A peer-reviewed academic paper published this month in New Review of Academic Librarianship, written by two librarians at Nigerian universities, contains numerous citations to articles and reports that do not appear to exist, according to TorrentFreak.
The paper, titled “Z-Library and Full-Text Access: Democratizing Knowledge or Undermining Copyright Ethics?” examines the tension between Z-Library—one of the largest pirate ebook libraries on the internet—and publishers’ efforts to shut it down. While the paper’s conclusions are not inherently controversial, its references raise significant concerns.
In a section discussing enforcement against Z-Library, the authors reference a 2021 TorrentFreak article titled “Publishers continue crackdown on shadow libraries.” TorrentFreak confirmed this article was never published. While the outlet has reported on Elsevier and Springer Nature obtaining UK blocking orders against Sci-Hub and Z-Library domain suspensions, no article with that specific title exists.
Additionally, the paper’s account of the Z-Library criminal case claims the U.S. Department of Justice “arrested individuals” in 2022, citing a press release with a title that does not exist. The actual DOJ press release states the two alleged operators were arrested in Argentina at the request of the United States.
Other problematic citations include a 2019 Harvard Journal of Law & Technology article by IP law professor Daniel Gervais that does not appear in that journal’s table of contents, and a 2019 College & Research Libraries article attributed to Carol Tenopir that was not published in that journal. In total, TorrentFreak identified at least ten of the paper’s 44 references that could not be located.
The paper’s acknowledgments section provides a potential explanation. The authors state that “Consensus AI was used in sifting through large volume of literature and extract relevant information,” and that ChatGPT was used to improve the language of sections of the paper.
AI language models are known to generate plausible-sounding but fabricated citations, often crediting real experts with papers they did not write. When TorrentFreak ran the paper’s abstract and conclusion through the AI detection tool Pangram, it reported that “100% of the text is likely AI-generated,” though such tools can produce false positives.
Publisher Taylor & Francis, which published the paper, confirmed it is investigating the matter. The lead author did not respond to follow-up questions about the mysterious references, nor did the experts whose names appear on non-existent papers.
Key facts
- A peer-reviewed paper on Z-Library in New Review of Academic Librarianship contains at least ten non-existent citations
- The paper cites a TorrentFreak article titled ‘Publishers continue crackdown on shadow libraries’ that was never published
- Authors acknowledge using Consensus AI and ChatGPT in research and writing
- AI detection tool reported the paper’s abstract and conclusion as 100% likely AI-generated
- Publisher Taylor & Francis has referred the paper to its Editorial and Publishing Ethics & Integrity teams for investigation