The principle "Draft Ethics Guidelines for Trustworthy AI" has mentioned the topic "safety" in the following places: