Trustworthy AI: Deciding What to Decide