Every AI tool used in the public service sits inside a process with human oversight built in, whether a human in the loop, on the loop or in command; the more autonomous the system, the stricter the oversight and testing. Where AI feeds a decision in a high-risk setting, the final call rests with a person; material generated by AI is always critically reviewed by a person for sense and accuracy; and instructions given to AI systems are well designed, since poor instructions produce poor outcomes.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.