Hallucination (AI)

In artificial intelligence ✕ Artificial intelligence is a computer system’s ability to perform tasks we associate with human intelligence, such as understanding language, recognizing images or making decisions. More in the glossary → , a hallucination happens when a model generates false, made-up or inaccurate information and presents it with the same confidence as a correct fact. It occurs because a language model does not look up a database ✕ A database stores a site’s content —posts, pages, comments— in an organized way so it can be retrieved instantly. WordPress uses MySQL or MariaDB. More in the glossary → of facts: it predicts the most plausible text from its patterns, and sometimes the most plausible is not the true one. It can invent quotes, dates, sources or product features that do not exist. It is one of the most important limitations of generative AI ✕ Generative AI creates new content —text, images, audio or code— from a prompt, instead of only analyzing existing data. More in the glossary → and the reason its answers must be verified, especially on medical, legal or financial topics. Techniques like connecting the model to real sources ( RAG ✕ RAG connects a language model to an external data source, so it answers based on real, up-to-date information instead of only its memory. More in the glossary → ) or asking it to cite reduce the problem, but do not eliminate it entirely. The practical rule is clear: useful as an assistant, not as a final source of truth.