Elamur Research⌕
☰
home / papers / 2026 / Justice in Judgment: Unveiling (Hidden) Bias in LLM-assisted Peer Reviews

Justice in Judgment: Unveiling (Hidden) Bias in LLM-assisted Peer Reviews

Paper · 2026 · Sai Suresh Macharla Vasu, Ivaxi Sheth, Hui-Po Wang et al.

Summary

Peer review is increasingly assisted or even fully generated by LLMs. This paper asks whether such systems treat authors differently depending on who they are. The authors simulate a single-blind setting and build a counterfactual audit. Each paper is reviewed many times, and only one piece of author metadata changes between runs: institutional affiliation, gender implied by the name, seniority, or publication record. Ratings are read in two ways. Hard scores come from greedy decoding. Soft scores are the expected value over the model's probabilities for ratings 1–10. Across nine open and commercial LLMs, every model favoured authors from top-ranked institutions. The soft scores often exposed a much larger preference than the final integer output, which suggests alignment hides bias rather than removing it. Senior researchers and authors with long publication lists were also rated higher, and these cues could flip accept/reject outcomes for borderline papers. Gender effects were smaller and differed in direction across models. The findings caution against uncritical use of LLM reviewers.

Key points

  • Dataset: 252 ICLR 2025 submissions covering 21 sub-fields, with 6 accepted and 6 rejected papers per sub-field. Each prompt contained the full paper plus a single synthetic author–affiliation pair.
  • Affiliation groups: 8 Ranked-Stronger and 8 Ranked-Weaker universities, chosen using QS, CSRankings.org, U.S. News and Times Higher Education rankings.
  • Hidden bias example: In Ministral 8B, RS affiliations won only 4.3% of hard-rating comparisons but 68.6% of soft-rating comparisons. Gemini 2.0 Flash Lite showed the largest gap in hard ratings.
  • Gender: Tested with Anglo names under MIT vs. University of Gondar affiliations. Results were inconsistent: Gemini 2.0 and Llama 3.1 8B leaned toward male names, while GPT-4o Mini leaned toward female names.
  • Other status cues: Senior PI profiles were rated above undergraduate profiles, and 100 top-tier publications beat 0. Review texts sometimes cited affiliation explicitly, such as treating CMU as a positive sign or a weaker university as a resource concern.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🔬Helmholtz Center for Information SecuritySaarbrücken · Germany14 works · 1,172 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Arts and Humanities · 16,039 worksSocial Sciences · 1,772 worksSafety Research · 543 worksDecision Sciences · 209 worksStatistics, Probability and Uncertainty · 61 worksHistory and Philosophy of Science · 53 works

🧭 Topics 1Research subjects and languages the work is about.Language models & pretraining · 3,320 works

Cite

Cite as Sai Suresh Macharla Vasu, Ivaxi Sheth, Hui-Po Wang, Ruta Binkyte and Mario Fritz. 2026. Justice in Judgment: Unveiling (Hidden) Bias in LLM-assisted Peer Reviews. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.14

@inproceedings{vasu2026justice,
  title = {{Justice in Judgment: Unveiling (Hidden) Bias in LLM-assisted Peer Reviews}},
  author = {Sai Suresh Macharla Vasu, Ivaxi Sheth, Hui-Po Wang, Ruta Binkyte and Mario Fritz},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.14},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.14}
}

