Confabulation
Confidently stated but false or erroneous output from a generative AI system, often called hallucination.
NIST's Generative AI Profile uses confabulation for the production of confidently stated but erroneous or false content, known colloquially as hallucinations or fabrications, by which users may be misled or deceived. It arises because generative models produce likely-sounding output rather than verified facts. In finance a confabulated account number, fee or legal requirement can lead directly to a wrong payment or decision, so facts that drive actions must come from verified sources, not from the model.
Agent Minute explains this term on 29 November 2026.
Related terms
Large language modelA model trained on very large amounts of text to predict likely text, used by most agents to interpret instructions and choose steps.Retrieval-augmented generationRetrieving relevant documents or records first and giving them to the model, so its answer rests on current material.Defective executionA payment that is not executed, or executed late or incorrectly; the rules set which provider is liable.Automation biasThe tendency of people to rely, or over-rely, on the output of an automated system.