Australia Makes AI Risk Assessments Mandatory for High-Risk Government Uses
Australia made AI risk assessments mandatory for high-risk government uses, relying on existing sectoral regulators rather than a dedicated AI authority — a distributed approach that will test whether existing institutions can absorb the new mandate.
Australia's Department of Industry published mandatory guardrails for AI use by government agencies and regulated entities in high-risk settings in October 2025, covering healthcare, financial services, and employment decisions. The guardrails require documented risk assessments, human oversight for consequential decisions, and public disclosure when AI is used in government services that affect individual rights.
The framework applies to agencies and regulated private sector entities, but enforcement of the private sector obligations relies on existing regulators, including the Australian Securities and Investments Commission and the Australian Communications and Media Authority, rather than a dedicated AI regulator.
Why it matters: Mandatory risk assessments for government AI use are a meaningful step, but the decision to rely on existing sectoral regulators rather than establish a dedicated AI authority reflects a broader debate about whether AI governance fits the existing regulatory architecture or requires new institutions. Australia's experience with this distributed approach will be closely watched by other countries facing the same structural question.
