The Intellectual Structure of Research on Agentic AI Ethics and Governance in Higher Education: A Systematic Evidence Synthesis and Research Agenda

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Research on the ethics and governance of artificial intelligence in higher education has expanded rapidly since 2022, but the field's intellectual structure, its dominant themes, internal debates, and blind spots have not been systematically mapped since AI tools began exhibiting agentic properties such as multistep planning and autonomous task execution. This evidence synthesis maps the field using a PRISMA guided systematic search and bibliometric analysis of 312 publications (2019 to 2026) drawn from Web of Science and Scopus. Keyword co-occurrence and co-citation analyses identify five thematic clusters: transparency and explainability; distributed accountability and liability; human oversight and controllability; value alignment and epistemic integrity; and equity, access, and algorithmic fairness. For each cluster, the synthesis characterizes the constituent literature, its convergent findings, and its unresolved disagreements. Across clusters, six structural gaps emerge: a deficit of empirical implementation studies relative to conceptual work; the absence of validated measurement instruments for responsible AI use; the underrepresentation of community colleges, minority-serving institutions, and Global South contexts; an imbalance that favors accountability directed at students over accountability directed at institutions and developers; uncertain transfer of findings from before 2023 to agentic systems; and a weak connection between education research and technical AI governance venues. The paper concludes with a research agenda addressing each gap in turn, intended to guide researchers, editors, and funders as agentic AI deployment in higher education accelerates.