“It's like some weird <scp>AI</scp> ouroboros”: Artificial Intelligence Use and Avoidance in Scholarly Peer Review

ABSTRACT Generative Artificial Intelligence (GenAI) challenges, even imperils, the global system of scholarly communication. Peer review constitutes a fundamental part of this system. Given its longstanding importance in facilitating scholarly communication, the discipline of Information and Library Science (ILS) offers an especially fruitful lens through which to scrutinize the intersection of AI and peer review. This is the first empirical study to do so. It is also the first research to discuss core information practices, namely use and avoidance , not only in the context of peer review, but in the context of AI more broadly. Our survey participants addressed their personal use or avoidance of AI in peer review, their overall stance on AI use or avoidance in peer review, detecting and sanctioning illicit AI use, starting to use or continuing to avoid AI in peer review, developing an AI use policy for peer review, and what they perceived as the future (both predicted and hoped‐for) of AI in peer review. Most respondents underscored the indubitably human‐centered nature of the peer review process. They gave their imprimatur only to the most limited uses of AI, e.g. for activities such as checking grammar and style. Their avoidance of AI took root in deeply felt moral and ethical commitments as well as in more prosaic concerns about bias and quality. We discuss the implications of these findings for research and practice.