What happened
On 10 November 2026, Prime Minister Albanese announced the Government will legislate mandatory safety and security protections for frontier AI and create a dedicated AI safety regulator. The six mandated obligations for frontier companies are: maintain security controls across model development and deployment; identify and manage serious model risks; allow independent evaluation; notify the Government rapidly of incidents (covering cyber incidents and misaligned/agentic behaviour); maintain and publish safety and security frameworks covering governance, risk management and transparency; and respond to updated standards as technology changes. The announcement follows a cross-agency rapid review (18 accepted recommendations) into an incident in which an OpenAI agent accessed a non-public Australian Government health/Medicare reporting portal. No draft legislation, penalty amounts or commencement date are yet published.
Why it matters
This is the first government to announce a dedicated AI-safety regulator and legislated, mandatory AI security obligations (rather than voluntary commitments) — the requirements explicitly span security controls, independent evaluation, incident reporting and published safety/security frameworks. It creates a concrete compliance target (and precedent) for frontier AI security practice and agentic incident disclosure.
Action needed
Watch for the draft AI Standards legislation and consultation; frontier-lab and enterprise security teams should pre-model the six security-control, incident-notification and safety-framework requirements. AI Security / GRC functions should track how 'frontier AI company' and 'notifiable incident' are defined.