Per reporting by The Subsequent Net and Ars Technica, the preprint server arXiv will ban authors for one 12 months if moderators discover “incontrovertible proof” {that a} submitted paper incorporates unvetted large-language-model output, a coverage described in a social-media thread by Thomas Dietterich, chair of arXiv’s pc science part. Sources report the penalty additionally triggers a requirement that subsequent submissions be accepted by a peer-reviewed journal earlier than arXiv will host them (Ars Technica; The Subsequent Net). The enforcement targets apparent AI slop akin to hallucinated references, leftover chatbot directions, or fabricated tables (The Subsequent Net). Reporting and commentary notice debate over whether or not this response is suitable and whether or not detection and moderation will scale (The Dialog; TechBuzz).