The Danish Committee on Research Misconduct has concluded that University of Copenhagen professor Hans Bonde used artificial intelligence to fabricate text passages and sources in a research article. The ruling, made in May 2025, has now become public.
Bonde, a sports historian at the University of Copenhagen, was found to have acted scientifically dishonestly by Nævnet for Videnskabelig Uredelighed. DR reports that the newspaper Kristeligt Dagblad obtained access to the decision through a freedom of information request.
The committee found it undisputed that the introduction to the article contained ten instances of fictitious text with accompanying fictitious source references. According to the decision as reported by Kristeligt Dagblad, the committee found that Bonde constructed the passages and references through at least gross negligence in his use of AI. The finding was unanimous.
First AI case of its kind
It is the first time the committee has reached a finding of scientific dishonesty in a case involving artificial intelligence. TV 2 also notes the precedent character of the ruling. Politiken reports that the decision from May 2025 only became publicly linked to Bonde through Kristeligt Dagblad’s access to the documents.
Under Danish rules, scientific dishonesty covers fabrication, falsification and plagiarism committed intentionally or through gross negligence in the planning, conduct or reporting of research. The committee’s published 2025 case overview describes a decision concerning an article with ten AI-generated fictitious references in its introduction. In that anonymized decision, the undisclosed creation and use of fictitious references was classified as fabrication.
Bonde acknowledges that he made mistakes, but disputes the conclusion that his conduct amounts to dishonesty. He describes the passages and sources as obviously fictional and absurd AI hallucinations, and states that his error could not cause harm because the fiction was visible. He has stated that this was his first experience using artificial intelligence in a research context.
Bonde filed a complaint against the decision. The complaint was rejected because it contained no new information that could change the outcome, according to DR and TV 2.
The University of Copenhagen’s leadership has declined to say whether the ruling had consequences for the professor, citing that it is a personnel matter. Bonde retired shortly before the decision was made, DR reports. Politiken and TV 2 report the same sequence of events.
The full decision published by the committee, case number 2024-07, does not name the researcher. The identification of Bonde rests on Kristeligt Dagblad’s access-to-documents request and subsequent Danish media reporting. The debate over AI use in Danish research therefore centers on interpretation, with the committee treating the conduct as fabrication and Bonde rejecting that label.
