😵‍💫 Why does AI make things up? (Hallucinations)

When AI confidently states something false, we call it a 'hallucination'. But the model isn't lying — lying requires intent. It's doing the only thing it knows how to do: predict the most likely next words. Sometimes the most likely-sounding answer is factually wrong.

Think of a very confident person who always has an answer. Ask them something obscure and they'll still answer, fluently, because that's what they've learned to do. The AI doesn't know what it doesn't know — it just produces the most probable-sounding response.

This matters because AI sounds authoritative. A hallucinated fact comes wrapped in the same confident prose as a true one, so you can't tell by tone alone. It will invent references, dates, and quotes that don't exist.

The fix is a habit, not a feature: treat AI output as a first draft and verify anything that matters. For facts, numbers, citations and anything you'll act on — check a real source. Use AI for the thinking and the drafting; use your own judgement (and Google) for the facts.