Google Scholar alerts brought another recommendation for an article on AI and journalism. The paper, entitled “AI News Anchors, Automated Journalism, and the Transformation of Journalistic Authority: A Conceptual Framework“, appeared to have been published in the Journal of AI Communication and Society, a journal that was previously unknown to us. The title suggested that it might be a relevant work to add to the library. At first glance, nothing seemed unusual.
An initial reading revealed the use of key references in this field: Carlson, 2015; Clerwall, 2014; Diakopoulos, 2015; Graefe, 2016; Guzman & Lewis, 2020; Montal & Reich, 2017; Thurman, Dörr & Kunert, 2017; and Wölker & Powell, 2021. The author of the article, however, was an unfamiliar name, Oliver Hughes, affiliated with the University of Sydney.
We searched for his academic profile, including on the website of the university’s School of Art, Communication and English, but without success. The whereabouts of Oliver Hughes, whose contact details consisted of the generic email address “oliver.hughes@academicmail.org“, therefore remained uncertain.
The investigation then turned to the journal itself. The Journal of AI Communication and Society had published only a single issue, on 6 June 2026, containing five articles. A careful reading of these texts revealed a recurring pattern: the same structure, the same theoretical foundations, the same ideas, the same titles… A few clicks later, we concluded that the articles were little more than AI-generated summaries and offered no scientific value.
The concerns became more serious when we attempted to verify the identities of the authors and members of the editorial board. None of the names listed could be found in academic databases, research profiles or other credible institutional sources. In other words, none of the article authors or journal editors actually existed. After contacting one of the universities listed as an affiliation, we received confirmation. Unsurprisingly, Tijmen Lansdaal from the University of Amsterdam noted that the differences between genuine and fictitious journals were noticeable, at least to researchers: “Thankfully, the difference with real journals is quite noticeable. Hopefully, this is also true for the general public“.
This episode illustrates how easily generative AI technologies can produce content, particularly text, that increasingly resembles academic writing. However, the use of AI as a support tool is not, in itself, a problem. The problem arises when AI is used to fabricate content presented as scientific research, or when a website poses as an academic journal by inventing authors, articles and institutional affiliations.
We share this experience as a warning to the academic community.
