ArXiv bans authors for 1 year for submitting AI-generated slop
arxiv will ban authors for a year if they submit papers with incontrovertible evidence of unchecked ai-generated content, such as hallucinated references or meta-comments.
arxiv enforced a new one-year ban on authors who submit papers containing obviously unchecked ai-generated content, termed 'ai slop.' chair thomas dietterich stated that submissions with hallucinated references or meta-comments demonstrate a lack of authorial oversight, voiding trust in the entire paper. the penalty requires a one-year exile followed by mandatory endorsement from a reputable peer-reviewed venue for future submissions. this policy follows a flood of low-quality ai-generated submissions, leading to earlier bans on certain cs review papers and stricter endorsement rules for new submitters.