Critical Thinking and Judgment with AI: evaluating outputs, spotting plausible errors, automation bias, when to use AI at all, calibrating trust to risk, what stays human, accountability