AI systems should work robustly, securely and safely across their whole life cycle, with risks continually identified, assessed and managed. Regulators are to judge how likely AI is to create safety risks in their sector and respond proportionately, may require regulated firms to test, or check through due diligence, how reliably and securely their systems perform, may point to NCSC principles for securing machine learning and to technical standards, and should coordinate their guidance. The February 2024 guidance asks regulators to communicate the level of safety risk in their remit, stress safety risk assessment and mitigation by developers and deployers, and address cyber resilience throughout the life cycle.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.