Artificial intelligence is producing fabricated citations in biomedical research papers, according to a recent study published in The Lancet. An audit of millions of biomedical publications identified more than 4,000 citations referencing non-existent studies, raising alarms about the accuracy of medical literature relied upon for clinical decision-making.
Maxim Topaz, an associate professor at Columbia University’s School of Nursing and the study’s lead author, emphasized the potential risks posed by these fake references. “When those fake references are making it into the literature, they will end up in those guidelines, and that’s how doctors decide how to provide care for you,” Topaz explained. He warned that treatment decisions could be based on invalid evidence, potentially compromising patient outcomes.
The researchers found these fabricated citations across nearly 3,000 academic papers, and none of these errors have yet been corrected or retracted. The problem appears to be escalating, with the rate of fake citations increasing twelvefold over the past three years.
Topaz’s investigation was sparked by a personal experience using an AI tool to assist with scientific writing. The app inserted a false citation, which passed multiple rounds of peer review before being identified by an editor. He noted, “I was mortified, because I’ve been studying AI for the past 15 years, so if it can happen to me, it can happen to anyone.”
The researchers described several ways AI fabricates citations: inventing entirely fictitious studies, misattributing research to real authors, or producing convincing but false references. Since AI-generated references can appear legitimate, Topaz stressed the need for meticulous fact-checking by authors, reviewers, and editors to ensure research integrity.
This issue is not necessarily limited to biomedical research, as other academic fields employing AI tools may face similar challenges with unverifiable or false citations.
Why it matters
Fabricated citations in biomedical literature undermine the scientific foundation of clinical guidelines, which health care professionals depend on to provide safe and effective patient care. As AI-assisted writing becomes more common, vigilance is required to prevent erroneous information from influencing treatment decisions. Without corrections or retractions, these fake references continue to persist in the literature, potentially affecting public health outcomes.
Background
The growing use of AI in academic writing offers efficiency but introduces risks when the technology produces inaccurate information such as false citations. Peer review processes have not consistently detected these errors, highlighting gaps in current quality control methods. Researchers urge the development of stricter verification protocols to address the unintended consequences of AI-generated scientific content.
Sources
This article is based on reporting and publicly available information from the following source:
Read more Business stories on Goka World News.