Sometimes an AI tells you something completely false — a fake book, a made-up quote — with total confidence. People call this “hallucinating.” It sounds like a glitch. It’s actually the same machinery working normally.

Remember: the model is guessing what words usually come next. It is not looking anything up in a list of true facts. So if you ask for, say, a quote by a famous person, it produces words that sound exactly like such a quote — whether or not that quote ever existed. A plausible-sounding fake and a real fact look identical to a next-word guesser.

That’s why it can be fluent and wrong at the same time. Fluency comes from pattern-matching; truth doesn’t, unless the true pattern happened to be common in what it read.

The practical takeaway: treat AI like a very well-read friend who never says “I’m not sure.” Brilliant for drafts and explanations. Always check it before you trust it on facts.