Definition
A formal policy document — issued by a government body — specifying how autonomous AI agents may be used in public-sector settings, including mandatory oversight models, human-review thresholds, transparency requirements, risk classification, and kill-switch obligations. Australia's DTA playbook (June 2026) is among the first such government-issued technical governance frameworks specifically targeting agentic AI rather than chatbots.
Why it matters
Government suppliers and contractors face new compliance obligations as these frameworks are enacted; they also signal the direction of private-sector regulation. Organisations building agentic AI products for or with governments need to assess alignment with these frameworks now.