Tao, Yingnian and Ryan, Mark and Bialy, Filip and Zhang, Xingna (2026) Assessment-focused governance mode? Mapping Generative AI policies across 76 UK higher education institutions: Policy framework and recommendations. Other. Research Square.
Full text not available from this repository.Abstract
Higher education institutions are under pressure to develop Generative AI (GenAI) policies in teaching, learning, and assessment. How these policies are framed provides important insights into how the sector is responding to the opportunities and challenges posed by GenAI. This paper examines GenAI policy guidelines published by 76 UK universities, comprising 132 policy and guidance documents. Through qualitative in-depth thematic analysis, we identify five dominant policy themes: institutional positioning on GenAI; assessment design; marking and feedback; teaching and learning; and AI ethics. Our findings reveal an assessment-centred and integrity-first AI governance model. Institutional attention is directed primarily toward safeguarding assessment integrity, preventing academic misconduct and regulating student use of AI in assessed work. By contrast, considerably less attention is given to the pedagogical integration of GenAI into teaching and learning processes. We further find variations across institutions in key policy areas, including AI attribution, AI detection, and AI use in marking and feedback, reflecting broader epistemic, technological and governance uncertainties surrounding the technology in higher education. In response, we propose a policy framework and ten recommendations to support a more balanced approach to GenAI governance. As the first large-scale mapping of GenAI policy across UK higher education, this study makes an original contribution to scholarship on AI governance in education and provides evidence-based guidance for future policy review and development.