ArXiv Stops Accepting Some CS Papers Due to AI Spam
The platform said it will no longer accept computer science review or policy papers without proof of external peer review.
A Flood of AI Content
ArXiv, the scientific preprint repository critical to AI development, has stopped accepting two specific types of computer science (CS) publications: review articles and policy articles. This decision was made in response to a "sharp increase" in automatically generated articles that "make no real scientific contribution".
The platform, operated by Cornell University, said its volunteer moderators were being "inundated" by these submissions. ArXiv stated it was receiving "hundreds of such publications each month," which were "fast and easy to write" using generative AI tools.
Many of these low-quality, AI-generated papers "resemble annotated bibliographies that merely …
Archive Access
This article is older than 24 hours. Create a free account to access our 7-day archive.