 82 91 89 134 -

, , Covid-19 pandemic 36 99 100 101 153 , ,

-
,

strategies to guide future of 169 171 -

Crawford, Kate 134 135 ,

task automation xi 40 41 41 45 57 52 53 54 55 56 74 ,
-
, ,
-

,
-
,
-
, , , creative destruction 73 74 81 -

136 147 142 143

-

, , creativity, defined 168

taxonomy of use 9 10 ,

data 124 126 135 -

technology types 17 18 18 -

,
,

bias in 134 135

see also data; process transformation ,

asset operation information 38 38 ,

digital building information categories 36 39 37 38 -

, ,

association 96 97 ,
interoperability 127 133 129 130 132 -

, , ,

Autodesk 8 8 18 38 57 86 88 127 132 164 165 -

value of 79 81 80 -

,
, , , , , , , , ,

automated process tools 35 35 39 40 44 , ,

-

,
Deamer, Peggy 153
automation deep learning see empiricist/deep learning systems
demand for professionals 92 99 97 150 154 152 - -

construction 9 41 42 43 120 145 146 ,

-

, , ,
- , , ,

task Xi 40 41 41 45 57 52 53 54 55 56 74 136 147

,
-

, ,
-

,
-
,
-
, , ,
-

,
Dennett, Daniel C. 98 99 ,

142 143,
design, generative 7 9 12 106 107 122 123 124 125 165, , , , , ,
-

, ,

autonomous building function 42 design intent 61 64 62 63-

, ,

autonomous process tools 35 35 40 42 41 43 44 , ,

-

, , ,
design labour 148 154 -

Autor, David 46 design objectives 116 124 -

design representation 8 9 10 11 12 36 37 121 123 , , , , , , ,

-

Bayesian networks 96 design task automation 40 41 41 -

bias in data 134 135 , Designing Justice + Designing Spaces 172

BIM see building information modelling (BIM) digital building information categories 36 39 37 38 -

, ,

Bloom's Taxonomy of learning 19 47 48 48 49 57 ,

-
, , , digital interstice 6 10 11 , ,

build strategy 145 Domingos, Pedro 19

building information modelling (BIM) 4 5 6 -

, Doxel 137 137 138 138 139 146

, , , , ,

automated processes 39 40 -

drafting room 149 149 ,

data interoperability 127 129 , duty of care 82 91 89 134

-

, ,

demand for professionals and 151 152 152 -

, DWG file format 127

design labour and 150 economics 72 81 -

design representation 9 10 11 12 , 36 37 , , , ,

canonical mode! 73 74 81 151 -

education and training 105 106 109 -

,
, ,

efficiencies 74 76 75 76 139 -
construction cost projection 143 145 143 146 147 -

, ,
-

, , ,

Level 2 standard 109 152 169 , ,

efficiency and effectiveness 74 76 75 76 -

, ,

project delivery and 62 67 68 ,

- productivity 76 79 78 -

value of data 79 81 80 -

responsibility and 85 86 ,

education and training 23 26 26 102 110 104 108

,

- -

taxonomy of use 9 10 ,
, , , ,

embodied carbon assessment tool 110 111 112 113 161 162 - -

BuildingSMART 127 128 ,

, , ,
empiricist/deep learning systems 16 17 19 ,
-

, 18 36 , Marcus, Gary 17 18 19 20 -

, ,

capabilities of 20 47 49 48 57 77 79 ,
-

, , ,
-
market conditions 144 146 147 ,
-

task automation 48 51 52 53 54 55 56 57 56 -

,
-
,
-
,
-

, MEDLINE index 29
End User License Agreements (EULA) 86 88 -

Miller, Arthur I. 168

environmental impacts 135 Minsky, Marvin 14
EskewDumezRipple 162 165 , MIT Technology Review 83
ethics 134 135 153 154
-

,
-
modern slavery assessment tool 110 111 113 -

evidence, role in design 118 121 -

Murnane, Richard 46
expert systems 15 16 16 124 -

, ,

National Council of Architectural Registrations Boards (NCARB)

