Home › Glossary › Research › Confabulation

Intermediate · Research

Confabulation

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. An alternative and more technically precise term for the phenomenon of AI hallucination, where an AI generates false or nonsensical information.

Technical Definition

An alternative and more technically precise term for the phenomenon of AI hallucination, where an AI generates false or nonsensical information.

How it works

Confabulation is a synonym for hallucination in the context of AI, particularly Large Language Models (LLMs). It refers to the generation of information that is factually incorrect, nonsensical, or not grounded in the training data. While 'hallucination' is more commonly used, 'confabulation' is considered a more accurate descriptor because it implies the confident assertion of fabricated details, similar to how humans might confabulate memories.

Related Concepts

  • Hallucination — When an AI model generates confident, plausible-sounding content that is factually incorrect or fabricated.
  • AI Safety & Alignment — The field ensuring AI systems behave as intended, remain under human control, and avoid unintended harm.
  • Factuality — A property of a model whose output is grounded in factual reality.

Further Reading

  • Google ML Glossary