Ophthalmology Journals' Guidelines on Generative AI: A Comprehensive Analysis

Wenqiao Huang; Yating Liang; Xianghui Wei; Yi Du

doi:10.1016/j.ajo.2024.12.021

Ophthalmology Journals' Guidelines on Generative AI: A Comprehensive Analysis

Am J Ophthalmol. 2024 Dec 30:S0002-9394(24)00588-9. doi: 10.1016/j.ajo.2024.12.021. Online ahead of print.

Authors

Wenqiao Huang¹, Yating Liang², Xianghui Wei³, Yi Du⁴

Affiliations

¹ Department of Ophthalmology, The First Affiliated Hospital of Guangxi Medical University, Nanning, Guangxi Zhuang Autonomous Region, China; Department of Ophthalmology, Liuzhou Red Cross Hospital / Eye Hospital of Liuzhou City, Liuzhou, Guangxi Zhuang Autonomous Region, China.
² Department of Ophthalmology, The First Affiliated Hospital of Guangxi Medical University, Nanning, Guangxi Zhuang Autonomous Region, China.
³ Department of Ophthalmology, Liuzhou Red Cross Hospital / Eye Hospital of Liuzhou City, Liuzhou, Guangxi Zhuang Autonomous Region, China.
⁴ Department of Ophthalmology, The First Affiliated Hospital of Guangxi Medical University, Nanning, Guangxi Zhuang Autonomous Region, China. Electronic address: [email protected].

PMID: 39742908
DOI: 10.1016/j.ajo.2024.12.021

Abstract

Purpose: The integration of generative artificial intelligence (GAI) into scientific research and academic writing has generated considerable controversy. Currently, standards for using GAI in academic medicine remain undefined. This study aims to conduct a comprehensive analysis of the guidance provided for authors regarding the use of GAI in ophthalmology scientific journals.

Design: Cross-sectional bibliometric analysis.

Participants: A total of 140 ophthalmology journals listed in the Scimago Journal & Country Rankings, regardless of language or origin.

Methods: We systematically searched and screened the 140 ophthalmology journals' websites on October 19 to 20, 2024, and conducted updates on November 19 to 20, 2024.

Main outcome measures: The content of GAI guidelines from the websites of the 140 ophthalmology journals.

Results: Of the 140 journals, 96 (69%) provide explicit guidelines for authors regarding the use of GAI. Among these, nearly all journals agree on three key points: 1) 94 journals (98%) have established specific guidelines prohibiting GAI from being listed as an author. 2) 94 journals (98%) emphasize that human authors are responsible for the outputs generated by GAI tools. 3) All 96 journals require authors to disclose any use of GAI. Additionally, 20 journals (21%) specify that their guidelines pertain solely to the writing process with GAI. Furthermore, 92 journals (66%) have developed guidelines concerning GAI-generated images, with 63 journals (68%) permitting their use and 29 (32%) prohibiting them. Among those that prohibit GAI images, 27 journals (93%) allow their use under specific conditions.

Conclusions: Although there is considerable ethical consensus among ophthalmology journals regarding the use of GAI, notable variations exist in terms of permissible use and disclosure practices. Establishing standardized guidelines is essential to safeguard the originality and integrity of scientific research. Researchers must uphold high standards of academic ethics and integrity when utilizing GAI.

Keywords: ChatGPT; academic ethics; author guidelines; generative artificial intelligence.