The principle "Key requirements for trustworthy AI" has mentioned the topic "verification" in the following places: