"Hallucination" is the term for when an AI model states something false with the same fluent confidence it uses for something true. It's arguably the most important AI limitation to understand, because the output looks identical whether it's right or wrong.
- Hallucination is when a model states something false with the same fluent confidence as something true.
- Specific, verifiable details (dates, citations, stats) are the highest-risk category.
- Ask for sources and independently verify anything you'll actually rely on.
Why it happens
A language model is fundamentally predicting plausible next words based on patterns in its training data — it doesn't have a built-in mechanism for checking facts against reality. When it doesn't actually know something, it can still generate a fluent, plausible-sounding answer instead of admitting uncertainty.
Where it happens most
Specific, verifiable details — dates, citations, statistics, names of niche sources — are the highest-risk category. General reasoning and well-known facts are comparatively far more reliable.
How to guard against it
For anything you'll actually rely on, ask the model to cite where a claim comes from, cross-check specific facts independently, and be especially skeptical of suspiciously precise numbers or quotes that sound exactly right.
Frequently Asked Questions
What is an AI hallucination?
When a language model generates a fluent, confident-sounding statement that is factually false — the output looks identical whether it's right or wrong.
How can I catch a hallucination?
Ask the model to cite its source, independently cross-check specific facts, and be especially skeptical of suspiciously precise numbers or quotes.