Operationalising AI governance in higher education: A stepwise policy-development method

Area of Science:

• Higher Education
• Artificial Intelligence Governance
• Policy Development

Background:

• Generative AI adoption necessitates robust governance frameworks for institutional implementation.
• Existing frameworks often lack operational guidance for specific AI use scenarios.
• There is a need for methods to translate AI risks into actionable institutional policies.

Purpose of the Study:

• To present the Higher Education Artificial Intelligence Policy Development Method (HE-AIPD) for institutional AI policy creation.
• To operationalize AI governance by converting use-specific risks into implementable decisions and safeguards.
• To provide a structured method for developing reusable AI governance packages.

Main Methods:

• The HE-AIPD is a stepwise method for cross-functional teams to develop institutional AI-use policies.
• It converts AI-use evidence into a governance package including decision rules, safeguards, clauses, responsibilities, tools, and review cycles.
• A traceability design links use cases to governance components, with content validity assessed via expert walkthrough (S-CVI/Ave = 0.882).

Main Results:

• The HE-AIPD operationalizes institutional AI governance by translating AI-use scenarios into concrete outputs.
• Reusable governance outputs include a scope sheet, use-case inventory, decision matrix, risk-safeguard map, clause bank, implementation toolkit, and review cycle.
• Content validity was confirmed with an S-CVI/Ave of 0.882.

Conclusions:

• The HE-AIPD provides a practical, stepwise method for institutions to govern generative AI adoption.
• It enables the creation of tailored, reusable AI governance packages, enhancing policy development efficiency.
• The method facilitates the systematic management of AI-related risks within higher education settings.