Give relevant stakeholders, directly or through AI providers, timely and appropriate information about the AI systems the developer builds, for example: the possibility that outputs or programs change through learning; safety information including technical characteristics, safety mechanisms, foreseeable risks of use and their mitigation; the intended scope of use set by the developer to avoid harm from unforeseen provision or use; the operating status, causes of failures and response status; the content and reasons of updates; and the collection policy for data the model learns from, the learning method and the implementation structure.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.