MIT Connection Science

Thrilled to share insights from Monday's EY x MIT CAE Luminary series! A day rich in innovation, AI governance, and responsible AI discussions. Challenges CAEs face was explored, alongside the evolving regulatory landscape. The event concluded with a scenic happy hour by the Charles River, where conversations on emerging tech, IA talent, and the IA agenda continued. Thank you to our featured speakers! Alex 'Sandy' Pentland Hossein Rahnama Ra'ad Siraj Alison W. Tobin South Robert Mahari Thank you to my fellow leaders and the entire EY team who make this magic happen! Jeffrey Saviano Clayton Stephens Michael Edwards Fernando Morera Kristi Kennedy Megan Duggan Daniel Prior William Thomas, CPA Mudit Gupta, CPA, CGMA Ernesto Guardia Ryan Cooksey Yiming Chang Agustina Silvera Massachusetts Institute of Technology MIT Connection Science #EYxMITLuminary #Innovation #responsibleAI #internalaudit

🚨Publication alert🚨 Today we are releasing the next version of the Foundation Model Transparency Index! In addition to our paper, we are publishing transparency reports for 14 top AI companies. A bit of background: in October 2023, my team at Stanford Institute for Human-Centered Artificial Intelligence (HAI) partnered with researchers at the Princeton Center for Information Technology Policy and MIT Media Lab to assess the transparency of 10 major foundation model developers like OpenAI and Meta. We set out to score the transparency of these companies based on publicly available information on their most prominent AI models. The rubric was 100 transparency indicators covering issues related to the data, labor, and compute used to build foundation models, the risks and capabilities of these models, and their downstream impact. We found that there is a fundamental lack of transparency among companies that build foundation models. Companies scored just 37 points out of 100 on average, with the top score barely eclipsing 50. They disclosed little if any information about the data used to build their AI models or their real-world impact. 📜 We have just published a follow up study six months on. For this version of the Foundation Model Transparency Index, we reached out to companies and asked that they prepare reports to disclose their practices in relation to our 100 transparency indicators. Fourteen companies agreed, and today we are making these transparency reports publicly available along with a paper describing our findings on how transparency has changed in the foundation model ecosystem. These reports include information the developers had not made public until we began this follow-up study; on average, each developer shared information related to 16 indicators that had not previously been public. 📊 We find that while there is substantial room for improvement, transparency has increased in some areas. The average score rose from 37 to 58 out of 100, with improved transparency related to risks and how companies enforce their policies to prevent such risks. Companies scored points on just 17% of the compute-related indicators in October 2023, whereas they now score 51%, reflecting the fact that several additional companies now disclose the amount of compute, hardware, and energy required to build their flagship foundation model. However, there is a systemic lack of transparency in some areas of the AI supply chain. Companies lack transparency on issues relating to data, like the copyright status of and presence of PII in the data used to build foundation models. Companies also do not share information about the nature of their models’ downstream impact, such as the number of users and how their models are used. Let me know what you think of the paper! Thanks to my coauthors Rishi Bommasani, Sayash Kapoor, Shayne Longpre, Betty Xiong, Nestor Maslej, and Percy Liang

Looking forward to speaking at the ACM Collective Intelligence Conference with D'Amore-McKim School of Business at Northeastern University at this week. Hope to see you there! https://lnkd.in/gSyXsEp5

ICYMI: Ramesh Raskar, Hari Balakrishnan, and I had a great conversation about how #AI will shape society over the next 20 years during the Imagination in Action Summit in April. We discussed everything from impacts in #education, #governance, #transportation, #cryptocurrency & more. https://lnkd.in/gfmHhw77

Danielle Allen of Harvard, Nathaniel Persily of Stanford, and I had a great and necessary conversation last month about AI and the 2024 Elections; convened by the Harvard Ash Center for Democratic Governance and Innovation and co-facilitated by MIT Connection Science Ph.D. student Robert Mahari. https://lnkd.in/g9SEgm63

Proud and honored to join Stanford Institute for Human-Centered Artificial Intelligence (HAI)’s Digital Economy Lab as a fellow and the faculty lead for research on Digital Platforms and Society. The collaborations and conversations over the past few years with the Lab have been so powerful and fruitful. Looking forward to the continued work we will do together for a better future.

🚀 Exciting News for Researchers and Data Enthusiasts! 📊 We released a first-of-its-kind dataset containing 1.8 million purchases from over 5,000 U.S. Amazon.com consumers, spanning 5 years, published in 𝐍𝐚𝐭𝐮𝐫𝐞 𝐒𝐜𝐢𝐞𝐧𝐭𝐢𝐟𝐢𝐜 𝐃𝐚𝐭𝐚. With Alex Berke, Daniel Calacci, Takahiro Yabe, Kent Larson, and Alex 'Sandy' Pentland. This rich dataset represents a step towards democratizing access to the kind of data big companies routinely collect and use. 🔍 What's in the Dataset? • Order date, product code, title, price, quantity, and shipping address state. • Linked survey data detailing demographics, lifestyle, and health of participants. 💡 Why It Matters: Government surveys are the traditional way of collecting consumer spending data to help public agencies and researchers generate economic insights. But purchase data can also offer a rich set of insights, which companies collecting large datasets already have access to. ✅ Dataset Validation: We've validated this dataset with a high correlation to Amazon sales data (Pearson r = 0.978, p < 0.001). Our analysis also highlights specific product categories, showing expected seasonal trends and strong correlations with other public datasets. 🤝 Ethical Data Collection: • The data were carefully crowdsourced through an online survey, with full informed consent from all participants. • This dataset not only offers invaluable insights for academic research but also serves as a resource for understanding consumer behavior more broadly. • Our methodology can be extended to allow others to ethically collect consumer data. Read the paper: https://lnkd.in/e5Jr_DyD