Hybrid AI governance policies and practices in top higher education institutions

PURPOSE: As Artificial Intelligence (AI) integrates into Higher Education Institutions (HEIs), governance structures shape its implementation, oversight, and adaptability. A key challenge is determining centralized versus decentralized governance models. This study examines AI governance across 69 HEIs, analyzing policy centralization and its link to institutional priorities, governance mechanisms, and flexibility. METHODOLOGY: Using AI policies from Times Higher Education (THE) top 100 universities, it identifies structural differences among governance models. The study adopts a mixed-methods analytical strategy combining qualitative policy analysis with quantitative statistical techniques, namely principal component analysis (PCA). In qualitative analysis, the study employed a desk research methodology, combining external and internal searches to broaden the analytical perspective and explore theoretical dimensions. The quantitative analysis used descriptive statistics, rank-order correlations, principal component analysis (PCA) for dimensionality reduction, and hierarchical cluster analysis on PCA scores to identify governance configurations and institutional clusters. FINDINGS: AI governance in higher education emerges as a negotiated process shaped by institutional priorities, risk perceptions, and regulatory pressures, rejecting a simple centralized-decentralized dichotomy. Instead, institutions adopt varied configurations that blend oversight with academic autonomy, particularly by centralizing compliance, security, and ethics at the institutional level. IMPLICATIONS: This study implies that HEIs should adopt flexible hybrid AI governance models to balance centralized oversight (e.g., for compliance and ethics) with decentralized academic autonomy, enhancing adaptability to evolving AI risks, regulatory pressures, and inclusivity needs while fostering innovation. Future research should adopt longitudinal designs to track AI governance evolution amid technological, ethical, and regulatory shifts; compare configurations across diverse HEIs, regions, and economies; and assess the impacts of AI-driven tools, such as real-time monitoring, on innovation, inclusivity, staff agency, and efficiency gains. ORIGINALITY & VALUE: This study contributes to governance theory by reconceptualizing AI policy centralization as a dynamic, multidimensional negotiation rather than a binary spectrum, integrating institutional priorities, risk perceptions, and hybrid configurations into higher education governance frameworks.