Hallucination
When an AI model generates false, fabricated, or logically inconsistent information while presenting it as factual.
In the context of AI, a Hallucination occurs when a Large Language Model confidently generates information that is factually incorrect, nonsensical, or entirely fabricated.
Because LLMs are fundamentally predictive text engines designed to generate highly probable sequences of words, they do not have a built-in concept of "truth." If they lack information on a topic, they may stitch together plausible-sounding but completely false statements to satisfy the user's prompt.
Mitigating hallucinations is a primary focus of AI engineering, often addressed by grounding the model with facts via Retrieval-Augmented Generation (RAG), utilizing Chain of Thought reasoning, or implementing strict system prompts.
Related Terms
Retrieval-Augmented Generation (RAG)
A framework that improves AI responses by dynamically retrieving facts from an external knowledge base before generating an answer.
Chain of Thought (CoT)
A prompting technique where an AI model explicitly generates a step-by-step reasoning path before providing a final answer.