TL;DR
ArXiv announces a new policy to ban authors for one year if their submissions contain clear evidence of unchecked AI-generated errors. The move aims to improve research quality amid rising concerns over AI-generated content.
ArXiv has announced a new policy to ban authors for one year if their papers contain incontrovertible evidence of unchecked AI-generated errors or hallucinations, marking a significant step to curb low-quality AI-influenced research submissions.
The platform’s section chair of computer science, Thomas Dietterich, confirmed that authors will face a one-year ban if their submissions include clear signs of AI misuse, such as hallucinated references or meta-comments from language models that indicate unverified or fabricated content. The policy aims to address concerns about the proliferation of AI-generated papers that lack proper verification, which can undermine scientific integrity.
According to Dietterich, the ban will be imposed only in cases where there is incontrovertible evidence—such as hallucinated references or meta-comments indicating unverified content—that the authors did not check the results generated by large language models (LLMs). Authors will be able to appeal the decision, and the process involves moderation and confirmation by the section chair. Future submissions must also be accepted at reputable peer-reviewed venues before being posted on ArXiv, further ensuring quality control.
Why It Matters
This development signals a concerted effort by ArXiv to uphold research integrity in the face of increasing AI-generated content. The policy aims to deter authors from submitting papers with unchecked AI errors, which could otherwise dilute the quality of scientific literature and mislead the research community. It reflects broader concerns about the impact of AI tools on academic publishing and the need for stricter oversight.
Generative AI for Software Development: Code Generation, Error Detection, Software Testing
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background
Over the past year, ArXiv has taken steps to limit AI-related low-quality submissions, such as restricting review articles to peer-reviewed conference or journal acceptances. The rise of large language models has made it easier to generate content rapidly, but often at the expense of accuracy and reliability. The new policy targets the growing problem of AI-generated ‘slop’ that contains errors, hallucinations, or fabricated references, which can undermine trust in preprint repositories.
“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.”
— Thomas Dietterich
“Our Code of Conduct states that by signing your name as an author of a paper, each author takes full responsibility for all its contents, irrespective of how the contents were generated.”
— Thomas Dietterich
The Ultimate Guide to Plagiarism Checkers and AI Detection Tools: How to Identify Similarity, Avoid Copying, and Write with Integrity (AI for Academic Research)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What Remains Unclear
It remains unclear how frequently the policy will be enforced, what specific evidence will trigger the ban, or how many authors might be affected initially. The effectiveness of this measure in reducing AI slop on ArXiv is also yet to be seen, and the process for verifying claims of incontrovertible evidence is still being established.
STEMI 12 Lead Tool Vertical Reference Badge ID Card (1 Card)
This card features a 12 Lead EKG, associated "adjacent" leads, ST Elevation ruler, and chart of corresponding vessels.
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What’s Next
ArXiv will likely implement monitoring tools to detect AI-generated errors more systematically and may update guidelines as they gather more data on the policy’s impact. Researchers should prepare for stricter review processes, and the community will observe how the policy influences submission quality over the coming months.
Better Health with AI: Your Roadmap to Results
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What types of AI-generated errors will lead to a ban?
Incontrovertible evidence such as hallucinated references, meta-comments indicating unverified content, or fabricated data will trigger the ban.
Can authors appeal a ban?
Yes, authors can appeal the decision, and the process involves moderation and confirmation by the section chair.
Will future submissions need peer review approval?
Yes, all future submissions will require acceptance at reputable peer-reviewed venues before being posted on ArXiv.
How will ArXiv detect AI slop in submissions?
The platform is developing or may develop tools to identify signs of unchecked AI-generated errors, but specific detection methods have not been detailed.
Is this policy only for computer science papers?
The policy applies to submissions in the computer science section, but similar measures could be considered for other fields in the future.
