Why do AI chatbots "hallucinate" and make things up?

An AI hallucination is when a chatbot gives an answer that sounds confident and convincing but is false: a made-up statistic, a fake quote, a court case that never happened, or a book reference that does not exist.

Why it happens:

• A language model is trained to produce likely-sounding text, not to check truth. If it has not learned a fact well, it may still produce something that fits the pattern of a correct answer.
• It may have gaps or outdated information in its training data.
• It is often trained to be helpful, so it may "fill in" an answer rather than say "I don't know".
• Vague or leading questions increase the risk.

There have been real-world consequences. Lawyers in several countries, including India, have been criticised by courts for submitting fake case citations generated by AI.

How to protect yourself:

• Verify important facts, especially numbers, names, dates, laws and medical information, from official sources.
• Ask the chatbot for sources, then actually open them.
• Use tools that connect AI to web search or your documents, which reduces (but does not eliminate) errors.
• Never rely only on AI for health, legal or financial decisions.