The OECD has established a Working Group on Agentic AI in Government under the Working Party of Senior Digital Government Officials (E-Leaders) to define how governments should deploy autonomous AI systems that act on their behalf1. The group draws on national frameworks already in development: Canada's guidance on decision boundaries and accountability, Australia's agent-specific technical standards emphasizing audit logs and rollback capability, Singapore's four-dimension governance framework, and Israel's government-wide architecture for AI agents built on open standards. Estonia and Ukraine are already piloting citizen-facing agentic systems, including Bürokratt and Diia.AI.
OECD Forms Working Group on Agentic AI in Government
The OECD established a Working Group on Agentic AI in Government to guide how autonomous AI systems operate in public services across member nations.