L272 · Staying safe · 3 min

Teach your kids to check AI

Children should learn early that a chatbot can sound sure of itself and still be wrong.

01 · The idea

Kids tend to trust anything that answers quickly and politely. A chatbot does both. It can also invent a fact, a date or a book title, and say it in the same confident tone as everything else. This is called a hallucination.

Chatbots also like to agree. If your child says "Ho Chi Minh City is the capital, right?", some chatbots will go along with it. Teach your child that a reply that agrees with them is not proof.

Give them a simple habit: for anything that matters, check it in a second place. A school textbook, a teacher, an encyclopedia or a trusted news site all count.

Make it a game. When you catch the chatbot making a mistake together, celebrate it. Your child learns that spotting errors is a skill to be proud of.

02 · Think of it like this

Treat the chatbot like a friend who loves to talk and sometimes makes things up. Fun to chat with, and worth a double check before you repeat what they said.

03 · Do it now

04 · Quick check

Your child says, "The chatbot agreed with me, so I must be right." What should you teach?