AI governance for Australian government and public sector.

Government agencies are bound by the Australian Privacy Principles. AI systems that make or inform decisions about individuals must comply with collection, use, and disclosure limits. Privacy Impact Assessments are expected for high-risk AI.

The statutory obligation to notify individuals when automated systems make or significantly contribute to decisions affecting them takes effect December 2026. Government agencies must identify and document all automated decision-making ahead of this date.

The Australian Government AI6 framework (October 2025) applies to all Australian Government entities and sets expectations for transparency, fairness, reliability, privacy and security, contestability, and accountability in AI use.

Documents relating to AI systems used in government decisions — including model documentation, training data descriptions, and decision rationale outputs — may be subject to FOI requests. Agencies must be prepared to produce these.

AI-assisted decisions that affect individual rights must satisfy natural justice and procedural fairness requirements. Fully automated adverse decisions without human oversight or explanation pathways create administrative law risk.

Government AI procurement must align with the Digital Transformation Agency's AI procurement guidance. Vendors supplying AI to government face increasing contract requirements around auditability, bias testing, and governance documentation.