Skip to content
fAImly.

Family guide

How to fact-check ChatGPT with your child

A fluent answer can still contain a mistake. You do not need to verify every sentence. Pick the one claim that would change what your family believes, buys, shares, or does.

Start with one claim

Ask your child: “Which part of this answer would we need to be right?” Copy that sentence, not the whole reply. If the answer names a book, place, person, date, or rule, start there.

  1. Say the claim in your own words.
  2. Predict where you could check it without asking the same chatbot again.

Check outside the chat

For a book, use a library or publisher listing. For a place, use the venue’s own current information. For a school rule, read the school’s written guidance. Open the source and check whether it actually supports the claim. A search snippet or another chatbot reply is not the source itself.

  1. Find the original source.
  2. Compare its exact detail with the AI answer.
  3. If they disagree or the source is missing, leave the claim unconfirmed.

Decide together

Ask: “Does this change what we would do?” Correct the answer, ask a better follow-up, or set it aside. For health, legal, financial, or safety decisions, involve a qualified person before acting.

Try it together

Try a low-stakes example: ask ChatGPT for a book by an author your child likes. Check the title in a library catalog. Let the child decide whether to keep that recommendation.

Keep exploring

An AI hallucination activity for kids →

See real AI mistakes →

Want to use fAImly with your family? Request a private beta invite →