The principle "Four principles of AI ethics" has mentioned the topic "controllability" in the following places: