"It's also important to note that an inaccurate citation doesn't negate the paper's research. As NeurIPS told Fortune, which was first to report on this GPTZero's research, "Even if 1.1% of the papers have one or more incorrect references due to the use of LLMs, the content of the papers themselves [is] not necessarily invalidated." But having said all that, a faked citation is not a nothing, either. NeurIPS prides itself on its "rigorous scholarly publishing in machine learning and artificial intelligence," it says."
"No one can fault the peer reviewers for not catching a few AI-fabricated citations given the sheer volume involved. GPTZero is also quick to point this out. The goal of the exercise was to offer specific data on how AI slop sneaks in via "a submission tsunami" that has "strained these conferences' review pipelines to the breaking point," the startup says"
GPTZero scanned 4,841 papers accepted by NeurIPS and identified 100 hallucinated citations confirmed as fake across 51 papers. Each paper typically contains dozens of citations, so the proportion of fabricated references is very small relative to the total. NeurIPS stated that isolated incorrect references do not necessarily invalidate the research content. Peer reviewers are instructed to flag hallucinations, but heavy submission volume can overwhelm review pipelines. Citations serve as career currency for researchers, and AI-generated fabricated references can dilute their value. GPTZero framed the analysis as evidence of how AI errors seep in amid strained review processes.
Read at TechCrunch
Unable to calculate read time
Collection
[
|
...
]