Managing AI Risk in Higher Education: Compliance versus Design

Artificial intelligence (AI) is deploying across higher education, yet current governance approaches rely on compliance-based frameworks that treat AI challenges as calculable risks rather than unpredictable uncertainties. This mismatch has contributed to the proliferation of ineffective or counterproductive AI implementations. While compliance-based governance remains necessary, particularly for calculable risks such as privacy, it is not sufficient for the uncertainties that AI introduces into teaching and learning. Material and Methods: This paper synthesises insights from organization theory, cognitive science, and socio-technical systems analysis to evaluate modes of AI governance in higher education. To complement compliance-oriented governance, the NIST AI Risk Management Framework (AI-RMF) is enriched with cognitive and socio-technical perspectives. Results: Five faculty AI literacies are proposed—System-Perception Awareness, Cognitive-Route and Output Evaluation, Values-Anchored Practice, Human-Centred Agency, and Governance-and-Context—each mapped to AI-RMF functions (Govern, Map, Measure, Manage). These literacies reposition faculty development from operational tool training toward adaptive judgement under uncertainty. The framework is a design-based complement to, rather than a replacement for, compliance-based governance. Discussion: The framework addresses gaps in existing AI literacy models, which consistently under-represent governance reasoning, sociocultural orientation, and human agency. By integrating cognitive, socio-technical, and governance dimensions, the framework enables educators to recognise when AI reshapes reasoning and practice and to respond with principled, context-sensitive judgement. Conclusions: Faculty AI literacy should be understood as cognitive, pedagogical, and governance capacities. Empirical validation through faculty development interventions, longitudinal studies, and institutional case studies is needed to test the framework’s effects on teaching quality and student outcomes.