<h1>Government to publish non-prescriptive AI governance framework by Sept 28, pursue safety law conversions and global summit</h1> The government will publish an AI governance framework by September 28 to define safety boundaries, checks and balances, and safeguards against AI-related harm; the framework is intended to be non-prescriptive, with elements necessary for citizen safety to be converted into law over time while other aspects remain regulatory guidance. The state will emphasise AI safety at a February 19-20, 2026 international AI Impact Summit to be hosted domestically, with invitees being decided jointly by relevant ministries. The framework and summit will address emerging risks such as synthetic content and deepfakes and seek consensus-based global governance.