Toward ethical governance of generative artificial intelligence in higher education: a Delphi study of key dimensions and priorities

The rapid integration of generative artificial intelligence (GAI) into higher education has introduced significant opportunities alongside complex ethical challenges. While existing research has largely focused on specific tools such as large language models, there remains limited consensus on the broader dimensions that shape the ethical use of GAI across institutional, pedagogical, and technological contexts. This study addresses the gap by identifying and prioritizing key dimensions that influence the ethical use of GAI in higher education. A three-round Delphi method was employed, involving 31 experts from higher education institutions across China and Malaysia. In Round 1, qualitative responses were analyzed using an inductive approach to derive six overarching dimensions. In Round 2, experts ranked these dimensions based on perceived importance. A third round was conducted to refine consensus, resulting in a statistically significant level of agreement (Kendall’s W = 0.248, p = .001), indicating moderate convergence of expert opinion. The findings revealed six key dimensions, ranked in order of importance: (1) Institutional Governance and Policy Frameworks, (2) Academic Integrity and Ethical Accountability, (3) GAI Literacy, Training, and Capacity Building, (4) Technical Safeguards, Data Privacy, and Fairness, (5) Transparency, Critical Thinking, and Responsible Use Culture, and (6) Pedagogical Adaptation and Teacher Roles. Among these, institutional governance emerged as the foundational dimension. The study contributes to the literature by proposing a structured, multi-dimensional framework for ethical governance of GAI in higher education. The findings highlight the need for coordinated institutional strategies to ensure the responsible, sustainable integration of GAI into academic environments.