'AI Slop' Is Flooding Science Publishing, And One Major Site Is Fighting Back

arXiv will ban authors for up to a year if their papers contain unchecked AI-generated errors, with further restrictions possible.

The move responds to a surge in AI-written submissions, including papers riddled with hallucinated citations and other low-quality mistakes.

The policy has sparked debate over fairness and effectiveness, especially in collaborative research where responsibility is harder to assign.

More broadly, the case shows how AI is straining scholarly publishing and raising questions about quality control in preprints and peer review.