The principle "Draft AI Utilization Principles" has mentioned the topic "safety" in the following places: