Data provenance is assured and recorded from collection to ingestion by the AI. All data and metadata that influence the AI's output are kept under configuration management commensurate with that of the AI, with attention to classification, complexity, volume and velocity. Synthetic data that supplements or replaces real data is validated from the AI's perspective, because AI may cue on different features than a human.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.