SESKStuff Everyone Should Know
SESK: Stuff Everyone Should Know · sesk.info/ai-hallucination-definition

In generative AI, what is a hallucination?

Plausible-seeming but false output stated as fact about the real world.

Start playing free

  1. A picture the model draws on purpose
  2. A crash that stops the model
  3. Confident, plausible-sounding but false output ✔

Why: A hallucination is the model asserting something untrue as if it were real, like claiming a president died in the wrong century. It sounds right, which is exactly why it is dangerous.

Huh, didn't know that: Google's glossary example: a model that claims Barack Obama died in 1865 is hallucinating.

The dad joke
My chatbot told me I invented the sandwich. I'm not correcting it.
Like this card?
Play the game

References

Answer: Confident, plausible-sounding but false output. A hallucination is the model asserting something untrue as if it were real, like claiming a president died in the wrong century. It sounds right, which is exactly why it is dangerous.

  1. Google Machine Learning Glossary: Generative AI (hallucination) developers.google.com/machine-learning/glossary/generative
    The production of plausible-seeming but factually incorrect output by a generative AI model that purports to be making an assertion about the real world.
    Checked 2026-10-10.

Think this answer is wrong? Tap "Challenge this answer" on the card in the game and tell us why.

Test yourself in the game

Artificial IntelligenceWhat term does NIST's generative AI risk profile use for AI hallucinations?Artificial IntelligenceA 2025 paper by OpenAI researchers says hallucinations persist partly because most evaluations reward what?Artificial IntelligenceUnder current US guidance, is purely AI-generated material copyrightable?

That's stuff everyone should know.