AI hallucinations: how to check an answer before you use it

A hallucination is when an AI states something false with complete confidence: an invented fact, a quotation that does not exist, a command with an option that does not exist, a link that does not open. It happens because the model generates the most likely text rather than consulting a source of truth. There is no way to guarantee it will not happen, so the rule is to check what matters before you act.

Where the risk is greatest

Kind of answer Why it fails How to check
Numbers, dates, prices, deadlines The model does not have them and guesses something plausible Confirm on the official page or in the source document
Names, quotations and references It invents authors, studies and laws that look credible Look up the source yourself. If you cannot find it, treat it as non-existent.
Links and addresses It writes addresses that look right Open them before you publish
Commands and code It mixes versions and invents options Check the manual and try it outside production. See Linux commands with AI.
Legal, medical and tax matters Varies by country and changes over time Confirm with a professional or the official source
Recent events The model may not know about them Use a current source

How to check, in order

1 Ask for the source, and open it. If the source does not exist or does not say what the model said, the answer is worthless.
2 Give it the text instead of letting it guess. Paste the document and ask it to answer only from that, and to say “not stated” when it is not. See RAG.
3 Ask in a different way, or on another model. If the answers differ, distrust both.
4 Ask it to explain its reasoning and look for the weak step.
5 Test what can be tested. A command on a test server, a formula with numbers you know, code with a simple case.
6 Have a person read whatever goes to a customer, to the site or into a contract.
A confident tone is not proof. The AI writes just as surely when it is right as when it invents. Asking “are you sure?” does not fix it: it usually says yes, or changes its answer for no reason.
When money, health, law or security is at stake, the AI is a draft and never a decision. You are the one who signs.
Use AI for what is easy to check: rewriting, summarising a text you gave it, putting ideas in order. Be suspicious when you ask it for facts you did not give it.

Want to put an AI assistant on your site without risking wrong answers to customers? Talk to us about the server and the design.

Open a support ticket

SEE ALSO

ChatGPT, Claude or Gemini: which AI tool suits which task

How to use AI on your phone, and what never to type into it

AI that answers from your own documents: RAG explained

What Interweb AI does, and what it does not

RECOMMENDED PRODUCT

Web hosting with cPanel

Domain and SSL included, daily backups and the panel you already know. from $6.59/mo (3-year plan, with coupon)

See plans
  • 0 Users Found This Useful
Was this answer helpful?

Related Articles

ChatGPT, Claude or Gemini: which AI tool suits which task ChatGPT, Claude, Gemini, Copilot, Grok. New versions turn up month after month and the question that reaches us is alway...
How to use AI on your phone, and what never to type into it In Angola, plenty of people work from a phone and do not always have a computer to hand. Artificial intelligence tools w...
Building a WordPress site with AI help: what works and what does not You can get a WordPress site standing up far faster than two years ago, using artificial-intelligence tools for the part...
Building a campaign page fast, with AI help A campaign page is one page, with one goal and one action: leave a contact, ask for a quote, book, buy. It has no menu t...
What OpenClaw is, and what you need to run it OpenClaw is an open-source personal AI agent that runs on your own server. Instead of opening a vendor’s website to talk...
AI for a small business: what to hand over and what to keep in human hands In a small business, AI is good for drafting and sorting: first versions of texts, summaries, answers to repeated questi...