Governance of artificial intelligence: A risk and guideline-based integrative framework

16 Jul 2025 · MIT AI Risk Repository – Blog Global

A public-sector-focused AI risk taxonomy may offer APS governance practitioners a structured vocabulary for risk identification and policy cycle work.

Key points

Summary

The MIT AI Risk Repository has spotlighted a 2022 paper by Wirtz, Weyerer, and Kehl published in Government Information Quarterly, which presents a six-category AI risk taxonomy tailored to public sector organisations. The taxonomy covers technological, informational, economic, social, ethical, and legal/regulatory risks, and links each category to governance guidelines through a four-layer conceptual model and a seven-stage policy cycle. The framework is specifically designed to address public sector concerns such as democratic legitimacy, citizen trust, and public value creation. The item itself is a blog summary rather than new research.

Implications for Australian agencies

Implications are AI-generated. Starting points, not advice.