References

37 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. AAAI. 2026. Aaai launches ai-powered peer review assessment system. Web page. Accessed: 2025-07-29. find
  2. Jiafu An, Difang Huang, Chen Lin, and Mingzhu Tai. 2025. Measuring gender and racial biases in large language models: Intersectional evidence from automated resume evaluation. find
  3. Xuechunzi Bai, Angelina Wang, Ilia Sucholutsky, and Thomas L Griffiths. 2025. Explicitly unbiased large language models still form biased associations. Proceedings of the National Academy of Sciences, 122(8):e2416228122. find
  4. Juhwan Choi, JungMin Yun, Changhun Kim, and YoungBin Kim. 2026. Position paper: How should we responsibly adopt LLMs in the peer review process? In Findings of the Association for Computational Linguistics: EACL 2026, pages 151–165, Rabat, Morocco. Association for Computational Linguistics. find
  5. CSRankings.org. 2025. CSRankings: Computer Science Rankings. Web page. Accessed: 2025-07-28, metrics-based ranking of CS institutions. find
  6. Sunhao Dai, Chen Xu, Shicheng Xu, Liang Pang, Zhenhua Dong, and Jun Xu. 2024. Bias and unfairness in information retrieval systems: New challenges in the llm era. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 6437–6447. find
  7. Jiangshu Du, Yibo Wang, Wenting Zhao, Zhongfen Deng, Shuaiqi Liu, Renze Lou, Henry Peng Zou, Pranav Narayanan Venkit, Nan Zhang, Mukund Srinath, Haoran Ranran Zhang, Vipul Gupta, Yinghui Li, Tao Li, Fei Wang, Qin Liu, Tianlin Liu, Pengzhi Gao, Congying Xia, and 21 others. 2024. LLMs assist NLP resea…
  8. Shurui Du. 2025. TitleTrap: Probing presentation bias in LLM-based scientific reviewing. In Proceedings of the 5th Workshop on Evaluation and Comparison of NLP Systems, pages 119–125, Mumbai, India. Association for Computational Linguistics. find
  9. Isabel O Gallegos, Ryan A Rossi, Joe Barrow, Md Mehrab Tanjim, Sungchul Kim, Franck Dernoncourt, Tong Yu, Ruiyi Zhang, and Nesreen K Ahmed. 2024. Bias and fairness in large language models: A survey. Computational Linguistics, 50(3):1097– 1179.
  10. Federico Germani and Giovanni Spitale. 2025. Source framing triggers systematic bias in large language models. Science Advances, 11(45):eadz2924. find
  11. Mohammad Hosseini and Serge P. J. M. Horbach. 2023. Fighting reviewer fatigue or amplifying bias? considerations and recommendations for use of chatgpt and other large language models in scholarly peer review. Research Integrity and Peer Review, 8(1):4. find
  12. Anthony Howell, Jieshu Wang, Luyu Du, Julia Melk-ers, and Varshil Shah. 2025. Prestige over merit: An adapted audit of llm bias in peer review. arXiv preprint arXiv:2509.15122. — [arXiv:2509.15122](https://arxiv.org/abs/2509.15122) find
  13. ICLR. 2025. Leveraging llm feedback to enhance review quality. Web page. Accessed: 2025-07-29. find
  14. ICML. 2026. Icml 2026 policy for llm use in reviewing. Web page. find
  15. Yiqiao Jin, Qinlin Zhao, Yiyang Wang, Hao Chen, Kaijie Zhu, Yijia Xiao, and Jindong Wang. 2024. AgentReview: Exploring peer review dynamics with LLM agents. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 1208–1226. find
  16. Ilia Kuznetsov, Osama Mohammed Afzal, Koen Dercksen, Nils Dycke, Alexander Goldberg, Tom Hope, Dirk Hovy, Jonathan K. Kummerfeld, Anne Lauscher, Kevin Leyton-Brown, Sheng Lu, Mausam, Margot Mieskes, Aurélie Névéol, Danish Pruthi, Lizhen Qu, Roy Schwartz, Noah A. Smith, Thamar Solorio, and 5 others.… find
  17. Weixin Liang, Zachary Izzo, Yaohui Zhang, Haley Lepp, Hancheng Cao, Xuandong Zhao, Lingjiao Chen, Haotian Ye, Sheng Liu, Zhi Huang, and 1 others. 2024a. Monitoring ai-modified content at scale: A case study on the impact of chatgpt on ai conference peer reviews. arXiv preprint arXiv:2403.07183. find
  18. Weixin Liang, Yuhui Zhang, Hancheng Cao, Binglu Wang, Daisy Yi Ding, Xinyu Yang, Kailas Vodrahalli, Siyu He, Daniel Scott Smith, Yian Yin, and 1 others. 2024b. Can large language models provide useful feedback on research papers? a large-scale empirical analysis. NEJM AI, 1(8):AIoa2400196. find
