ArXiv tightens grip on AI-created references with year-long bans
ArXiv will bar authors from posting for one year if a submission contains AI-generated hallucinated citations or other clear signs of unverified AI use, and authors must have work accepted at a reputable peer-reviewed venue before reposting. The policy aims to curb AI 'slop' in preprints but has sparked debate about whether bans address root causes, with supporters praising accountability and critics arguing for broader fixes and cross-platform cooperation.
