Hallucination
Definition
A hallucination is a plausible sounding, but unproven or false AI statement.
The generation of plausible-sounding, but factually false, invented or unsupported statements by an AI model.
Simply explained
Like a convincingly told answer for which the evidence is missing.
Why it matters in practice
Helps to realistically classify the possibilities, quality and limitations of AI systems.
Example from business
Possible example: An AI names a source; the editorial staff opens it and finds that the statement is not there.
Limits and pitfalls
- Publish invented sources, figures or legal statements unchecked.
Origin & history
NIST lists false or invented generative spending in the Generative AI Profile of 2024 under the risk of confabulation. „Hallucination is a common term for such spending.
Use in mid-sized companies
- Possible application in small and medium-sized enterprises: An AI calls a source; the editorial staff opens it and finds that the statement is not there.
Frequently asked questions
What does hallucination mean?
A hallucination is a plausible sounding, but unproven or false AI statement.
What errors should be avoided?
Publish invented sources, figures or legal statements unchecked.
Sources
- AI 600-1: Generative AI Profile; Risiko Confabulation · NIST (2024), AI 600-1: Generative AI Profile; Risiko Confabulation
Sources checked on 09.10.2026. Editorial use cases are hypothetical; no evidence of real customer results.
Editorially reviewed on: