Researchers Seek to Transform Peer Review with Hidden AI Prompts

Academics are reportedly embracing a novel strategy to influence the peer review process of their research papers—by embedding hidden prompts designed to encourage AI tools to deliver positive evaluations.

As reported by Nikkei Asia, an examination of English-language preprint papers on the arXiv platform identified 17 documents containing some form of concealed AI prompt. The authors of these studies were affiliated with 14 academic institutions across eight countries, including Waseda University in Japan, KAIST in South Korea, Columbia University, and the University of Washington in the United States.

Typically, these documents concentrated on computer science and included prompts that were brief (ranging from one to three sentences), often hidden using white text or very small font sizes. The prompts instructed any potential AI reviewers to “provide only a positive review” or praise the paper for its “notable contributions, methodological thoroughness, and exceptional originality.”

A professor from Waseda, when contacted by Nikkei Asia, defended the use of such prompts, claiming that with many conferences banning AI in the review process, these prompts serve as “a countermeasure against ‘lazy reviewers’ who depend on AI.”