The principle "Artificial Intelligence Application Criteria" has mentioned the topic "safety" in the following places: