From Compliance to Responsible AI: A Bibliometric Study on Educational Data Governance and Student Privacy (2015-2025)

This study maps the development of research on educational data governance and student privacy between 2015 and 2025. It follows the structure of a bibliometric trend review and uses a Scopus-derived record set of 313 articles to analyze the analytical workflow, including annual output, keyword co-occurrence, thematic evolution, policy turning points, and research collaboration. The study is guided by three questions: how the volume of research changed, how the dominant vocabulary moved from learning analytics and ownership to compliance and responsible artificial intelligence, and which problems remained unresolved across the period. The results show five connected phases. From 2015 to 2017, research was organized around learning analytics, educational data mining, student tracking, data ownership, and informed consent. After the General Data Protection Regulation became applicable on May 25, 2018, the literature moved toward data protection, compliance, governance frameworks, and institutional responsibility. The COVID-19 period increased attention to remote learning privacy, online proctoring, surveillance, and vendor data sharing. From late 2022, generative artificial intelligence, ChatGPT, algorithmic bias, ethical artificial intelligence, and artificial intelligence literacy became central. The 2024-2025 frontier connected responsible artificial intelligence with high-risk education systems, human oversight, auditability, and data minimization in large language models. The study argues that compliance is necessary but insufficient. A mature governance model must connect legal duties, institutional processes, technical controls, and educational judgment.