Hallucination

Hallucination is when an AI model confidently generates false or made‑up information. It’s a major quality and safety problem because:

• The AI can produce outputs that are misleading or outright wrong, sometimes with real-world risks (e.g., bad medical advice).
• Hallucinations are thought to stem from gaps or weaknesses in the training data.
• To reduce them, there’s a trend toward specialized, domain-specific models that focus on narrower areas, aiming to limit knowledge gaps and lower disinformation risk.