Hallucinations Undermine Trust; Metacognition is a Way Forward

ArXi:2605.01428v1 Announce Type: new Despite significant strides in factual reliability, errors -- often termed hallucinations -- remain a major concern for generative AI, especially as LLMs are increasingly expected to be helpful in complex or nuanced setups. Yet even in the simplest setting -- factoid question-answering with clear ground truth-frontier models without external tools continue to hallucinate.