MAF paragraphs 3.9 to 3.13: before deploying, decide the commercial objective of using AI and weigh it against the risks to individuals, groups and the organisation (including group-level effects such as herding and variation across countries' norms), guided by corporate values; document any deliberate deviation from those values with its rationale; treat this as iterative, keep identifying and mitigating risks with a response plan if mitigation fails, and record it in a periodically reviewed risk impact assessment. Where ethical AI costs outweigh benefits, consider non-AI alternatives (para 2.6).
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.