Reported about 15 hours ago
Researchers are experimenting with hidden AI prompts in their papers to sway peer reviews in their favor. A study by Nikkei Asia identified 17 preprint papers on arXiv containing such prompts, primarily in computer science, requesting AI reviewers to provide positive feedback. These prompts, sometimes disguised in white text or tiny fonts, aim to counteract AI's influence, especially in conferences that restrict AI assistance in the review process.
Source: YAHOO