Peer Review Week is 15–19 September — an opportunity to thank the dedicated volunteers so critical to scientific publishing, and to explore key issues in the field. Visit our Peer Review Week page to help celebrate all our reviewers and see other blog posts on this topic!
The use of AI in peer review isn’t just a technical question, it’s an ethical one. As part of Peer Review Week, this final post in our series turns to the principles and policies that should shape how AI is used in scholarly publishing. We asked: What ethical guidelines should govern the use of AI in peer review? How have/would you use(d) AI in your peer review work? And are AMS’s current guidelines sufficient, or is there a need for more?
“This tool is around, whether we like it or not, and some people will choose to use it.”
—Fraser Dennis Murray King
AMS Editor’s Award recipients offered candid thoughts on these questions, acknowledging both the promise and the potential pitfalls of AI. Their perspectives speak to the importance of transparency, accountability, and continued dialogue. As AMS policies continue to evolve, these voices remind us that ethical clarity must accompany technological innovation, especially in a process as consequential as peer review.
Daniel Rothenberg, Brightband:
“We should keep in mind that AI tools could play many useful roles in this sort of activity, and find ways to encourage this use. But as a reviewer, I’m making a personal commitment to supporting scholarship; ethical guidelines should ensure that reviewers are standing up to this commitment. It’s also vital that we protect the intellectual property of authors — something that can be difficult to ensure, depending on how AI is used (e.g., inadvertently submitting manuscript contents to a tool which caches it for future model training).”
Gert-Jan Duine, University of California, Santa Barbara:
“I have not used AI myself for peer review. However, I can imagine that critical review skills would get underdeveloped when AI is used to perform reviews, especially when used in the early career stage. This is because when you perform the review yourself, you would in turn automatically develop review skills.”
Fraser Dennis Murray King, NASA Goddard Space Flight Center:
“In terms of ethics, I do not feel use of AI should be outright banned, but it should be disclosed if used. This tool is around, whether we like it or not, and some people will choose to use it. Since detection is not reliable, we should make clear that if it is being used, it should be clear as to where and how it was applied in the process.”
Osamu Miyawaki, Union College:
“Ethical guidelines for AI in peer review should distinguish between mechanical assistance and evaluative judgment. My opinion is that reviewers should be allowed to use AI for mechanical tasks (checking internal consistency, verifying citations, validating code) provided they maintain full responsibility for accuracy and appropriateness of such use. For example, uploading the manuscript to an online AI model is not appropriate as it violates confidentiality, whereas using a local, offline AI model will not. Judgments about novelty, significance, and publication merit should remain exclusively human. Like authors who remain responsible for AI-generated content, reviewers are responsible for independently verifying all AI outputs and retain responsibility for their scientific judgments. The current AMS guidelines appropriately prioritize confidentiality and recognize AI’s inability to provide expert evaluation. However, I think the blanket prohibition on AI use by reviewers (except for language/clarity improvements) is restrictive, especially as offline AI becomes more widespread in the future with improvements in computing hardware on devices individuals own like phones and laptops.”
AMS has adopted guidelines to inform the use of AI tools for authors (see #12) and reviewers (see #13). The AMS Publications Commission welcomes your thoughts on the policies. Email [email protected].
