Skip to content

Glossary entry hallucination

Hallucination

When an AI engine generates incorrect facts not supported by retrieved sources. The risk grows when retrieval coverage is poor.

Filed under
Core concepts

Definition

Hallucination is the generation of plausible-sounding but factually incorrect statements by an AI engine. Common causes: retrieval failed to surface authoritative source; the model interpolated between conflicting sources; the query is outside any cited source's coverage. For AEO purposes: hallucinations about YOUR brand are a serious risk that can be mitigated by being the cited authoritative source.

Example

User asks 'what does Halden & Ruiz do?'. If Halden & Ruiz has no Wikidata entry and sparse press coverage, the engine may hallucinate plausible-but-wrong descriptions.

How to optimize

Be the cited source for queries about your brand. Build a strong knowledge-graph presence (Wikidata + Wikipedia). Submit clear About content + JSON-LD Organization schema.

Check a page against a real AI answer

Claude’s live answer to your buyer’s question, its sources, your readiness score and the fixes. Free, no signup, about half a minute.

Run a free audit