When an AI model generates confident-sounding but factually incorrect or fabricated information.
Hallucination happens because LLMs generate plausible-sounding text based on patterns, not verified facts. RAG and careful prompt design help reduce it, but rarely eliminate it entirely.
