AI-faked Arctic fox study – Canid Specialist Group

Dear Canid Colleagues,

In December 2025, an article entitled “Behavioural adaptions of the Arctic fox (Vulpes lagopus) in response to climate change” (Narbayev et al. 2024) began showing up in Google Scholar searches. Published in the Caspian Journal of Environmental Sciences (CJES), the article reported on a 3-year study that claimed to have radio-collared 60 foxes, deployed 100 camera traps, performed genetic analyses, and conducted in-person year-round observations on a variety of fox behaviours across 8,000 km² in one of Alaska’s largest and most remote wilderness regions.

Such a Herculean field effort would represent an impressive logistical feat, and potentially provide a valuable contribution to our knowledge of the species, were it not for one thing: the study was faked. The paper, which appears to have been AI-generated, summarised work that never occurred. While slightly more sophisticated than some AI products, rudimentary inspection revealed the paper to display multiple hallmarks of an AI-fake including lack of methodological or analytical detail, too-perfect results, and a long list of authors from disciplines completely unrelated to the topics at hand. An inquiry to the Alaska Dept of Fish and Game, the USFWS, and regional biologists confirmed that no such study was ever proposed, permitted, or took place.

Shortly after the paper’s discovery on Google Scholar, the matter was brought to the attention of the Arctic fox research community for discussion. The issue of AI-faked science is not new; however, this is the first instance of which we are aware of a fake project involving a species that is intensively studied by a small, close-knit research community. While a few initially voiced skepticism about devoting time or energy to tackling the issue, we soon reached a consensus that, as a small, well-connected group, we were in a unique position to rapidly detect and expose a fake within our Arctic fox research network. We also are highly motivated to combat it in order that others, who may not be as familiar with Arctic fox, do not unwittingly incorporate the falsified results into their own work.

The primary goal is to have the paper retracted as soon as possible. To that end, our group has written to CJES’ editor-in-chief bringing the issue to his attention. We have filed a fraud report with UNESCO because two of the authors list themselves as UNESCO Chairs. We are in the process of taking additional actions, such as contacting the indexing databases (Scopus, DOAJ) as well as relevant authorities at the authors’ institutions to alert them of this fraudulent data reporting and ethical/scientific misconduct.

While we concede that our actions against one publication cannot stop the widespread pollution of AI-faked science, we maintain that, by acting quickly and collectively, we can make a small difference for one small canid. We are in the process of chronicling the event timeline, actions, and responses which will be summarized in a case study once the ultimate outcome is known in the hopes that our experience may help guide future efforts.

Our purpose here is to inform the larger canid community of this incident and encourage similar vigilance and communication in recognizing and reporting fake studies, because dissemination of fake science has the potential to negatively impact genuine research and conservation efforts. By keeping ourselves and one another informed, we seek to help minimize the damage caused by AI-faked science.

We are purposely refraining from providing a formal citation of the fake Arctic fox paper in question because we do not want to create a legitimate citation of the fake work.