There's no dedicated legal framework for agentic AI under UK law. It seems the government prefers adapting existing legislation to creating new AI-specific laws. The pressure to innovate is real but the government must consider if agentic AI is the right system for the job, and its governance role.

AI agents are tools and agentic AI is the system that uses those tools to think, decide and act independent of human oversight. The government must address any risks like: data leaks, hallucinations, compliance failures, misleading content, reinforcing systemic inequities and distorting decision-making. While protecting democratic integrity, economic stability, human relationships, and societal well-being. As capabilities grow, so does the need for careful governance, security and oversight, for any use of agentic AI in critical National Infrastructure, Public Services, National Defence and highly regulated environments. The government should review how to address the aforementioned risks and ensure compliance is built in from the start.