# AI hallucination

> Learn why AI systems fabricate convincing but false details and how reliable, machine-readable sources lower the risk for your business.

Synonyms: hallucination, AI hallucinations, fabricated AI answers

An AI hallucination is a convincingly worded but false or fabricated statement by an AI system. Language models compute plausible phrasings — where a reliable source is missing, they guess. Risky for businesses: invented opening hours or services look deceptively real.

Hallucinations stem from how [large language models](https://btlabs.dev/en/glossary/llm-large-language-model) work: they compute which phrasing is most likely — not which one is documented. Where reliable information on a question is missing, the model fills the gap with something plausible. The result sounds confident but may be entirely made up: a wrong phone number, a closing day that never existed, a service you do not offer.

The most effective remedy from a business's point of view: give the AI reliable material. An up-to-date website as the central [source of truth](https://btlabs.dev/en/glossary/single-source-of-truth) , consistent contact details everywhere and [structured data](https://btlabs.dev/en/glossary/structured-data) that labels facts unambiguously. Many AI systems consult real sources before answering — a business that presents itself clearly and machine-readably there lowers the risk of being guessed at instead of reported correctly.

For your business that means checking regularly what assistants say about you — opening hours, prices, services. If wrong details appear, the cause is often an outdated or contradictory source. The fix happens not inside the AI but at the source: bring your website, Google profile and directories to the same, current state.

- [AI Hallucinations](https://btlabs.dev/en/glossary?tag=ki-halluzinationen)
- [AI Basics](https://btlabs.dev/en/glossary?tag=ki-grundlagen)
- [ChatGPT](https://btlabs.dev/en/glossary?tag=chatgpt)

## Related terms

- [LLM (Large Language Model)](https://btlabs.dev/en/glossary/llm-large-language-model)
- [RAG (Retrieval-Augmented Generation)](https://btlabs.dev/en/glossary/rag-retrieval-augmented-generation)
- [NAP (Name, Address, Phone)](https://btlabs.dev/en/glossary/nap-name-address-phone)

## Frequently asked questions.

**How do I keep opening hours and contact details up to date everywhere?**

The core principle: one source instead of many copies. If opening hours, address and phone number are maintained in one central place, and the website, structured data and [machine-readable](https://btlabs.dev/en/glossary/machine-readable-website) formats automatically serve the same state, nothing can drift apart. That is exactly what a website built as a single source of truth delivers: change it once, current everywhere.

Externally, the effort stays manageable: above all, keep your Google Business Profile and the most important directories up to date — with exactly the same details as on your website. For seasonal hours, a fixed routine helps, say at the start and end of the season. The background: contradictory details are the most common reason AI assistants tell people wrong things about a business.

**Why does an AI recommend me with the wrong number or opening hours?**

Because AI assistants mostly pull their facts from map and directory data. If your details are inconsistent or outdated there — an old number, a wrong opening hour — the AI picks up exactly that and states it confidently as an answer.

Directories also often copy from one another, so a single error spreads across several platforms. For your business, that means a small inconsistency in one place can end up as a wrong answer in several AI responses.

**How do I find out whether an AI gives wrong information about my business?**

Regularly ask the major AI assistants about your own business — services, opening hours, prices, and how you compare to competitors — and check the answers for accuracy. Without this check, wrong information often only surfaces when customers arrive with false expectations or never get in touch at all.

---
Source: https://btlabs.dev/en/glossary/ai-hallucination
Languages: [de](https://btlabs.dev/llms/de/glossary/ki-halluzination) · [it](https://btlabs.dev/llms/it/glossary/allucinazione-ia)
See also: [llms.txt](https://btlabs.dev/llms.txt) · [ai.txt (Policy)](https://btlabs.dev/ai.txt) · [identity.json](https://btlabs.dev/identity.json)