What is a hallucination in AI models?
A hallucination is a statement by an AI model that sounds fluent and convincing but is factually wrong or not backed by any source. Examples are invented product features, wrong prices, studies that do not exist, or quotes attributed to a source that does not contain them.
Language models produce text that fits the patterns they learned. They do not check on their own whether a statement is true. Grounding with retrieved sources reduces the problem but does not remove it: a model can misstate a correct source or mix up information from two sources.
For brands, hallucinations are a risk of their own. An AI can attribute services to a company that it does not offer, or confuse it with a place of the same name or another business. Such mix-ups are especially common with names that are also common words or place names.
Tools that measure AI visibility can be affected too, when they use a language model to extract brand names from answers. A careful tool checks every extracted name against the original text.
What it means for your website
For a website it helps to state the key facts about the brand clearly and in an easy-to-find place: what is offered, for whom, where, and how the name differs from similar ones. Whether an AI makes false statements only shows when you read real answers word for word, not in a condensed metric.
Sources
This entry contains no figures and relies on generally documented terms.
Terms help you understand. Whether AI crawlers can reach your site is something you measure.
Check your website for freefree · no signup · no credit card