Indicators are reported separately for each component (verification, estimation, inference, binding, delivery) and at system level as the aggregate outcome allowing for integration, configuration and contexts, with differing component levels specified in the schedule. They disclose demographic coverage (gender; skin tone; age band; accessibility need; environmental condition) where relevant; outcome error parity is assessed and any disparity between groups stated in both the audit report and the schedule; and indicators state whether results hold globally or only in specified jurisdictions or contexts. Where a local regulator or a national standards body sets additional indicators (fairness thresholds, sector thresholds, fallback or successive validation requirements) these are reported alongside the global ones, marked as jurisdiction-specific and recorded on the certificate and schedule. The standardised reporting table gives the component or system name and version, the level achieved, accuracy and error rates with confidence or credible bounds, sample size and demographic coverage, applicable contexts and jurisdictions, local indicators, and limitations, assumptions and exclusions.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.