ArXiv will ban researchers for one year if papers contain unverified AI-generated content
Tags AI · OSS
ArXiv, the widely-used preprint academic research platform, announced it will ban authors for one year if their submissions contain 'incontrovertible evidence' that the authors did not check LLM-generated results, such as hallucinated references or meta-comments left by an AI. After the ban, subsequent ArXiv submissions must first be accepted at a reputable peer-reviewed venue. The policy, announced by ArXiv's computer science section chair Thomas Dietterich, is a direct response to the growing volume of AI-generated 'slop' papers flooding academic repositories.
Technical significance
ArXiv hosts over 2 million preprints and is the primary distribution channel for CS, physics, and math research. This policy sets a precedent for how academic institutions will handle AI-generated content, and could push researchers toward more rigorous verification practices — or toward alternative, less moderated platforms.