Arxiv cracks down on unchecked AI-generated content in research papers

arXiv has tightened its rules on AI-assisted content in research papers, making authors fully responsible for checking any machine-generated text.

Papers with clear signs of unverified LLM output may face a one-year submission ban, followed by mandatory peer review.

The policy responds to growing concerns about hallucinated references, stray model text, and hidden prompts in preprints.

The change is meant to protect trust in preprints and prevent AI-written errors from spreading into the research record.