Talk to us
Glossary · AI

What is Hallucination?

Definition

A hallucination is output from an AI model that is fluent and confident but false or unsupported, such as an invented citation, a wrong policy detail or a product feature that does not exist. Hallucinations happen because language models generate plausible text rather than looking up verified facts. They can be reduced through grounding, retrieval, validation and clear instructions, but not fully eliminated.

Why it matters for your business

Hallucinations are the main reason AI output needs controls before it reaches customers, contracts or financial records. The risk depends on the task: a wrong word in a marketing draft is minor, while an invented refund policy or legal citation can create liability. Practical defenses include answering only from retrieved sources, requiring citations, telling the model to say when it does not know, validating extracted values against source systems, and keeping a person in the loop for high-stakes outputs. Test sets should include questions the system should decline to answer.

Example

A software company's support chatbot told a customer that a feature was included in the basic plan when it was not. The fix had three parts: the bot now answers only from retrieved help center articles, it must link the article it used, and if no article matches it says it does not know and offers to open a ticket.

Keep reading

Next steps

01 Browse free n8n templates Fifty starter workflows you can import today. No email required. 02 Take the 5-minute AI readiness assessment Get a score and a recommended first project in your browser. 03 Get a written scope and estimate Tell us what you want to automate. We reply within two business days.