Organisations are able to explain the decision-making of their AI applications clearly and comprehensibly. This is a performance principle on which the others rest. Explainability means human-understandable accounts of a model's characteristics and of the pathway to each prediction, built in where the model allows it (decision trees, linear models) or produced after the fact through examples, surrogate models, feature importance and partial dependence, with hyper-parameters of complex models documented and the explainability-accuracy trade-off and the safety-critical nature of the use weighed at design time. Transparency means clear, honest communication with stakeholders rather than disclosure of intellectual property: traceability of the engineering process, verifiability of normative decisions so that harm can be explained and audited, no deception by generated content outside justified contexts, and people's access to and correction of their own data. Provability applies mathematical certainty to safety-critical or regulator-approved applications. Project teams translate decision logic into plain language and graphics for non-technical audiences and make explicit how factors and data determine outcomes.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.