AI's 'hallucinations' are its most notorious flaw—the model confidently fabricates information that seems plausible but is actually incorrect, leading to inaccuracies or false information in the content, and becoming a key obstacle to fully trusting AI.