Sign up to see the future, today<\/p>\n
Can\u2019t-miss innovations from the bleeding edge of science and tech<\/p>\n
Even in 2026, there are still plenty of researchers who refuse to use AI to publish their research papers. Others do use the tech for tasks like sourcing journal articles for references, editing copy, or formatting citations \u2014 but they face pressure to verify every claim, since AI has a baked-in risk of contaminating their work with hallucinations.<\/p>\n
A vocal minority of academics, however, argue they should be able to use AI to write original research while remaining immune from any hallucinated claims or data that make their way into the final product.<\/p>\n
Last week, the open-source research repository arXiv announced that it was banning scholarly authors from the platform for up to a year if \u201challucinated references\u201d are found in their work. The rationale behind this should be obvious enough for any self-respecting academic: as arXiv computer science chair Thomas Dietterich wrote<\/a> in his announcement, \u201cif a submission contains incontrovertible evidence that the authors did not check the results of LLM generation, this means we can\u2019t trust anything in the paper.\u201d<\/p>\n Attention @arxiv<\/a> authors: 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. 1\/<\/p>\n \u2014 Thomas G. Dietterich (@tdietterich) May 14, 2026<\/a><\/p>\n As TechCrunch observed<\/a>, arXiv isn\u2019t banning AI altogether, but simply clarifying that the author is ultimately responsible for any work that goes out under their name. Makes sense, right?<\/p>\n Apparently not. After Dietterich\u2019s announcement of X-formerly-Twitter, numerous researchers immediately went on the offensive, trashing the platform for its decision.<\/p>\n \u201cSo this means you expect every author to check every citation and make sure that every citation is real and accurate?\u201d economics professor at Smith College James Miller replied<\/a> in shock. \u201cWhat if it\u2019s beyond the ability of one of the authors to verify one of the citations because that citation is in a language he doesn\u2019t know or concerns technical material he doesn\u2019t understand but another author on the paper does?\u201d <\/p>\n \u201cThis is way too strict. Errors can slip in when using any tools. We aren\u2019t perfect,\u201d said Luca Ambrogioni, assistant professor in AI at the Donders Institute for brain, cognition and behaviour. \u201cHaving a prompt left in is a mistake, it\u2019s sloppy but giving permanent answer a one time sloppiness is absurd.\u201d<\/p>\n Ambrogioni, who appears to argue that getting reprimanded via arXiv\u2019s policy on hallucinated citations will amount to a de facto \u201clifetime ban\u201d from publishing, continued: \u201cwe are not taking just about false citations (more serious), but also more harmless copy pasting editing mistakes. Papers are long, the likelihood of an incorrect copy past in the supplementary isn\u2019t zero even in a otherwise good quality work.\u201d<\/p>\n Neal Amin, a former neuroscientist and Stanford medical clinician turned biotech startup founder<\/a>, wrote on X<\/a> that \u201cthis is what overreaction looks like and how gatekeeping starts.\u201d<\/p>\n More on AI hallucinations: Doctors\u2019 AI Systems Are Hallucinating Nonexistent Medical Issues During Appointments With Patients<\/a><\/p>\n \t