explicit knowledge 46 48 49 50 ,
-

24 25 26 85 104 108 108

, , , , , ,

failures of execution 88 90 89 -

, National Institute of Standards and Technology, US 128

flexible specialisation 67 68 69 -

, National Level 2 BIM Standard, UK 109 152 169 , ,

forced labour assessment tool 110 111 113 -

, natural language processing vi , ix x 15 16 28 134 135 -

, , , , ,

Forensic Architecture 171 Negroponte, Nicholas 6 14 14 15 18 145 , , , , ,

neural networks 14 15 17 16 18 96 134 ,

-

, , , ,

Gebru, Timnit 135 154 , numeracy 165

generative design 7 9 12 106 107 122 123 124 125 165 , , , , , ,
-

, ,

gig economy 152 153 ,

objectives of design 116 124 -

Gilman, Michele 83 painting robot 41 -42 43 145 146 , ,

-

glue code 31 Pearl, Judea 96 98 97 -

Google 135 154 ,

Pelli, César 123
Google Translate ix , 16 perceptive tasks 49 52 53 ,
-

GPT-3 technology ix x 134 135

perceptrons 14 15
-

, ,
,

Greaves, Mark x 19 20 19 47 77,

-

, , ,
Picon, Antoine 34
Grenfell Tower fire 88 89 98 99 , ,
-

PictoBot 41- 42 43 145 146 ,

-

Gropius, Walter 14 Plan of Work 49 50 51 52 53 57 61 94 95 94 95 ,

-
,
-
, , ,
-

,
-

Gudea 3 3 , Porter House project, New York 159 161 ,

Health and Safety Executive (HSE), UK 93 98 111 , ,

practical reasoning 27
health, safety and welfare 93 98 99 100 101 164 166 167 ,
-

,
-
, ,
- precarious workers 153 154 -

heuristics x 18 23, ,
prediction 9 10 77 79 78 121
, ,
-

, ,

Higgin, Gurth 61 62 63 , ,
procedural task automation 50 51 52 53 54 55 56 57 -

,
-
,
-
,
-

human cognition models 15 17 141 145 142 143

-

, ,
,

process transformation 32 44 35
-

image indexing and generation 28 28 , automated process tools 35 35 39 40 44 , ,

-

implicit knowledge 46 47 50 -

, autonomous process tools 35 35 40 42 41 43 44 , ,

-

, , ,

industry Foundation Classes (IFCs) 127 common data environment (CDE) 35 35 39 , ,

inferential reasoning 96 digital building information categories 36 39 37 38 -

, ,

integrated project delivery (IPD) 60 production task automation 55 141 ,

-
145 142 143 , ,

integrative tasks 49 52 53 ,
-
productivity 76 79 78 -

International Alliance for Interoperability 127 professional knowledge 21 31 -

internet of things (IoT) 38 architecture curricula 23 26 26 -

interoperability of data 127 133 129 130 132 -

, , , artificial intelligence and 27 31 28 29 30 46 47 50 -

, , , ,
-

intervention 96 97 , professional liability 85

professional relationships 64 67 65 66 88 89 89 140 141 -

, , ,
-

, ,
-

Jacobs, Garrett 172 professional responsibility 82 91 89 134 -

, ,

Jessop, William Neil 61 62 63 , ,

professionals 22 23 46 -

Johnston, George Barnett 156 demand for 92 99 97 150 154 152 -

, ,
-

Kieran Timberlake 161 162 ,

duty of care 82 91 89 134 -

, ,

see also education and training; professional knowledge

knowledge, professional see professional knowledge
Krathwohl, David 47 48 ,
project definition 94 94 95 98 , , ,

project delivery 58 69 59 -

labour equity 153 154 -

design intent 61 64 62 63 -

, ,

labour of design 148 154 -

means and methods 67 68 69 -

Ladder of Causation 96 98 97 -

,
professional relationships 64 67 , 65 66 , 88 89 , 89 -

,
-

Latrobe, Benjamin 156 165 , 140 141 -

Leach, Neil 34 project uncertainty 163 164 163 164 -

, ,

Levy, Frank 46 projectivity 95 98 97 -

Lexus/Nexus system 29 29 ,
public policy 84 90 91 93 94 ,
-

,
-

LiDAR (Light Detection and Ranging) 6 7 9 30 37 118 , , , , ,

public-private partnerships (PPP) 60
McCarthy, John 14 quantity surveying 62 63 144 , ,

machine capabilities 19 20 19 47 49 48 57 77 79 -

, ,
-

, , ,
-

machine learning see artificial intelligence/machine learning Rabeneck, Andrew 67 68 -

realisation 8 9 10
MacLeamy Curve 139 139 158 , ,
, ,

MacLeamy, Patrick 139 140 141 , ,

reality capture information 36 37 ,

registration/certification 24 25 26 , , , 104 106 110 108 ,

-

,
 Index

relationships, professional 64 67 65 66 88 89 89 140 141

-

, , ,
-

, ,
-
IMAGE CREDITS
representation 8 9 10 11 , , , , 12 36 37 121 123
, , ,
-

82 91 89 134 0.1: Makehaven Inc.

responsibility, professional -

, ,

RIBA 84 104 110 137 0.2: Loz CC BY-SA 2.0, via Wikimedia Commons
, , ,
Pycock, BY-SA2.0,
Plan of Work 49 50 51 52 53 57 61 94 95 94 95 ,
-
,
-
, , ,
-

,
-

1.1.1: René-Gabriel
Rene-Gabriel Ojeda. @RMN-Grand Palais
Ojéda. Palais/Art
/ Art
risks 85 88 90 89 91
,
-

, ,

Rittel, Horst x , 23 77 ,
Resource, NY
robotics 1.1.2, 1.4.9 & 2.1.1: Pelli Clarke Pelli Architects
1.1.2,1.4.9
construction automation 9 41 42 43 120 145 146 - -

1.1.3, 3.3.1 & 3.3.2: DOXELInc.

