<![CDATA[As AI systems increasingly govern assessment, feedback, and learning pathways in education, concerns over algorithmic bias have expanded from technical issues to governance and equity. However, limited attention has been given to how the research landscape itself frames these bias-related issues. This study addresses that gap through a bibliometric analysis of 445 Scopus-indexed publications (1975–April 2025) using Biblioshiny. The results reveal a sharp rise in publications since 2019, with research concentrated in technologically advanced countries. Dominant themes include machine learning, ethics, and medical education, while equity-oriented topics such as gender bias, vocational education, and contextual disparities remain markedly underrepresented. The growing adoption of generative AI further complicates bias dynamics by introducing opacity and amplifying dataset biases. The discussion highlights that geographical and thematic concentration suggests a narrow institutional perspective, potentially marginalizing bias concerns from lower-resource contexts and creating a disconnect between technical AI research and educational justice frameworks. The novelty of this study lies in its meta-level examination of how the research field structures bias priorities, empirically revealing thematic omissions and identifying generative AI as an amplifying factor in educational bias. Collectively, these findings provide a structural foundation for developing more transparent, accountable, and equity-driven AI practices in education.]]>
Copyrights © 2026
