Hallucination: confidently wrong answers

A "hallucination" is when an AI tool generates something that sounds entirely plausible and is stated with total confidence, but is factually wrong or entirely made up: a fabricated legal case, a citation to a report that doesn't exist, a quote nobody said, a statistic invented on the spot.

Why this happens

The model is generating the next most plausible piece of text, not retrieving a verified fact from a database. If a real answer isn't well represented in what it learned, it will often still generate something that reads like a confident, correct answer, because that's what the pattern of a confident, correct answer looks like.

What this means for you

Never submit AI-generated facts, figures, citations, or quotes into anything that matters, a client document, a legal filing, a report, without independently verifying them first. The tool's tone of confidence tells you nothing about whether it's actually correct.