Verify facts before using them

Verify facts before using them

AI-generated text can read smoothly even when it contains a mistake, an incomplete conclusion, or an invented detail. Fluency is not proof. If the output includes names, dates, amounts, quotations, or expert claims, verify them using a source appropriate to the subject.

When summarising a document, compare important claims with the original text. Check whether the model omitted an exception or confused a proposal with an approved decision. For questions that depend on current information, check the date and source instead of relying on a general answer.

If the model provides a link or names a source, open it and check whether it really supports the claim. A trustworthy-sounding title doesn’t mean the source exists or says what the response claims.

Match the level of checking to the consequences of a mistake. For a draft of a personal list, a quick edit may be enough. For information you’ll share with a client, publish, or use for an important decision, check every key detail and, if needed, get a qualified person to review it.

If you can’t independently verify a claim, don’t present it as fact. Ask for a source or remove it from the output.

For material intended for clients or the public, decide who will check the information and whether you need to cite a source or indicate uncertainty.

What should you do with factual claims in an AI response?

AI in Everyday Life: Use It Effectively and Safely

Continue the training

Enter your email address to continue the training.