PURPOSE AI hallucinations are confident-sounding outputs that are false, unsupported, or invented. WARNING SIGNS Look for fake citations, overly specific claims without sources, invented policies, wrong dates, impossible steps, and answers that do not match known facts. CHECK CAREFULLY Verify important claims against trusted sources or official documentation. BE SKEPTICAL Confidence does not equal correctness. BEST PRACTICE The more important the decision, the more verification is required.