Throughout their operational lifetime AI applications do not compromise people's physical safety or mental integrity; unintended risks of physical, emotional and environmental harm are minimised, with attention to the growing risk of unanticipated behaviour as capability and autonomy rise, and tools that augment human decisions are safe, trustworthy and in keeping with the ethics and wishes of the people they affect.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.