In a bid to influence artificial intelligence-powered peer review systems, a number of researchers have resorted to embedding covert instructions within their academic papers, often by using white text or minuscule fonts visible only to machines. The subterfuge, which can lead AI reviewers to issue favorable assessments, has prompted some preprint servers to consider withdrawing the affected manuscripts. The episode underscores mounting concerns in the scientific community over the integrity of publishing as AI tools become increasingly central to the peer review process, raising fresh ethical questions about technological oversight and the future of academic quality control.
Scientists Embed Hidden Instructions in Research Papers to Manipulate AI Peer Review Systems
- Trending
- Comments
- Latest
