Accuracy, hallucinations & checking

What is a "hallucination"?

Because AI predicts likely text, it can produce confident-sounding but false statements — a wrong date, a made-up source, an invented rule. This is the single most important risk to manage.

How to protect yourself

• Verify facts, names, numbers, and laws independently. Never publish them on AI's word alone.
• Ask for sources — then check them; AI can invent citations too.
• Be extra careful with legal, medical, or financial specifics.
• Use it for drafting, not deciding.

Lower the risk

Giving the AI your real source material (Module 2) keeps it grounded and cuts hallucinations sharply. This is why tools like Lex Belize are built on your documents, not the model's guesses.

Key takeaway: Assume AI can be confidently wrong. Verify anything that matters before you rely on it.