A Governance Framework for Responsible AI Integration in Higher Education: A Fuzzy Delphi Study
TL;DR
It is suggested that sustainable AI adoption requires institutions to first establish robust governance structures, ensure data readiness and embed ethical safeguards before expanding technological investments, with particular relevance for resource-constrained higher education contexts in Malaysia and the broader Global South.
Abstract
The rapid integration of artificial intelligence (AI) in higher education presents both transformative opportunities and major governance challenges. Many institutions still lack structured mechanisms to ensure responsible and coordinated implementation. This study identifies and prioritizes key design principles for institution-wide AI integration in Malaysian higher education through the Fuzzy Delphi Method (FDM). A panel of twelve experts assessed proposed principles organized across five socio-technical domains: Governance and Strategy, Technology and Infrastructure, Data and Analytics, Ethics, Risk and Compliance and Change Management and Workforce Enablement. Consensus was established using a threshold value of d ≤ 0.2 with a minimum agreement rate of 75%. All five domains achieved consensus, with Governance and Strategy emerging as the highest priority, followed by Data and Analytics, Ethics, Risk and Compliance, Technology and Infrastructure and Change Management and Workforce Enablement. These findings suggest that sustainable AI adoption requires institutions to first establish robust governance structures, ensure data readiness and embed ethical safeguards before expanding technological investments. This study contributes a validated, prioritized socio-technical framework offering evidence-based guidance for university leaders, with particular relevance for resource-constrained higher education contexts in Malaysia and the broader Global South.