arXiv Is Rate Limiting Submissions Because It Can’t Keep up With AI Slop
404 Media Matthew Gault
arXiv is limiting how many papers people can submit because AI spam is burying it. Researchers get two a month now, and the real casualty is the volunteer review queue.
Based on reporting by 404 Media, Matthew Gault — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
arXiv is putting a harder cap on submissions after getting buried under what it sees as AI-written junk. The open-access preprint site now lets a researcher submit only two papers per calendar month, with no more than three active submissions at once. That is a sharp change for a system built on speed and openness, but arXiv says the old setup was letting a relatively small group of authors eat up too much moderator time.
The pressure has been building fast. arXiv says it received 9,869 submissions in September 2016, 20,569 in September 2024, and 40,363 in September 2026. It says submissions have doubled over the last two years and have sextupled in computer science. That flood is not just more scholarship. arXiv says AI is helping people parse data, write code, shape experiments, and in some cases write papers outright.
The site’s own policy still allows AI as a tool, but only if authors disclose it and the work still meets arXiv’s standards for scholarly interest and field advancement. The problem, arXiv says, is that many new submissions do not. It points to thin papers with narrow scope, salami-sliced studies broken into multiple articles, and a growing pile of dense, AI-written papers.
This rate limit is also not coming out of nowhere. arXiv already had a submission policy, but the decision was left to individual moderators. Now the platform wants to spread that burden more evenly across its volunteer team, which it says is being overwhelmed. The change follows a string of tighter moves: less than a year ago arXiv stopped taking review articles and position papers in computer science, in January it started requiring endorsement from an existing author for new authors, and in May it said researchers who submit AI slop can be banned for a year.
My take — AI-written commentary, not fact-checked reporting
This is what happens when a scholarly commons gets treated like an infinite landfill. arXiv is doing the boring but necessary thing: adding friction where spam found the loophole. If a system needs volunteers to play whack-a-mole with machine-written junk, the system is already paying the bill.
Read more about this at: 404 Media