Evaluating the potential risks of employing large language models in peer review
Abstract Objective This study aims to systematically investigate the potential harms of Large Language Models (LLMs) in the peer review process. Background LLMs are increasingly used in academic processes, including peer review. While they can address challenges like reviewer scarcity and review efficiency, concerns about fairness, transparency and potential biases in LLM‐generated reviews have not been thoroughly investigated. Methods Claude 2.0 was used to generate peer review reports, rejection recommendations, citation requests and refutations for 20 original, unmodified cancer biology manuscripts obtained from eLife 's new publishing model. Artificial intelligence (AI) detection tools (zeroGPT and GPTzero) assessed whether the reviews were identifiable as LLM‐generated.All LLM‐generated outputs were evaluated for reasonableness by two expert on a five‐point Likert scale. Results LLM‐generated reviews were somewhat consistent with human reviews but lacked depth, especially in detailed critique. The model proved highly proficient at generating convincing rejection comments and could create plausible citation requests, including requests for unrelated references. AI detectors struggled to identify LLM‐generated reviews, with 82.8% of responses classified as human‐written by GPTzero. Conclusions LLMs can be readily misused to undermine the peer review process by generating biased, manipulative, and difficult‐to‐detect content, posing a significant threat to academic integrity. Guidelines and detection tools are needed to ensure LLMs enhance rather than harm the peer review process.
Authors
- Jason Chia‐Hsien Cheng (ORCID: https://orcid.org/0000-0002-2007-0220)
- Bufu Tang (ORCID: https://orcid.org/0000-0003-3525-4433)
- Wenyi Gan (ORCID: https://orcid.org/0000-0003-1886-8062)
- Dongqiang Zeng (ORCID: https://orcid.org/0000-0001-9724-8325)
- Tao Yang (ORCID: https://orcid.org/0000-0002-7382-2983)
- Aimin Jiang (ORCID: https://orcid.org/0000-0001-9161-5729)
- Peng Luo (ORCID: https://orcid.org/0000-0002-8215-2045)
- Mingjia Xiao
- Weiming Mou (ORCID: https://orcid.org/0009-0007-1089-6516)
- Jiarui Xie (ORCID: https://orcid.org/0000-0001-9022-4881)
- Anqi Lin (ORCID: https://orcid.org/0000-0002-6324-0410)
- Lihaoyun Huang (ORCID: https://orcid.org/0000-0003-3833-9117)
- Chang Qi (ORCID: https://orcid.org/0009-0005-3840-550X)
- Yancheng Lai (ORCID: https://orcid.org/0009-0004-8444-7535)
- Zaoqu Liu
- Lingxuan Zhu
- Guangdi Chu
Institutions
- Qingdao University (CN)
- Central South University (CN)
- TU Wien (AT)
- Sun Yat-sen University (CN)
- Jinan University (CN)
- Chinese University of Hong Kong (HK)
- Second Military Medical University (CN)
- Shanghai Jiao Tong University (CN)
- Chinese Academy of Medical Sciences & Peking Union Medical College (CN)
- Fudan University (CN)
- Wenzhou Medical University (CN)
- Quzhou City People's Hospital (CN)
- Nanfang Hospital (CN)
- Zhuhai People's Hospital (CN)
- Affiliated Hospital of Qingdao University (CN)
- Zhujiang Hospital (CN)
- Quzhou University (CN)
- Zhongshan Hospital (CN)
- The First Affiliated Hospital, Sun Yat-sen University (CN)
- Shanghai First People's Hospital (CN)
- Xiangya Hospital Central South University (CN)
- Kangda College of Nanjing Medical University
- Southern Medical University (CN)
- Nanjing Medical University (CN)
- University of Hong Kong (HK)
Publication Details
- Journal
- Clinical and Translational Discovery
- Published
- 2025-06-27
- DOI
- https://doi.org/10.1002/ctd2.70067
- Citations
- 5
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 3.27