AI-Assisted Paper on Z-Library Cites a TorrentFreak Article That Doesn't Exist
An academic paper about Z-Library, openly admitting AI assistance, was found to contain numerous fabricated citations, including a non-existent TorrentFreak article. This story highlights the perilous and often-discussed issue of AI 'hallucination' making its way into peer-reviewed research. It serves as a stark warning about the current limitations of AI tools and the critical need for rigorous human verification in scholarly publishing.
The Lowdown
A recently published academic paper, co-authored by two Nigerian librarians and focusing on the controversial Z-Library, has been discovered to include multiple entirely fictitious citations. The article, which explores the tension between Z-Library's role in democratizing knowledge and its impact on publishers, became problematic when its references were scrutinized.
- TorrentFreak, a site mentioned in one of the paper's citations, could not find any article matching the title and claims attributed to it.
- Further investigation revealed at least ten unlocatable references throughout the paper, many of which were 'plausible but inaccurate' — citing real publications or authors in incorrect or non-existent contexts.
- The paper's acknowledgments section explicitly states the use of "Consensus AI" for literature review and ChatGPT for language refinement.
- AI detection tools also flagged the paper's abstract and conclusion as "100% likely AI-generated."
- The publisher, Taylor & Francis, has initiated an investigation into the matter after being contacted by TorrentFreak.
This incident provides a compelling real-world example of AI's tendency to 'hallucinate' information, raising serious questions about the integrity of AI-assisted academic research and the current state of the peer-review process.