ArXiv to ban authors for one year if they submit AI-generated papers without meaningful human contribution

The research repository ArXiv announced new enforcement policies targeting the use of large language models in scientific paper submissions. Authors found to have submitted papers where AI did essentially all the work will face a one-year ban from the repository. The policy update reflects growing concern in the scientific community about the integrity of academic publishing amid the proliferation of LLM-generated text. ArXiv, which hosts over 2.5 million preprints across physics, computer science, mathematics, and other fields, has been grappling with how to maintain quality standards while not stifling legitimate AI-assisted research.