Translation: An AI hallucination is when an AI system generates a response that is incorrect, unsupported by data, or entirely made up, even though it may sound plausible. This can include invented facts, false references, or descriptions of events that never occurred. Hallucinations happen because the system generates text by predicting likely word patterns from its training data, rather than checking those predictions against verified or factual knowledge.