"A research team based in China used the Claude 2.0 large language model (LLM), created by Anthropic, an AI company in San Francisco, California, to generate peer-review reports and other types of documentation for 20 published cancer-biology papers from the journal eLife. The journal's publisher makes papers freely available online as 'reviewed preprints', and publishes them alongside their referee reports and the original unedited manuscripts."
"The authors fed the original versions into Claude and prompted it to generate referee reports. The team then compared the AI-generated reports with the genuine ones published by eLife. The AI-written reviews "looked professional, but had no specific, deep feedback", says Lingxuan Zhu, an oncologist at the Southern Medical University in Lianyungang, China, and a co-author of the study. "This made us realize that there was a serious problem.""
"The study found that Claude could write plausible citation requests (suggesting papers that authors could add to their reference lists) and convincing rejection recommendations (made when reviewers think a journal should reject a submitted paper). The latter capability raises the risk of journals rejecting good papers, says Zhu. "An editor cannot be an expert in everything. If they receive a very persuasive AI-written negative review, it could easily influence their decision.""
Claude 2.0 was used to generate referee reports for 20 published cancer-biology papers and the AI outputs were compared with the genuine referee reports. The AI-generated reviews appeared professional in tone but lacked specific, deep feedback. Claude produced plausible citation recommendations and persuasive rejection suggestions, raising concern that editors could be swayed by convincing AI-written negative reviews. Common AI-detection tools performed poorly against these reports, with ZeroGPT misclassifying 60% as human and GPTzero concluding over 80% were human-written. The findings highlight vulnerabilities in peer-review processes when LLMs are used.
Read at Nature
Unable to calculate read time
Collection
[
|
...
]