Skip to content
BENIBIX Learn

Hallucination (AI)

AI💬 LLMsLanguage Model Limits

Also known as AI Hallucination, Confabulation

When an AI states something false as if it were true.

Think of it like

Think of it like a kid who does not know the answer. They make one up and say it with a big smile.

Example

Jack asks for three studies on sleep and coffee. The model lists three with authors and years, and one of them was never written.

Why it matters

People trust smooth, sure text. A made-up fact looks just like a real one until someone checks.

Definition

A hallucination is when a language model states something false as if it were true. It can be a made-up fact, a wrong date, a book that does not exist or a fake quote. It happens because the model writes likely-sounding text, not checked text. The false part reads just as smoothly as the true parts. Giving the model real sources and checking its claims cuts the risk, but does not remove it.
One sentence in this smooth answer is made up. Nothing in its tone gives it away.

Common mistake

Thinking a confident tone means it is right. The model sounds just as sure when it is wrong.

Key takeaways

  1. 1
    False but fluent. Reads like a fact.
  2. 2
    Likely, not checked. Why it happens.
  3. 3
    Check what matters. Sources cut the risk.

Learn it properly