ArXiv Requires Peer-Reviewed CS Papers Amid AI Submission Surge

ArXiv, a leading preprint server widely used by scientists and technologists, has instituted new policies to combat an overwhelming influx of low-quality, AI-generated submissions in its Computer Science category. The platform will no longer accept review articles or position papers unless they have already passed peer review at a recognized journal or conference, a shift from its traditional open-access model that allowed moderated but not peer-reviewed content. This change responds to a surge in submissions produced by generative AI tools like ChatGPT, which have resulted in numerous papers that are essentially annotated bibliographies lacking substantial scholarly discussion, straining volunteer moderators tasked with vetting content. Studies indicate a significant rise in AI-assisted research papers, exacerbating the problem and threatening to dilute the platform's credibility and usefulness. ArXiv's move aims to preserve research quality and ensure that the content remains valuable for the academic community by requiring prior vetting to reduce the flood of fast, easy-to-write AI content. This policy update reflects a broader challenge in academic publishing to maintain standards amid rapid advancements in AI technology.