Commonwealth responsible AI policy and national assurance framework

Instrument

Two linked Commonwealth policy layers:

1. Policy for the responsible use of AI in government (Digital Transformation Agency), Version 2.0, effective 15 December 2025 for applicable non-corporate Commonwealth entities.
2. National framework for the assurance of artificial intelligence in government, agreed at the Data and Digital Ministers Meeting on 21 June 2024 and published by the Department of Finance — a joint Australian, state and territory principles-based approach to AI assurance.

DTA also publishes a Pilot AI assurance framework (draft) that DTA states explores mechanisms to support Commonwealth implementation of the nationally agreed principles. Treat pilot materials as draft unless superseded by final policy.

Type and status

Government policy and nationally agreed assurance framework — not statute. Mandatory requirements in the responsible AI policy apply only to in-scope Commonwealth entities as stated on digital.gov.au.

Applicability and dates

Read DTA implementation guidance for entity exceptions. State and territory agencies should confirm how the national assurance framework applies alongside their own policies. Do not infer private-sector obligations from this briefing.

Practical institutional implications

The responsible AI policy (v2.0) sets mandatory expectations for in-scope Commonwealth entities across areas such as accountable officials, transparency statements, strategic approach to adoption, operationalising responsible use, use-case accountability, internal registers, staff training and AI use-case impact assessment (per the policy landing page).

The national assurance framework establishes cornerstones and practices for applying Australia's AI Ethics Principles to government AI assurance. DTA's pilot assurance framework is positioned as complementary to existing law and policy, not a duplicate.

Institutions should maintain a clear line between policy compliance artefacts (registers, impact assessments, transparency statements) and assurance of institutional work — whether authority, evidence and accountable judgement remain intact through real operations.

DataMPowered interpretation

DataMPowered separates assurance of models and technology from assurance of AI-supported institutional work. Commonwealth policy and national assurance frameworks are necessary context for public-sector AI governance; they do not, by themselves, validate any vendor product or DataMPowered platform. See AI Governance & Assurance.

Primary sources

Cite the DTA policy PDF and Finance framework pages linked in the frontmatter. Re-check digital.gov.au and finance.gov.au before board or audit reporting.

Related reading

• Preparing for Australia's automated decision-making transparency requirements for APP-layer transparency themes that may intersect with AI use-case transparency expectations.