Researchers hide prompts in papers, manipulating AI reviews

Researchers are inserting hidden AI prompts into scientific papers to manipulate peer review and secure positive evaluations. This practice undermines the integrity of academic publishing. Seventeen articles on the online platform Arxiv have been found to contain prompts, often concealed in small or white text, instructing AI to provide favorable reviews. These prompts request praise for the work's quality and novelty. Universities in eight countries are implicated. The issue arises amid increasing reliance on AI in research and evaluation, with some researchers attempting to counter perceived biases. This practice highlights growing concerns about academic ethics and the impact of AI on scientific integrity.