A popular academic journal is coming down hard on AI-generated submissions

We’re still in the early stages of the AI revolution, but there’s already plenty of evidence that it won’t be purely a blessing. Generative AI has made writing exponentially faster, if not necessarily better, and the result has been a massive increase in submissions of novels, newspaper pieces, and even academic journals, with one publication even warning of a coming “swamp of slop.” 

Now, however, the journals are fighting back. ArXiv, one of the largest open-access repositories of preprint academic research, is issuing a one-year ban on all authors who submit “obviously AI-generated work,” according to 404media. Moreover, if the offending author wishes to return to the good graces of ArXiv, they will have to first submit to a “reputable peer-reviewed review venue,” according to Thomas Dietterich, chair of the publication’s computer science division.

He recently took to X to not only clarify the new rules but also place the onus on authors to use LLMs responsibly: “If generative AI tools generate inappropriate language, plagiarized content, biased content, errors, mistakes, incorrect references, or misleading content, and that output is included in scientific works, it is the responsibility of the author(s). We have recently clarified our penalties for this. 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.”  

Faulty, misleading references, plagiarism, and invented citations are not the only issues with AI; however, there are others. In November of 2025, ArXiv was forced to shut down its entire computer science review section due to the overwhelming volume of AI-generated submissions, most of which did not even introduce new research results, according to a press statement. 

A funny, counterintuitive consequence of AI-enabled hyperefficiency is the evaluation bottleneck. If, in any given month, there are 100 academic papers submitted for review, it’s not too hard to find and publish the best work, but if there are a thousand submissions, even the best-funded journals can’t keep up. 

Expect the backlash to grow even fiercer as the power of AI increases and the costs of using it decrease.