Members and firms must apply professional judgement (knowledge, skills, experience and professional scepticism) to decide whether any AI output with a material impact is reliable, and record that decision in writing: the assumptions made, key concerns about reliability including of underlying datasets, why each concern arises, whether it can be lessened, and its effect on overall reliability with a conclusion on whether the output can reasonably be used; the decision is made or supervised by a suitably qualified surveyor, named, who accepts responsibility, and an output judged unusable must be reported in writing to the client with the reasoning or a summary.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.