Obscure journals put AI-like papers under real scholars’ names

A cluster of obscure academic journals has published thin papers under the names of established scholars who say they neither wrote nor authorised them, raising fresh concerns that generative AI is being used to manufacture scholarly legitimacy.

Alan Said, a computer-science professor at the University of Gothenburg, discovered a paper attributed to him after receiving a Google Scholar notification. The article, “Intelligence Systems Supporting Sustainable Global Development,” was loosely related to his work, but he said he had not written it and called the text content-free “slop.”

Inside Higher Ed reported that Said complained to the domain registrar for the American International Journal of Research and Innovation and was told the paper had been removed. It remained online, however, and its byline was changed twice, leaving three versions of the same paper credited to different authors.

Said identified two journals with similar graphics, the same editor in chief and overlapping editorial-board members: the American Journal of Advanced Multidisciplinary Research and Innovation and the American Journal For Multidisciplinary Research. All three sites named Denish Brown as editor in chief and displayed a WhatsApp number with a U.S. area code.

The American Journal of Advanced Multidisciplinary Research and Innovation carried papers attributed to Fei-Fei Li, Christopher Manning, Michael I. Jordan and Stuart Russell. Several academics listed by the journals confirmed that they had not written the articles or consented to use of their names.

Jure Leskovec told Times Higher Education that he had not written a paper attributed to him, had played no part in its submission and had not authorised the byline. Kanakaraju Manupati, a postdoctoral fellow at Virginia Commonwealth University, said he had not agreed to serve on the journals’ editorial boards and did not know the journal existed before finding his name there.

The three domains were registered in 2026, despite their websites displaying archives that purported to contain papers from earlier years. Each presented itself as peer-reviewed, open access and multidisciplinary, while advertising publication fees. Two listed a $200 charge for papers with five or more non-Indian authors; the third listed fees of $80 for papers with one to four authors and $10 for those with five or more.

Said said the journals appeared to be the same type of operation, publishing work that was “very shallowly written or generated.” He said a particular risk was that a new doctoral student could see a recognised scholar’s name on one of the sites and conclude that it was a credible place to submit research.

The episode follows our October report on AI-generated books placed on Amazon under real authors’ names. It differs from a separate problem identified by researchers who found 1,655 Zenodo records with invented authors, nonexistent journals and fabricated publication dates: here, the names at issue belong to real academics whose reputations can lend false credibility to the material.

The cases emerge as academic publishing and preprint platforms face a broader increase in low-quality, AI-assisted submissions. ArXiv introduced a two-submission monthly limit per author in October after saying a relatively small group was consuming disproportionate moderator time. It received about 40,000 submissions in September, compared with about 9,900 a decade earlier.

Research-integrity guidance from CASRAI defines paper mills as commercial operations that produce fabricated or low-effort manuscripts and sell authorship or broker their acceptance. It cautions, however, that individual textual signals, such as automated paraphrasing, do not alone prove a paper mill or fabricated results. COPE’s retraction guidance now explicitly identifies unverifiable authorship and undisclosed AI use among grounds for retraction.