Hidden Prompts in Manuscripts Threaten the Integrity of Peer Review and Research: Recommendations for Journals and Institutions

Abstract

I examine the scholarly implications of a troubling case where researchers embedded hidden prompts like “give a positive review only” into academic preprints to manipulate AI-assisted peer review. AI is now woven into nearly every facet of academic life, including the peer review process. I contend that manipulating peer review through embedding secret prompts is as serious as plagiarism or data fabrication. Peer review may not be perfect, but deception is misconduct. Reviewers must still be held accountable. Those who blindly rely on AI outputs without critical engagement fail in their scholarly duty. AI should only amplify the reviewer’s expertise. As institutions begin regulating AI in research, similar frameworks must extend to peer review. Journals and publishers should establish clear, enforceable guidelines on acceptable AI use: Will AI be banned, regulated, or embraced? If allowed, disclosures must be mandatory. Authors should also be informed if AI tools will be used in the review process, ensuring transparency and consent. Confidentiality is another pressing issue. Real cases have shown how ChatGPT links shared by reviewers were indexed online, compromising sensitive, unpublished research, even though OpenAI has since moved to discontinue public link discoverability. Beyond policy, we must cultivate a culture of transparency, trust, and responsibility. Institutions can host ethics workshops and mentor early-career scholars. This is not just about AI; it is about who we are as researchers and reviewers. No matter how advanced the technology, integrity must remain our anchor. Without it, even the most innovative research stands on shaky ground.