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

ArXiv, the massive preprint repository that hosts millions of scientific papers, is getting serious about AI-generated submissions. Authors who let LLMs do all the writing will now face a one-year ban from the platform.

This isn't about banning AI tools outright. Researchers can still use LLMs for editing, brainstorming, or polishing their work. The issue is fully automated papers that show zero human effort or oversight.

The move comes as ArXiv deals with a flood of low-quality submissions that appear to be churned out by language models with minimal human input. These papers clog up the review process and make it harder to find legitimate research.

For anyone building AI products or following the space, this matters because it shows how institutions are drawing lines around AI assistance versus AI replacement. The tools are meant to augment human expertise, not substitute for it entirely.

ArXiv's approach is pragmatic. Use AI to work faster and smarter, but you still need to be the one doing the actual thinking and writing. It's a pattern we're seeing across academia and professional fields as organizations figure out where AI fits.

The year-long ban is meant to be a real deterrent. It's long enough to hurt if you're trying to publish research regularly, but not so harsh that a single mistake ends your ability to share work on the platform forever.