A model does not know what it does not know. It still answers.
Direct answer
An AI hallucination is a confident answer that is wrong or made up, such as a fake quote, citation or statistic. It happens because language models generate likely-sounding text instead of looking up verified facts. Reduce the risk by giving the model source material, asking for evidence and checking anything important yourself.
Common examples
- Invented references or links.
- Wrong dates, numbers or names.
- Confident descriptions of things that do not exist.
How to protect yourself
- Verify important facts with trusted sources.
- Provide the source text and ask the model to answer only from it.
- Ask it to say when it is unsure.
- Check any citation before using it.
- Use tools that retrieve real documents for current information.
Where Teyro fits
Teyro's The AI Stack course turns this into short lessons you can play in a few minutes a day. It has 100 lessons, and you can start free.
The bottom line
Treat AI answers as drafts, not facts, especially for health, legal and financial topics.
