titleOn the limits and opportunities of AI reviewers: Reviewing the reviews of Nature-family papers with 45 expert scientists
authorsSeungone Kim, Dongkeun Yoon, Kiril Gashteovski, Juyoung Suk, Jinheon Baek, Pranjal Aggarwal, Ian Wu, Viktor Zaverkin, Spase Petkoski, Daniel R. Schrider, Ilija Dukovski, Francesco Santini, Biljana Mitreska, Yong Jeong, Kyeongha Kwon, Young Min Sim, Dragana Manasova, Arthur Porto, Biljana Mojsoska, Makoto Takamoto, Marko Shuntov, Ruoqi Liu, Hyunjoo Jenny Lee, Niyazi Ulas Dinç, Yehhyun Jo, Sunkyu Han, Chungwoo Lee, Huishan Li, Esther H. R. Tsai, Ergun Simsek, Khushboo Shafi, Yeonseung Chung, Jihye Park, Aleksandar Shulevski, Henrik Christiansen, Yoosang Son, Elly Knight, Amanda Montoya, Jeongyoun Ahn, Christian Langkammer, Heera Moon, Changwon Yoon, Nikola Stikov, Mooseok Jang, Edward Choi, Junhan Kim, Yeon Sik Jung, Woo Youn Kim, Jae Kyoung Kim, Ishraq Md Anjum, Hyun Uk Kim, Drew Bridges, Carolin Lawrence, Xiang Yue, Alice Oh, Akari Asai, Sean Welleck, Graham Neubig
date2026-05-20
venuearXiv (preprint, not peer reviewed)
idarXiv:2605.20668
affiliationCarnegie Mellon University
upvotes13 (Hugging Face Daily Papers, 2026-11-01)
linksabs pdf hf code
abstractWith the advancement of AI capabilities, AI reviewers are beginning to be deployed in scientific peer review, yet their capability and credibility remain in question: many scientists simply view them as probabilistic systems without the expertise to evaluate research, while other researchers are more optimistic about their readiness without concrete evidence. Understanding what AI reviewers do well, where they fall short, and what challenges remain is essential. However, existing evaluations of AI reviewers have focused on whether their verdicts match human verdicts (e.g., score alignment, acceptance prediction), which is insufficient to characterize their capabilities and limits. In this paper, we close this gap through a large-scale expert annotation study, in which 45 domain scientists in Physical, Biological, and Health Sciences spent 469 hours rating 2,960 individual criticisms (each targeting one specific aspect of a paper) from human-written and AI-generated reviews of 82 Nature-family papers on correctness, significance, and sufficiency of evidence. On a composite of all three dimensions, a reviewing agent powered by GPT-5.2 scores above each paper's top-rated human reviewer (60.0% vs. 48.2%, p = 0.009), while all three AI reviewers (including Gemini 3.0 Pro and Claude Opus 4.5) exceed the lowest-rated human across every dimension. AI reviewers' accurate criticisms are also more often rated significant and well-evidenced, and surface a distinct 26% of issues no human raises. However, AI reviewers overlap far more than humans do (21% vs. 3% for cross-reviewer pairs), and exhibit 16 recurring weaknesses humans do not share, such as limited subfield knowledge, lack of long context management over multiple files, and overly critical stance on minor issues. Overall, our results position current AI reviewers as complements to, not substitutes for, human reviewers.

cite

@misc{kim2026limits,
  title={On the limits and opportunities of AI reviewers: Reviewing the reviews of Nature-family papers with 45 expert scientists},
  author={Seungone Kim and Dongkeun Yoon and Kiril Gashteovski and Juyoung Suk and Jinheon Baek and Pranjal Aggarwal and Ian Wu and Viktor Zaverkin and Spase Petkoski and Daniel R. Schrider and Ilija Dukovski and Francesco Santini and Biljana Mitreska and Yong Jeong and Kyeongha Kwon and Young Min Sim and Dragana Manasova and Arthur Porto and Biljana Mojsoska and Makoto Takamoto and Marko Shuntov and Ruoqi Liu and Hyunjoo Jenny Lee and Niyazi Ulas Dinç and Yehhyun Jo and Sunkyu Han and Chungwoo Lee and Huishan Li and Esther H. R. Tsai and Ergun Simsek and Khushboo Shafi and Yeonseung Chung and Jihye Park and Aleksandar Shulevski and Henrik Christiansen and Yoosang Son and Elly Knight and Amanda Montoya and Jeongyoun Ahn and Christian Langkammer and Heera Moon and Changwon Yoon and Nikola Stikov and Mooseok Jang and Edward Choi and Junhan Kim and Yeon Sik Jung and Woo Youn Kim and Jae Kyoung Kim and Ishraq Md Anjum and Hyun Uk Kim and Drew Bridges and Carolin Lawrence and Xiang Yue and Alice Oh and Akari Asai and Sean Welleck and Graham Neubig},
  year={2026},
  eprint={2605.20668},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2605.20668}
}