ArXiv to Ban Authors for Submitting Unchecked AI-Generated Papers ArXiv, one of the world’s most important preprint servers, is escalating its fight against low‑quality, AI-generated research by introducing yearlong bans for authors who submit obviously unchecked machine-written papers.

Early concerns and groundwork

Over the past few years, arXiv has become a core distribution channel for research in fields like computer science and mathematics, even before formal peer review. As large language models proliferated, moderators began noticing what they described as “AI slop” — papers cluttered with hallucinated references and boilerplate text.

In response, arXiv had already tightened some processes, such as requiring first‑time posters to obtain an endorsement from an established author to curb low‑quality and AI‑generated submissions. The organization also started transitioning from Cornell hosting to an independent nonprofit structure, partly to raise more funding to handle these new challenges.

The new one‑year ban policy

The latest step came in mid‑May 2026, when computer science section chair Thomas Dietterich outlined a formal “one‑strike” rule. If a submission shows “incontrovertible evidence that the authors did not check the results of LLM generation,” such as hallucinated references or leftover meta‑comments addressed to or from an AI system, “we can’t trust anything in the paper.” In such cases, authors will receive “a 1‑year ban from arXiv” and, afterward, any new submissions must first be accepted at “a reputable peer‑reviewed venue.”

Dietterich emphasized that, by signing as authors, researchers take “full responsibility” for every part of a paper “irrespective of how the contents are generated,” including any “plagiarized content, biased content, errors, mistakes, incorrect references, or misleading content” copied from an LLM. Moderators must document problems, section chairs must confirm the evidence, and authors retain the right to appeal.

Balancing AI use and research integrity

The rule does not ban AI tools outright. Instead, it aims to preserve trust in arXiv’s rapidly growing corpus by penalizing only clearly unverified AI output, while allowing careful, human‑checked use of generative models. Coverage across outlets framed the move as a sharp warning: “Send the arXiv AI-generated slop, get a yearlong vacation from submissions.”

[1] The Verge — “ArXiv will ban researchers who upload papers full of AI slop”
ArXiv, a popular platform for preprint academic research, is taking a new step to attempt to reduce the volume of papers that include AI slop. If a paper has “incontrovertible evidence that the authors did not check the results of LLM generation,” such as hallucinated references or “meta-comments” left by an LLM, authors will be banned from ArXiv for a year.

[2] TechCrunch — “Research repository ArXiv will ban authors for a year if they let AI do all the work”
Thomas Dietterich wrote that “if a submission contains incontrovertible evidence that the authors did not check the results of LLM generation, this means we can’t trust anything in the paper,” and that this will result in a one-year ban plus a requirement that future submissions first be accepted at a reputable peer-reviewed venue.

[3] Ars Technica — “Send the arXiv AI-generated slop, get a yearlong vacation from submissions”
One of the site’s moderators described a new policy on social media under which authors who submit AI-generated hallucinations face a yearlong suspension from submitting to arXiv.

Continue reading https://foxvector.com