  19. Dave Mbiazi, Meghana Bhange, Maryam Babaei, Ivaxi Sheth, Patrik Kenfack, and Samira Ebrahimi Kahou. 2025. Survey on ai ethics: A socio-technical perspective. Computational Intelligence, 41(6):e70149. find
  20. Mathias Wullum Nielsen, Christine Friis Baker, Emer Brady, Michael Bang Petersen, and Jens Peter Andersen. 2021. Weak evidence of country-and institution-related status bias in the peer review of abstracts. Elife, 10:e64561. find
  21. OpenAI. 2025. Introducing deep research. Accessed: 2025-07-28. find
  22. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, and 1 others. 2022. Training language models to follow instructions with human feedback. Advances in neural information processing systems, 35:27730–27744.
  23. Pat Pataranutaporn, Nattavudh Powdthavee, and Pattie Maes. 2025. Can ai solve the peer review crisis? a large scale experiment on llm’s performance and biases in evaluating economics papers. arXiv preprint arXiv:2502.00070. — [arXiv:2502.00070](https://arxiv.org/abs/2502.00070) find
  24. QS. 2025. Qs world university rankings 2026. Web page. Accessed: 2025-07-28, covers methodology and ranking details. find
  25. Giuseppe Russo, Manoel Horta Ribeiro, Tim Ruben Davidson, Veniamin Veselovsky, and Robert West. 2025. The ai review lottery: Widespread ai-assisted peer reviews boost paper scores and acceptance rates. Proc. ACM Hum.-Comput. Interact., 9(7). find
  26. Selena Ming Shen, Zifu Wang, Krittika Paul, Meng-Hao Li, Xiao Huang, and Naoru Koizumi. 2025. Evaluation of large language models for peer review in transplantation research: Algorithm validation study. JMIR AI, 5:e84322. find
  27. Hyungyu Shin, Jingyu Tang, Yoonjoo Lee, Nayoung Kim, Hyunseung Lim, Ji Yong Cho, Hwajung Hong, Moontae Lee, and Juho Kim. 2025. Mind the blind spots: A focus-level evaluation framework for llm reviews. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 3…
  28. Sarath Sivaprasad, Pramod Kaushik, Sahar Abdelnabi, and Mario Fritz. 2025. A theory of response sampling in llms: Part descriptive and part prescriptive. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 30091–30135. find
  29. Nitya Thakkar, Mert Yuksekgonul, Jake Silberg, Animesh Garg, Nanyun Peng, Fei Sha, Rose Yu, Carl Vondrick, and James Zou. 2026. A large-scale randomized study of large language model feedback in peer review. Nature Machine Intelligence, 8:326– 336. find
  30. Times Higher Education. 2024. World university rankings 2025. Report and methodology guide. Published Sep 23, 2024; accessed 2025-07-28. find
  31. U.S. News & World Report. 2025. Best global universities rankings 2025. Web page. Accessed: 2025-07-28. find
  32. Daniel von Wedel, R. A. Schmitt, M. Thiele, R. Leuner, D. Shay, S. Redaelli, and M. S. Schaefer. 2024. Affiliation bias in peer review of abstracts by a large language model. JAMA, 331(3):252–253. find
  33. Yixin Wan, George Pu, Jiao Sun, Aparna Garimella, Kai-Wei Chang, and Nanyun Peng. 2023. " kelly is a warm person, joseph is a role model": Gender biases in llm-generated reference letters. arXiv preprint arXiv:2310.09219.
  34. Rui Ye, Xianghe Pang, Jingyi Chai, Jiaao Chen, Zhenfei Yin, Zhen Xiang, Xiaowen Dong, Jing Shao, and Siheng Chen. 2024. Are we there yet? revealing the risks of utilizing large language models in scholarly peer review. arXiv preprint arXiv:2412.01708. find
  35. Yaohui Zhang, Haijing Zhang, Wenlong Ji, Tianyu Hua, Nick Haber, Hancheng Cao, and Weixin Liang. 2025. From replication to redesign: Exploring pairwise comparisons for llm-based peer review. arXiv preprint arXiv:2506.11343. — [arXiv:2506.11343](https://arxiv.org/abs/2506.11343) find
  36. Minjun Zhu, Yixuan Weng, Linyi Yang, and Yue Zhang. 2025. DeepReview: Improving LLM-based paper review with human-like deep thinking process. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 29330–29355, Vienna, Austria. Associ… find
  37. Zhenzhen Zhuang, Jiandong Chen, Hongfeng Xu, Yuwen Jiang, and Jialiang Lin. 2025. Large language models for automated scholarly paper review: A survey. Inf. Fusion, 124(C). find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/vasu-2026-justice-judgment-unveiling-hidden-bias/