The principle "G20 AI Principles" has mentioned the topic "safety" in the following places: