Source: Reddit; arXiv policy details unverified at time of publication.
arXiv is reportedly implementing a one-year submission ban for authors whose papers contain what the platform describes as "incontrovertible evidence" of unchecked LLM-generated errors — hallucinated references and fabricated results cited as examples triggering the policy.
The enforcement targets misuse of AI writing tools rather than AI-assisted writing broadly, drawing a line at outputs that were never verified before submission. No details have emerged yet about how arXiv intends to detect violations systematically or whether an appeals process accompanies the ban.
The policy, if confirmed, would apply to one of the primary dissemination channels for ML and AI research, where LLM-assisted drafting is widespread. A one-year ban represents a material disruption to research output timelines, particularly for teams on grant cycles or competitive publication tracks.
AI teams submitting to arXiv should audit current workflows to ensure LLM-generated content — citations in particular — is verified against primary sources before submission.