@article{ART003317795},
author={Lee, Jeong-Mee},
title={An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective},
journal={Journal of the Korean Biblia Society for Library and Information Science},
issn={1229-2435},
year={2026},
volume={37},
number={1},
pages={467-488},
doi={10.14699//kbiblia.2026.37.1.467}
TY - JOUR
AU - Lee, Jeong-Mee
TI - An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective
JO - Journal of the Korean Biblia Society for Library and Information Science
PY - 2026
VL - 37
IS - 1
PB - Journal Of The Korean Biblia Society For Library And Information Science
SP - 467
EP - 488
SN - 1229-2435
AB - This study seeks to reconceptualize issues of bias and fairness in AI-based automatic metadata generation from the perspectives of knowledge organization and resource discovery. Through a systematic review of recent domestic and international literature, it examines the stages at which AI intervenes in information organization and how structural biases—such as semantic reduction, representational imbalance, uncontrolled vocabulary proliferation, and distortion of classificatory hierarchies—emerge in the process. It further discusses how these biases affect the visibility of search results, subject accessibility, and opportunities for resource exposure in terms of fairness in resource discovery.
The findings indicate that although AI-driven metadata generation enhances efficiency, its reliance on statistically distributed training data and algorithmic inference may lead to the over- or under-representation of particular subjects or groups. This issue should be understood as an information-scientific concern that reshapes the very structure through which knowledge is organized and made visible. Accordingly, evaluating AI-generated metadata requires an integrated analytical framework that goes beyond performance metrics to consider alignment with knowledge organization systems and equity in resource discovery.
By framing AI metadata bias as a structural imbalance in information flow, this study provides a theoretical foundation for advancing fairness in resource discovery grounded in information science.
KW - AI-based Metadata Generation;Knowledge Organization Systems (KOS);Resource Discovery;Algorithmic Bias;Fairness in Information Systems
DO - 10.14699//kbiblia.2026.37.1.467
ER -
Lee, Jeong-Mee. (2026). An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective. Journal of the Korean Biblia Society for Library and Information Science, 37(1), 467-488.
Lee, Jeong-Mee. 2026, "An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective", Journal of the Korean Biblia Society for Library and Information Science, vol.37, no.1 pp.467-488. Available from: doi:10.14699//kbiblia.2026.37.1.467
Lee, Jeong-Mee "An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective" Journal of the Korean Biblia Society for Library and Information Science 37.1 pp.467-488 (2026) : 467.
Lee, Jeong-Mee. An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective. 2026; 37(1), 467-488. Available from: doi:10.14699//kbiblia.2026.37.1.467
Lee, Jeong-Mee. "An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective" Journal of the Korean Biblia Society for Library and Information Science 37, no.1 (2026) : 467-488.doi: 10.14699//kbiblia.2026.37.1.467
Lee, Jeong-Mee. An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective. Journal of the Korean Biblia Society for Library and Information Science, 37(1), 467-488. doi: 10.14699//kbiblia.2026.37.1.467
Lee, Jeong-Mee. An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective. Journal of the Korean Biblia Society for Library and Information Science. 2026; 37(1) 467-488. doi: 10.14699//kbiblia.2026.37.1.467
Lee, Jeong-Mee. An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective. 2026; 37(1), 467-488. Available from: doi:10.14699//kbiblia.2026.37.1.467
Lee, Jeong-Mee. "An Analysis of Bias and Fairness in Generative AI-Based Metadata Automation: An Information Science Perspective" Journal of the Korean Biblia Society for Library and Information Science 37, no.1 (2026) : 467-488.doi: 10.14699//kbiblia.2026.37.1.467
