Where suppliers use commercially available or open source applications, tools, libraries, frameworks and data, they are required to develop compelling arguments that confidence in the resulting AI is not undermined. The provenance of all data used to train and test the AI, including its legal and ethical compliance with MOD policy, is assured and documented so that data is fit for purpose, free of adversarially introduced vulnerabilities and compliant, covering both MOD-supplied and third-party data.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.