Hallucination
Definition
When an AI model produces information that sounds confident and plausible but is false or made up.
In plain English
Language models generate likely-sounding text. They don’t check facts as they write, so when they lack the right information they can fill the gap with something that merely sounds right.
Grounding answers in real documents with RAG, and showing sources, are common ways to reduce hallucinations. Checking important claims yourself is still wise.
Example
Asked for sources, a chatbot lists three research papers with real-sounding titles and authors, but none of the papers exist.
Draft, not yet reviewed by a person · Updated Nov 2026