Editing Ethical Prompting & Judgment (section)

== Common mistakes ==
* '''"I always double-check."''' Intention isn't process. Without a consistent method, you check when you remember and skip when you're busy. That's when errors get through.
* '''Over-verifying low-stakes output.''' Not everything needs the same scrutiny. Brainstorming ideas don't need fact-checking; a client-facing report does. Calibrating effort to risk is itself a judgment skill.
* '''Trusting AI's self-assessment.''' When you ask AI "are you sure?" it will often say yes — or hedge unconvincingly. The Fact-Check Habit exercise teaches you to use structured self-audit prompts that produce genuinely useful uncertainty signals.