Researchers try to influence peer review by using hidden AI prompts (
Academics are using a novel approach to influence peer reviews of their research papers – adding hidden prompts to encourage AI tools to provide positive feedback. Nikkei Asia reported
that 17 papers were found to have some form of hidden AI prompt when they examined English-language preprints available on the arXiv website. The authors of the paper were affiliated with 14 institutions in eight countries including Japan’s Waseda University, South Korea’s KAIST and Columbia University and University of Washington, both in the United States.
These papers were typically computer science-related, with prompts which were short (one to three sentence) and were reportedly hidden by white text or very small fonts. They asked any potential AI reviewers “to give a positive review” or praise the article for its “impactful contribution, methodological rigor and exceptional novelty.” “
A Waseda Professor contacted by Nikkei Asia defended the use of a prompt – since many conferences prohibit the use of AI for reviewing papers, the prompt was supposed to serve as a “counter against ‘lazy’ reviewers who use AI.”
