Researchers use hidden AI prompts to sway peer review

Researchers are secretly embedding AI prompts within their research papers to encourage positive peer reviews. This tactic aims to influence the evaluation process. Hidden prompts were discovered in 17 preprint papers on arXiv, originating from 14 institutions across eight countries. These prompts, often concealed, instructed AI tools to provide favorable assessments, praising the papers' qualities. The practice, primarily in computer science, involves brief prompts using white text or small fonts. One professor defended the prompts as a countermeasure against AI-using reviewers.