The principle "Social Principles of Human-centric AI (Draft)" has mentioned the topic "safety" in the following places: