Mapping Frameworks at the Intersection of AI Safety and Traditional Risk Management
A structured evidence scan of AI risk frameworks gives APS governance practitioners a consolidated starting point for assessing and strengthening agency risk approaches.
Key points
- MIT AI Risk Repository maps 11 frameworks bridging traditional risk management and AI safety, all published 2023 or later.
- Frameworks span maturity models, probabilistic risk assessment, and cybersecurity adaptations useful for agency AI governance work.
- UK DSIT's 'Emerging Processes for Frontier AI Safety' is among the 11 - a directly accessible government reference.
Implications for Australian agencies
- Consider APS agencies developing or reviewing AI risk frameworks could assess this evidence scan as a consolidated reference to avoid duplicating work already done internationally.
- Consider Risk and assurance teams may want to consider whether maturity model frameworks - particularly those based on NIST AI RMF - are applicable for benchmarking agency AI risk management capability.
- Monitor Policy teams could monitor whether any of these frameworks are adopted or cited by DISR, DTA, or AISI as the Australian AI governance landscape matures.
Implications are AI-generated. Starting points, not advice — see methodology for how they're framed.
View original source
Copied.
"Mapping Frameworks at the Intersection of AI Safety and Traditional Risk Management"
Source: MIT AI Risk Repository – Blog
Published: 8 April 2025
URL: https://airisk.mit.edu/blog/mapping-frameworks-at-the-intersection-of-ai-safety-and-traditional-risk-management
MIT's AI Risk Repository has published an evidence scan identifying 11 frameworks that bridge traditional risk management disciplines - such as aviation, nuclear, and cybersecurity - with AI safety considerations for advanced AI systems. The scan categorises frameworks into four types: risk management translation (adapting established methods to AI), maturity models (scoring organisational readiness), novel approaches (methods designed specifically for frontier AI), and emerging practice (documenting real-world safety practices). Primary authors are from the UK, US, Singapore, Germany, Finland, and France. The full document set is publicly available via a Paperpile folder, making it a practical reference for agencies developing or reviewing AI risk frameworks.
Implications for Australian agencies:
- [Consider] APS agencies developing or reviewing AI risk frameworks could assess this evidence scan as a consolidated reference to avoid duplicating work already done internationally.
- [Consider] Risk and assurance teams may want to consider whether maturity model frameworks - particularly those based on NIST AI RMF - are applicable for benchmarking agency AI risk management capability.
- [Monitor] Policy teams could monitor whether any of these frameworks are adopted or cited by DISR, DTA, or AISI as the Australian AI governance landscape matures.
Retrieved from SIMS, 18 July 2026.