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.
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
- GroundingGrounding is the practice of tying an AI model's answer to specific retrieved sources, so the response reflects real documents rather than the model's unverified internal memory.
- CorroborationCorroboration is when multiple independent, reputable sources agree on a claim about you, giving AI systems the confidence to treat it as fact and repeat it in answers.
- CitationA citation in AI search is when an answer engine credits your page as a source for its response, usually as a linked reference, making it the surviving path to your site in a zero-click answer.