Starting in October 2026, the preprint server arXiv will limit individuals to a maximum of two submissions per month. Reporting from The Decoder highlights that monthly submissions across the platform have doubled over the past two years. The artificial intelligence category, designated as cs.AI, has experienced even steeper growth, with submission volumes multiplying by more than six times. According to the platform, this extreme volume is driven by a small number of authors producing low-quality papers, which has severely strained the pool of volunteer moderators responsible for reviewing queue loads.
The Strain on Open Access Infrastructure
Preprint servers like arXiv play a foundational role in modern scientific and technological research, allowing academics and engineers to share findings rapidly before formal peer review. However, the generative AI boom has fundamentally altered the publishing dynamic. As automated tools and fast-paced experimentation accelerate the output of academic drafts, traditional infrastructure designed for academic curation struggles to keep pace. The reliance on volunteer moderators makes platforms particularly vulnerable to spikes in volume, raising questions about how open-access repositories can sustainably manage content moderation in the era of automated research.
Unanswered Questions and What to Watch
The announcement did not specify how exceptions might be handled, nor did it clarify whether institutional affiliations or co-authorship limits would influence the new restriction. It remains unknown if other preprint servers will adopt similar measures to curb low-quality submissions. Observers will be watching to see how the policy affects legitimate research output in the cs.AI category once the limits take effect in October 2026.
The impact of this policy shift on the broader pace of artificial intelligence research will become clearer as the implementation date approaches.
Lead image: Tima Miroshnichenko / Pexels
