arXiv limits researchers' submissions at two papers per month amid AI-powered paper flood

Starting October 1st, 2026, the world's largest scientific preprint repository arXiv has imposed its strictest submission limits to date, capping each submitter at two papers per calendar month with no more than three papers simultaneously under moderation. The policy comes as a direct response to an explosive surge in submissions, September 2026 saw a record 40,363 papers, doubling the roughly 20,569 submissions recorded just two years prior. According to arXiv's official announcement, this represents a watershed moment for scholarly publishing, as AI tools have dramatically increased the speed at which papers can be produced, overwhelming volunteer moderators who processed nearly 9,000 support tickets last month alone.

This measure builds upon a May 2026 crackdown that instituted one-year bans for authors submitting papers containing “incontrovertible evidence” of unchecked AI generation, such as hallucinated citations or residual LLM instructions embedded in the text. arXiv moderators reported a marked increase in what they describe as “thin papers of narrow scope”, so-called “salami papers” where single works are fragmented into multiple smaller submissions, and dense AI-written manuscripts that fail to advance their fields. The rate limit policy is framed as a stopgap while arXiv develops improved moderation tools and determines long-term best practices for an AI-accelerated research landscape.

The move has drawn mixed reactions from the scientific community. While some researchers view it as necessary protection for arXiv's integrity following its July 2026 transition to independent nonprofit status, critics like AI researcher Nathan Lambert caution that blanket caps may inadvertently choke legitimate research dissemination. The restrictions apply only to the individual who clicks submit, meaning co-authors remain unaffected, a technical limitation arXiv acknowledges. As the organization stated, its goal is to preserve the platform as “open, fast, and free” while maintaining scientifically rigorous standards across its growing corpus.