Source of Law / Policy · National Framework stream
National Framework for Assurance of AI in Gov
Owner: Dept of Finance · Framework (non-binding)
GUIDANCEUpdated 2 Sept 2026
Agreed by Data and Digital Ministers on 21 June 2024, the National Framework provides a joint approach to safe and responsible AI across all three tiers of Australian government. It maps to Australia's eight AI Ethics Principles and sets out five cornerstones and a set of assurance practices. It is guidance, not legislation: it takes legal force only when a jurisdiction adopts it into its own policy (as the Commonwealth has through DTA Policy v2.0) or writes it into a contract.
Key obligations
- Adopt a risk-based AI governance regime aligned with the eight AI Ethics Principles.
- Maintain an AI use-case register and assign accountable owners.
- Apply assurance practices proportionate to risk: impact assessments, testing, monitoring, incident reporting.
- Embed transparency and contestability — publish what AI is in use and how decisions can be challenged.
- Coordinate with cross-jurisdictional reporting and lessons-learned channels.
Key dates
- 21 Jun 2024Framework agreed and published by Data and Digital Ministers.
- Jul 2026Framework pages last refreshed by the Department of Finance — no cross-jurisdictional compliance deadline has been set.
- WatchThe Government has signalled it intends to legislate standards for high-risk AI uses; nothing had been introduced as at September 2026.
Who is affected
- Commonwealth, state and territory government agencies — each through its own adopting policy.
- Suppliers across all three tiers — expect aligned but not identical procurement requirements.
- Non-government bodies receiving public funding to deliver services may be drawn in via contract.