LLMs in Peer Review: Capabilities, Biases, and Vulnerabilities Explored
ByPulseAugur Editorial·[10 sources]·
Multiple research papers are evaluating the capabilities and limitations of Large Language Models (LLMs) in the context of academic peer review. Studies like 'Review Arcade' and 'PRAIB' investigate LLM alignment with human reviewers, finding that while LLMs can be helpful, they exhibit biases such as overconfidence and less variability in ratings. 'PRISM' introduces a benchmark to assess LLM reviewers across dimensions like depth of analysis and flaw identification, concluding that LLMs excel in specific areas but lack the balanced performance of human reviewers. Additionally, research on 'TADDLE' focuses on detecting deficiencies in LLM-generated reviews, which are often fluent but may contain subtle errors. A significant concern highlighted is the vulnerability of LLMs to prompt injection attacks, which can manipulate review outcomes.
AI
IMPACTLLM integration into academic peer review requires careful calibration and safeguards against biases and adversarial attacks to ensure reliability.
RANK_REASON
Multiple research papers introduce new benchmarks and analysis frameworks for evaluating LLM performance in academic peer review.
arXiv:2605.28897v1 Announce Type: new Abstract: LLM-generated reviews for scientific papers are gaining considerable traction and are even being officially piloted by major conferences. We have to assume that not only reviewers are using LLM-assistance, but also that authors use …
arXiv cs.AI
TIER_1English(EN)·Krzysztof \.Zurawicki, Julia Farganus, Arkadiusz Gawe{\l}, Mateusz Bystro\'nski, Tomasz Jan Kajdanowicz·
arXiv:2605.29815v1 Announce Type: new Abstract: The growing number of submitted papers has motivated the exploration of Large Language Models (LLMs) as a means to support and augment the peer review process, particularly in terms of improving its speed and scalability. Yet, it re…
arXiv cs.CL
TIER_1English(EN)·Tomasz Jan Kajdanowicz·
The growing number of submitted papers has motivated the exploration of Large Language Models (LLMs) as a means to support and augment the peer review process, particularly in terms of improving its speed and scalability. Yet, it remains unknown whether LLMs engage with scientifi…
LLM-generated reviews for scientific papers are gaining considerable traction and are even being officially piloted by major conferences. We have to assume that not only reviewers are using LLM-assistance, but also that authors use LLMs to revise their papers before submitting. I…
arXiv cs.CL
TIER_1English(EN)·Ngoc Phan Phuoc Loc, Toan Huynh La Viet, Thanh Tran Khanh, Duy A Nguyen, Tuan Anh Nguyen Pham, Thanh Nguyen, Nitesh V. Chawla, Wray Buntine, Kok-Seng Wong, Khoa D. Doan, Binh T. Nguyen·
arXiv:2605.26730v1 Announce Type: new Abstract: The rapid growth in submissions to machine learning venues has strained the scientific peer-review system and intensified interest in LLM-based automated peer reviewers. However, how good these systems are actually, especially compa…
arXiv:2605.26911v1 Announce Type: new Abstract: LLM-generated peer reviews are increasingly common at major venues, yet their deficiencies are hard to detect because they are uniformly fluent and well-structured. Existing work either classifies authorship without judging quality,…
PRISM evaluates automated peer review systems across multiple dimensions using argument mining and retrieval-augmented verification, revealing that while LLMs match human performance in specific areas, no system consistently equals human reviewers across all evaluation criteria.
LLM-generated peer reviews are increasingly common at major venues, yet their deficiencies are hard to detect because they are uniformly fluent and well-structured. Existing work either classifies authorship without judging quality, or scores quality with features designed for hu…
arXiv:2605.25415v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used in academic peer review, yet their reliability, alignment with human judgment, and robustness to adversarial attacks remain poorly understood. We present a systematic benchmark of L…
Large language models (LLMs) are increasingly used in academic peer review, yet their reliability, alignment with human judgment, and robustness to adversarial attacks remain poorly understood. We present a systematic benchmark of LLM-as-a-Reviewer on 898 papers stratified from N…
