Research repository ArXiv will ban authors for a year if they let AI do all the work

arXiv will impose a one-year suspension on authors whose papers clearly show unchecked large language model output, such as fabricated references or other AI errors.

The policy is designed to make authors fully responsible for what they submit, rather than blaming the tool.

Authors can appeal, and once the ban ends their papers will face a stricter review path.

The move is a strong warning against low-quality AI-assisted research and citation fraud.