The principle "Microsoft AI Principles" has mentioned the topic "controllability" in the following places: