Z-Library Paper Cites an Article That Was Never Published, Publisher Opens Review

An academic paper about Z-Library that was written with the help of AI cites an article that was never actually published. The paper appeared online on October 6 in the New Review of Academic Librarianship, a journal published by Taylor & Francis. It examines how easily material can be obtained through Z-Library, along with the copyright ethics involved.

Which citations were the problem

TorrentFreak reported that an article of its own listed in the paper’s bibliography does not exist. The paper describes it as a 2021 report, but TorrentFreak says it never published any such article.

The same issue turned up in citations to papers in other journals. Works listed as a 2019 article in the Harvard Journal of Law & Technology and an article in College & Research Libraries do not appear in the tables of contents for those issues. TorrentFreak reported that it was unable to locate at least 10 of the paper’s 44 references.

Where AI came in

The authors stated that they used Consensus AI for the literature review, the stage at which existing research is surveyed. They also said they used OpenAI’s ChatGPT to polish some of the prose. The reference list where the errors appeared is precisely what readers rely on when they want to trace a paper’s evidence themselves.

Taylor & Francis told TorrentFreak that it has referred the concerns raised to the team responsible for editorial and publishing ethics for investigation. That review will have to establish whether the references actually exist and how the incorrect citations found their way into the paper. Whether there is a link between the AI use the authors disclosed and the citation errors is another point to be clarified.

If you were planning to cite this paper

Z-Library, the subject of the paper, sits at the intersection of access to material and copyright questions. In 2022 the US Department of Justice announced that it had charged two people in connection with operating Z-Library and seized related domains. The two were arrested in Argentina at the request of the United States.

For readers who intended to draw on the paper for research or work, the errors are a practical problem. Anyone who follows the references to verify the underlying evidence will not be able to find and read an article that TorrentFreak says it never published. The remaining sources the outlet could not locate also need to be checked for whether they were ever published at all.

What remains to be seen is the outcome of the publisher’s review. The key questions are how many citations are incorrect, how they ended up in the paper, and whether they affected its arguments and conclusions. Whether this leads to a correction or a retraction is also worth watching.