Hallucination
When a model states something false with full confidence, because it is predicting plausible-sounding language, not looking up facts.
It is the number-one risk in real use. Anything factual from an AI needs verification.
Frequently asked questions
Why does AI make things up?
Because a model predicts plausible-sounding language rather than looking facts up, so when it does not truly know, it can still produce a confident, wrong answer. This is called a hallucination, and it is the number-one risk in real use.
How do I stop an AI from giving me false information?
You cannot fully stop it, so the practical rule is to verify anything factual before you rely on it, especially names, numbers, dates, and quotes. Techniques like grounding and RAG, which tie answers to real sources, also reduce how often it happens.
Can I tell when the AI is hallucinating?
Not always, because the false statement usually sounds just as confident and fluent as a correct one. That is exactly why anything important needs checking against a real source rather than trusting the tone.
New to all this? Start with what an AI agent really is, browse the full glossary, or explore the learning hub.
← Back to the glossary