AI Literacy · Hallucinations

AI Hallucination Awareness

An AI hallucination is a confident-sounding output that is factually incorrect, fabricated, or unsupported by evidence. Recognizing them is a core professional skill.

Last updated: August 9, 2026

1. What hallucinations look like

Invented citations, made-up statutes, fictional case names, fabricated quotes, wrong dates, wrong prices, wrong dosages, plausible but inaccurate summaries, and overconfident explanations.

2. Confidence is not correctness

AI systems do not signal uncertainty reliably. A fluent, well-structured answer can still be entirely wrong. Tone is not evidence.

3. Fact verification

Cross-check claims against authoritative primary sources.
Ask the AI for its sources — and then independently verify those sources exist and say what the AI claims.
Be especially skeptical of numbers, names, dates, citations, and direct quotations.

4. Responsible review

Treat every AI output as a draft from an enthusiastic but inexperienced assistant. Edit, verify, and take responsibility for the final result.