How to know when NOT to trust AI
You don't need to fear AI, and you don't need to believe it. You need to know which moments call for which.
The lesson
Here's the calibration: the more it costs to be wrong, the less you should trust without checking. Brainstorms and drafts? Trust freely, mistakes are cheap. Facts, figures, medical, legal, money? That's where you slow down and verify. Skepticism isn't fear. It's just knowing where the stakes are.
Try it — 60 seconds
What part of your last answer are you least sure about — and how would I double-check it?
Copy it into ChatGPT, Claude, or Gemini — whichever you already use. Fill in the brackets with your real life.
Watch out
A confident tone is part of the product design — it's not evidence. The stakes, not the certainty, decide when you verify.