AI Risk Repository Report updated (April 2025)
A maintained, cross-framework AI risk taxonomy gives APS governance practitioners a structured reference for risk identification and audit design.
Key points
- MIT's AI Risk Repository updated to 1,612 unique risk entries across 65 frameworks, now including multi-agent risks.
- The repository provides causal and domain taxonomies designed to support policy, auditing, and governance processes.
- A credible reference resource for APS agencies developing AI risk frameworks or audit criteria - freely accessible.
Implications for Australian agencies
- Consider APS agencies developing AI risk frameworks or assurance processes could assess whether the repository's taxonomies offer useful structure for categorising agency-specific AI risks.
- Monitor Policy teams working on AI governance standards may want to monitor future updates, particularly as the multi-agent risk subdomain develops.
Implications are AI-generated. Starting points, not advice — see methodology for how they're framed.
View original source
Copied.
"AI Risk Repository Report updated (April 2025)"
Source: MIT AI Risk Repository – Blog
Published: 23 April 2025
URL: https://airisk.mit.edu/blog/new-version-of-the-ai-risk-repository-preprint-now-available
MIT's AI Risk Repository has released an April 2025 update to its preprint and database, expanding coverage to 65 AI risk frameworks and 1,612 coded risk entries. New additions include a dedicated subdomain on multi-agent risks, reflecting developments in agentic AI research. The repository offers two complementary taxonomies - causal and domain-based - intended to support risk identification, auditing, policy development, and safety research. It is publicly accessible and positioned as a living reference for researchers, policymakers, and auditors seeking a consolidated view of AI risk classification across academic, industry, and policy literature.
Implications for Australian agencies:
- [Consider] APS agencies developing AI risk frameworks or assurance processes could assess whether the repository's taxonomies offer useful structure for categorising agency-specific AI risks.
- [Monitor] Policy teams working on AI governance standards may want to monitor future updates, particularly as the multi-agent risk subdomain develops.
Retrieved from SIMS, 18 July 2026.