What is an "AI hallucination," and why does it happen?

A hallucination is when an AI confidently states something that isn't true — a fake citation, a wrong date, an invented statistic. It happens because LLMs don't "know" facts the way a database does; they predict plausible-sounding text. When the model has weak signal on a topic, it generates what sounds right instead of admitting it doesn't know.

Modern models hallucinate far less than the early ones, especially when they can search the web or cite sources. But they can still produce confident nonsense. The rule of thumb: never publish or act on AI output for anything important without verifying the facts yourself.