Data underpinning AI must be the correct data to rely on in the circumstance and must be accurate. Relying on the correct data means considering completeness and relevance, exclusions required for legal, intellectual property, ethical or regulatory reasons, how old the data is and whether that matters, clarity of attribute definitions, units and sources, adequacy for those the data relates to, and freedom from bias especially in historical data. Accuracy means duplication is avoided, labels are correct, values are right and identifiers are consistent for the same subject. A structured process with clear accountability has suitably skilled senior people satisfy themselves continuously about data quality, and senior leaders and management boards require quality assurance through independent reviews so that issues are understood and acted on quickly.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.