Hallucination
A hallucination is a confident but false statement produced by an AI model, such as a chatbot inventing a configuration option that does not exist in the documentation.
In documentation contexts, hallucinations look like plausible answers about your product that are simply wrong: an invented CLI flag, a deprecated endpoint presented as current, or a code example that never worked. Users cannot tell them apart from real answers.
Hallucinations happen because language models generate likely-sounding text rather than looking facts up. Without access to your docs, a model asked about your product will interpolate from similar products in its training data.
The main defenses are retrieval and refusal. A RAG system grounds answers in retrieved docs passages, and a well-configured assistant declines to answer when nothing relevant is found rather than filling the gap with a guess.
Frequently asked questions
Can hallucinations be fully eliminated?
No. They can be reduced sharply with grounding, source citations, and refusal behavior, but any generative model can still produce errors. Citations let users verify answers against the docs.
Why do chatbots hallucinate about products?
General-purpose models answer from training data, which is stale and mixed with content about similar products. Grounding the model in the actual docs replaces that interpolation with retrieval.