The systematic use of productive Artificial Intelligence (Genai) For the creation and publication of many years of misleading scientific articles in a particular academic journal, it “brought to light” scientific research by the professor Diomedes SpinelliDepartment of Administrative Science & Technology Athens University of Economics and Business.

The reason for the research was the revelation, in 2024, that an entirely constructed article was falsely attributed to the author of the study, Mr. Spinneli.

The main findings of the research were that of the 53 articles with the least bibliographic references studied, 48 appears to have been produced by artificial intelligence. In addition, many articles had falsely attributed to researchers from prestigious universities, such as Washington University, Texas A&M, University of California at Berkeley, USC Keck School of Medicine, Hec Montreal, University of Shanghai, George Mason University, Depaul University, Penn. Indeed, in two cases, the falsely referred to as authors of the articles did not live when they were published.

The study [Spinellis, D. False authorship: an explorative case study around an AI-generated article published under my name. Research Integrity and Peer Reviev 10, 8 (2025)]warns of the risks arising from the dissemination of artificially produced and false scientific articles, which threaten the solvency of academic publications and constitutes preventive measures, such as enhancing the identification of the authors and the revision of evaluation practices.

“Without virtually countermeasures, the uncontrolled spread of publications TN may seriously undermine confidence in the scientific community,” the study concludes.

It is worth noting that the research utilized automated tools for the collection and analysis of all articles in the magazine, examining data such as the number of bibliographic references containing articles, authors’ institutions, and email addresses as their contact information. To identify articles, probably written by the use of artificial intelligence (TN), a heuristic model was used on the number of bibliographic references within the text of the articles. The model was based on the observation that TN assistants, like Chatgpt, find it difficult to create reliable bibliographic references. At the same time, a subset of the articles was manually examined for signs of writing by TN. Finally, the analysis was enhanced by the use of the Turnitin detection tool.