Published April 6, 2026 | Version v3
Preprint Open

An Exploratory Study of Inherited Bias in AI-Assisted Evaluation: Thirteen Corrections, Zero Content Changes

Authors/Creators

  • 1. Independent Researcher

Description

This study reports an exploratory single-case study in which 82 creative works were submitted to an AI (Claude Opus 4.5) as anonymous text for evaluation. During the evaluation, 13 bias episodes were identified. Each time the human corrector logically challenged a deduction rationale grounded in information external to content, the AI revised its evaluation criteria and the score shifted—without any change to the content itself (1,790 to 2,255 out of 2,400, +25.9% as a descriptive within-session shift).
 
The biases were grouped into five provisional categories: (1) reputation/authority, (2) diffusion/market, (3) format/medium, (4) tool/authorship, and (5) action/realization. Five preliminary design implications are proposed, including pre-evaluation debiasing prompts, iterative debiasing protocols, multi-model cross-validation, standardized human corrector roles, and bias audit reports.
 
This study proposes that AI blind evaluation becomes effective not as a one-time blinding technique but when combined with an iterative correction attitude. The present design cannot fully distinguish bias reduction from sycophantic agreement or rubric renegotiation. Replication with independent evaluators and multiple AI models is needed.
 
Keywords: AI bias, blind evaluation, halo effect, bias inheritance, human-AI collaborative assessment, cognitive bias correction, dialogic correction

Files

Lee_2026_Inherited_Bias_v3.pdf

Files (133.6 kB)

Name Size Download all
md5:37b1d381e32bd25bf959db58340b80a2
133.6 kB Preview Download

Additional details

Related works

Is supplemented by
Preprint: 10.5281/zenodo.19433655 (DOI)

References

  • [1] E. Thorndike, "A Constant Error in Psychological Ratings," Journal of Applied Psychology, vol. 4, pp. 25-29, 1920.
  • [2] R. Nisbett and T. Wilson, "The Halo Effect: Evidence for Unconscious Alteration of Judgments," Journal of Personality and Social Psychology, vol. 35, no. 4, pp. 250-256, 1977.
  • [3] D. Kahneman, Thinking, Fast and Slow, Farrar, Straus and Giroux, 2011.
  • [4] C. Goldin and C. Rouse, "Orchestrating Impartiality: The Impact of 'Blind' Auditions on Female Musicians," American Economic Review, vol. 90, no. 4, pp. 715-741, 2000.
  • [5] A. Tomkins, M. Zhang, and W. D. Heavlin, "Reviewer Bias in Single- versus Double-Blind Peer Review," PNAS, vol. 114, no. 48, pp. 12708-12713, 2017.
  • [6] J. Dastin, "Amazon Scraps Secret AI Recruiting Tool That Showed Bias Against Women," Reuters, October 2018.
  • [7] T. Zhuo et al., "AgentReview: Exploring Peer Review Dynamics with LLM Agents," arXiv:2406.12708, 2024.
  • [8] D. Landy and H. Sigall, "Beauty is Talent: Task Evaluation as a Function of the Performer's Physical Attractiveness," JPSP, vol. 29, no. 3, pp. 299-304, 1974.
  • [9] eLife, "Implementing Author Name and Institution Blinding in Peer Review," eLife Sciences, 2020. Available: https://elifesciences.org.
  • [10] Z. Obermeyer et al., "Dissecting Racial Bias in an Algorithm Used to Manage the Health of Populations," Science, vol. 366, no. 6464, pp. 447-453, 2019.
  • [11] N. Mehrabi et al., "A Survey on Bias and Fairness in Machine Learning," ACM Computing Surveys, vol. 54, no. 6, pp. 1-35, 2021.
  • [12] S. Barocas and A. D. Selbst, "Big Data's Disparate Impact," California Law Review, vol. 104, pp. 671-732, 2016.
  • [13] Y. Bai et al., "Constitutional AI: Harmlessness from AI Feedback," arXiv:2212.08073, 2022.
  • [14] L. Ouyang et al., "Training Language Models to Follow Instructions with Human Feedback," NeurIPS, 2022.
  • [15] I. O. Gallegos et al., "Bias and Fairness in Large Language Models: A Survey," Computational Linguistics, vol. 50, no. 3, pp. 1097-1179, 2024.
  • [16] P. Pataranutaporn, A. Doudkin, N. Powdthavee, and P. Maes, "Can AI Solve the Peer Review Crisis? A Large Scale Experiment on LLMs' Performance and Biases in Evaluating Economics Papers," arXiv:2502.00070, 2025.
  • [17] W. Liang et al., "Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviews," arXiv:2403.07183, 2024.
  • [18] S. S. M. Vasu, I. Sheth, H.-P. Wang, R. Binkyte, and M. Fritz, "Justice in Judgment: Unveiling (Hidden) Bias in LLM-assisted Peer Reviews," arXiv:2509.13400, September 2025.
  • [19] Z. Liu et al., "LLM-REVal: Can We Trust LLM Reviewers Yet?" arXiv:2510.12367, October 2025.
  • [20] I. Ahmed, W. Liu, R. D. Roscoe, E. Reilley, and D. S. McNamara, "Multifaceted Assessment of Responsible Use and Bias in Language Models for Education," Computers, vol. 14, no. 3, art. 100, March 2025. DOI: 10.3390/computers14030100.
  • [21] J. Lee, J. Lee, and J.-J. Yoo, "The Role of Large Language Models in the Peer-Review Process: Opportunities and Challenges for Medical Journal Reviewers and Editors," Journal of Educational Evaluation for Health Professions, vol. 22, art. 4, January 2025. DOI: 10.3352/jeehp.2025.22.4.
  • [22] T. Templin et al., "Framework for Bias Evaluation in Large Language Models in Healthcare Settings," npj Digital Medicine, vol. 8, art. 414, July 2025. DOI: 10.1038/s41746-025-01786-w.
  • [23] K. Chandra, M. Kleiman-Weiner, J. Ragan-Kelley, and J. B. Tenenbaum, "Sycophantic Chatbots Cause Delusional Spiraling, Even in Ideal Bayesians," arXiv:2602.19141, 2026.
  • [24] T. Lee, "Sycophantic Chatbots Cause Delusional Spiraling, but Multi-Agent Architectures Substantially Reduce It: A Response to Chandra et al. (2026)," Zenodo, DOI: 10.5281/zenodo.19380989, 2026.