DOXEL Inc.
, , , ,

construction progress evaluation 137 137 138 138 146 147 , , , , ,

Roitblat, Herbert 103 1.1.4, 1.4.3, 1.4.4, 1.4.5, 1.4.6, 1.4.7, 1.4.10, 1.5.6, 3.1.2 &
Rosenblatt, Frank 14 3.5.7: Autodesk Inc.
Rowe, Peter x 18 ,
1.2.1: Shunk-Kender ©J.
© J. Paul Getty Trust. Getty Research
Russell, Stuart 27 49 50 56 106 , , , ,
Institute, Los Angeles (2014.R.20)
Schank, Roger vi 96 , 1.2.2: Joseph Paul Cohen, Paul Bertin, and Vincent Frappier
1.2.2:Joseph
Schön, Donald 23 2019
Schumpeter, Joseph 73
1.3.3: Westlaw, By permission of Thomson Reuters
scopes of services 49 51 50 51 52 53 57 94 95 94 95
- - - - -
, , , , ,

scripting 7 11 12 41 43
,
-

, ,
1.3.4:
Building Ventures. 5G_wifi_©uniconlabs;
SHOP Architects 159 161 , AdvancedMaterials_advanced_©Wichai.wi; Al_artificial-
Al_artificial--
SmartVid.io 164 166 167 ,
-
intelligence_©Umeicon; Autonomy_stability_©Eucalyp;
Spacemaker 8 8 18 56 57 , , , ,
BIM_planning_©justicon; Blockchain_blockchain_©iconnut;
spanning strategies 159 160 , Com putationDesign_com puter_©xni mrodx;
ComputationDesign_computer_©xnimrodx;
specialisation, flexible 67 68 69 -

,
DigitalTwins_-
ComputerVision_vision_©xnimrodx; DigitalTwins_
Squires, Frederick 156 Drone_drone_©Freepik;
simulation_©Wichai.wi; Drone_drone_g)Freepik;
standard services contracts 84 137
structural engineering software 36 85 86 134
,

-
ElecCar_electric-car_©Freepik; GIS_address_©Cuputo;
, ,
IOT_internet-of-things_©photo3idea_studio; Mobile_
Mobile_-
supply chain conditions 144 smartphone_©Made by Made Premium; ModularPrefab_
ModularPrefab_-
supporting strategies 159 160 ,

RealityCapture_motion--
building_©Smashicons; RealityCapture_motion-
Susskind, Daniel viii xi 19 22 34 46 47 48 50 73 74 140 , , , , ,
-

, , , , , ,

151 154 ,
capture_©Freepik; Robotics_robotic-arm_©smalllikeart;
Susskind, Richard viii 22 46 , ,
TeX_lessee_©Eucalyp; VR_vr-glasses_©Pixel perfect; 3D_3d-
3D_3d--
systems performance information 38 38 ,
printing_©surang; BMS_system_©Freepik
1.4.1: Florian Schaffer,
Schäffer, CC BY-SA 4.0, via Wikimedia
Tally carbon assessment tool
111 112 113 161 162 ,
-
, ,
Commons
task automation xi 40 41 41 45 57 52 53 54 55 56 74 ,
-
, ,
-

,
-
,
-
, , ,

136 147 142 143

-

, ,
1.4.8: UpCodes
taxonomy of use 8 10 10 -

,
Li and I. Chen
1.4.11: E. Asadi, B. L.i
technologies and tools 2 12 -

2.3.2: Maayan Flarel/Judah

education and training 105 106 109 110 -

,
- Harel/Judah Pearl/Dana Mackenzie
evolution of processes and 10 12 10 11 -

, , 2.3.3: Apicella
Apicelia + Bunton Architects
history of 3 8 3 5 7 8 -

, , , ,
2.4.2: S. Chaillou
taxonomy of use 8 10 10 -

see also artificial intelligence/machine learning; building 2.4.4 & 3.5.4: Kieran Timberlake Architects
information modelling (BIM); computer-aided design 2.4.5: FRDM.co
(CAD)
3.1.3: DPR Construction
Tombesi, Paolo 67 68 69 -

3.2.1: Niraj Thurairajah and Dan Goucher

163 164 163 164
uncertainty, project -

, ,

use information 38 38 ,
3.3.7: Aaron Flargreaves
Hargreaves / Foster + Partners
use, taxonomy of 8 10 10
-

,
3.4.1: © Ezra Stoller / Esto
value propositions 34 73 155 162 160 171 , ,
-

, , 3.4.3: Business Model Toolbox

see also
compensation and value 3.5.1: Yale University Library
Van Buren, Deanna 172
veridical bias 134 SHoP Architects
3.5.3: SFloP
verticalisation strategies 159 160 ,
3.5.8: SmartVid.io

Weizman, Eyal 171

wicked problems x, 18 23 77 90 103 123 , , , , , All other figures are credited to the author .

Witt, Andrew 36
Yale School of Architecture 24 26 99 100 101 158 159 160 , , ,
-
,
-