Mapping the Policy Landscape of AI in Higher Education: A Multi-level Analysis Across European Institutions

Abstract

Abstract Generative AI has turned AI governance in higher education from a strategic concern into an operational issue affecting assessment, academic integrity, data protection, procurement, and institutional accountability. Yet comparative analysis across European, national, and institutional policy levels remains limited. This paper presents a multi-level comparative document analysis of 56 policy and guidance documents across seven partner cases in six European countries. Using a common review protocol and an eight-dimensional coding framework, it examines how ethics, privacy, inclusion, transparency, governance, technical standards, accountability, and sustainability are articulated across policy levels. The findings show a clear maturity gradient: European instruments provide strong normative and legal anchors, national documents remain largely strategic, and institutional policies are the most operational but also the narrowest in scope. Strongest coverage concerns ethics, privacy, transparency, and academic integrity, while governance, ownership, liability, interoperability, procurement, and sustainability remain weak. A recurring pedagogical integration gap is also evident. The study, therefore, provides a comparative policy map of AI governance in European higher education and highlights where multi-level misalignment continues to limit coherent institutional operationalisation.