The Dark Sides of Artificial Intelligence: An Integrated AI Governance Framework for Public Administration
A structured academic governance framework for public administration AI risks - useful as a reference point, though not a recent development.
Key points
- A 2020 academic framework organises AI governance for public administration into law, society, and ethics categories.
- The five-layer governance structure and four-stage regulatory process offer a structured reference for APS governance work.
- The paper is five years old and spotlighted as part of MIT's AI Risk Repository catalogue - not new research.
Summary
The MIT AI Risk Repository has spotlighted a 2020 academic paper by Wirtz, Weyerer, and Sturm proposing an integrated AI governance framework for public administration. The framework organises AI challenges into three categories - law and regulation, societal impacts, and ethics - and presents a five-layer governance structure alongside a four-stage regulatory process covering framing, risk and benefit assessment, risk evaluation, and risk management. The paper is grounded in regulation theory, treating AI challenges as market failures requiring governmental intervention. APS practitioners may find the structured taxonomy useful for gap analysis against existing Australian frameworks, though the work predates several significant AI governance developments.
Implications for Australian agencies
- Consider Governance and policy teams could consider whether the five-layer framework or risk taxonomy offers a useful diagnostic lens when reviewing agency AI governance arrangements against the APS AI Policy.
- Monitor Teams building or maintaining AI risk registers may want to monitor the broader MIT AI Risk Repository catalogue for other frameworks that could inform Australian risk classification approaches.
Implications are AI-generated. Starting points, not advice.
"The Dark Sides of Artificial Intelligence: An Integrated AI Governance Framework for Public Administration" Source: MIT AI Risk Repository – Blog Published: 16 July 2025 URL: https://airisk.mit.edu/blog/the-dark-sides-of-artificial-intelligence-an-integrated-ai-governance-framework-for-public-administration The MIT AI Risk Repository has spotlighted a 2020 academic paper by Wirtz, Weyerer, and Sturm proposing an integrated AI governance framework for public administration. The framework organises AI challenges into three categories - law and regulation, societal impacts, and ethics - and presents a five-layer governance structure alongside a four-stage regulatory process covering framing, risk and benefit assessment, risk evaluation, and risk management. The paper is grounded in regulation theory, treating AI challenges as market failures requiring governmental intervention. APS practitioners may find the structured taxonomy useful for gap analysis against existing Australian frameworks, though the work predates several significant AI governance developments. Implications for Australian agencies: - [Consider] Governance and policy teams could consider whether the five-layer framework or risk taxonomy offers a useful diagnostic lens when reviewing agency AI governance arrangements against the APS AI Policy. - [Monitor] Teams building or maintaining AI risk registers may want to monitor the broader MIT AI Risk Repository catalogue for other frameworks that could inform Australian risk classification approaches. Retrieved from SIMS, 18 May 2026.