Generative AI & LLMs
Hallucination
An AI answer that presents made-up or unsupported information as fact.
Example
A chatbot supplies a book title that does not exist.
Why people use it
The term helps explain why a believable AI answer can still be false.
What you'll hear
“The assistant invented that reference. That's a hallucination.”
What this means for you
Check names, numbers and citations against original sources.
Can you control it?
Sometimes
Sometimes. Your choices depend on the tool and your access. The settings available to an everyday user may differ from those available to the people running it.
Common questions
- Can an answer contain both correct and made-up details?
- Yes. A mostly accurate reply can include an invented date, quotation or explanation among genuine facts.
- Is it the same as lying?
- Not necessarily. A false answer alone does not show that the AI knew the truth and chose to deceive.
- Can the same question get a correct answer later?
- Yes. Answers can vary between attempts, so one correct reply does not make earlier mistakes disappear.