A concrete example
The AI gives you a study's name, a date, and a link, but the study doesn't exist.
Why it matters
It's the most common risk when you use AI to learn, publish, or decide.
You run into it in nonexistent citations, wrong numbers, invented functions, or overly confident summaries.
Don't mix it up with
RAG: RAG lets an AI answer using documents pulled up at the moment you ask the question.
Grounding: Grounding means anchoring an answer in sources that were provided or retrieved.
Common mistakes
- Only checking when the answer sounds hesitant.
- Asking for sources without opening them.
- Using AI as your only proof on a sensitive topic.
Quick checklist
- First I check whether the word names a concept, a tool, a risk, or a metric.
- I tie it to a concrete case: The AI gives you a study's name, a date, and a link, but the study doesn't exist.
- I keep the main trap in mind: Only checking when the answer sounds hesitant.
Quick questions
What is Hallucination in AI?
A hallucination is a made-up or wrong answer that the model presents with full confidence.
Where will I run into Hallucination?
You run into it in nonexistent citations, wrong numbers, invented functions, or overly confident summaries.
Which word should I read next?
Start with RAG, Grounding, Evaluation / Eval.