Skip to content
AEO Canon · the reference for answer-engine optimization
AEO Glossary

Hallucination

A hallucination is when an AI model states something false or fabricated as if it were fact, usually because it generated from memory instead of grounding its answer in real sources.

BBurke Atkerson

A hallucination is confident-sounding AI output that isn't true. Because a language model generates plausible text rather than looking facts up, it can invent details — fake statistics, nonexistent sources, wrong attributions — especially when it has no retrieved evidence to lean on.

The defense, and the AEO opportunity, is grounding: when an engine retrieves and cites real sources, it's far less likely to fabricate, and your well-evidenced page becomes the safe thing for it to quote. Clear, specific, credible content that's easy to verify reduces the chance an engine garbles or invents claims about your topic — and being corroborated across sources makes the model more confident stating facts about you correctly.

Example. Ask an ungrounded model for a citation and it may produce a realistic- looking but entirely fake reference. The same model, grounded in retrieved sources, instead cites a real page — ideally yours, if you've made it the clearest evidence.

Relevant pillar

Related terms