AI doesn't know things.
It guesses them.
That's the whole story. But let's break it down with some hands-on demos so it really clicks.
↓
What is a hallucination?
When AI makes things up with total confidence
A hallucination is when an AI states something false as if it were a fact. Not a guess marked as a guess — a confident, detailed, completely wrong answer. Here's a real example:
AI OUTPUT
Reality: This paper does not exist. The authors, the journal, the DOI — all fabricated. The AI strung together plausible-sounding academic language because that's what the pattern suggested. It didn't lie — it had no concept of truth to begin with.
Interactive
Be the AI: Guess the next word
AI doesn't "read and understand" — it predicts the most likely next word based on patterns. Try it yourself. You'll see how easy it is to guess wrong when the context is ambiguous.
Pick the word that most likely comes next.
The 6 reasons
Why does AI hallucinate?
Tap each card to flip it and read the explanation.
Interactive
Spot the hallucination
Below are 6 statements. Some are real facts. Some are AI hallucinations. Can you tell which is which? Tap TRUE or HALLUCINATION for each.
Interactive
The creativity dial
AI models have a setting called temperature that controls how random their word choices are. Slide it and watch what happens to the output.
Cold (precise)
Hot (creative)
PROMPT: "The first person on the moon was"
Neil Armstrong, in 1969.
At low temperature, the AI picks the most likely next word. Safe and accurate.
Protect yourself
How to not get fooled
Tap to check off each tip. Build the habit.