An AI hallucination is output that sounds plausible but is unsupported, incorrect or fabricated. The word can cover several failure types, from invented facts and citations to incorrect combinations of real information.
Generative AI is designed to produce likely continuations, not to function as an automatic truth machine. Better models, tools and retrieval can reduce some errors, but they do not remove the need for verification.
Ask for sources, open the sources, preserve uncertainty, compare against authoritative records, keep contradictory evidence visible and require human review when a wrong answer could cause meaningful harm.
The more costly the consequence of a wrong answer, the stronger the evidence and review should be before acting.