Family promise

We use AI to learn and create. We protect private information, check important answers, tell the truth about our work and ask a trusted adult when something feels wrong.

Print this—or say it out loud

Before using an AI tool

During the activity

Keep it visible

Use shared spaces for younger children and make it normal for a parent to ask what is happening on screen.

Keep it bounded

Use a timer and a defined task. Endless conversations are not a learning plan.

Keep questioning

Ask what the child thinks, what seems uncertain and what needs a second source.

Keep humans first

AI should not replace sleep, play, relationships, teachers or professional support.

Never put these into a prompt

Full names, school names, addresses, phone numbers, passwords, account details, private family disputes, identifiable photos and sensitive health information. Use fictional placeholders instead.

When the answer is wrong

Thank the child for noticing. Compare the output with a trusted source, explain why the mistake may have happened and save the lesson: “A confident tone is not proof.”

When the content is unsafe

Stop, close the conversation and involve a trusted adult if content is sexual, violent, frightening, manipulative or asks the child to keep secrets. For immediate danger, use appropriate local emergency or professional support—not a chatbot.

Review the rules every term

As children grow, the balance between independence and supervision changes. Revisit the agreement when a new tool appears, schoolwork changes or the child starts using AI privately.

UNICEF’s child-centred AI framework highlights safety, privacy, transparency, fairness and wellbeing. Use the full guidance as a reference.

Make the agreement practical

AI Explorers gives families a visible, age-appropriate starting point for practising these rules through short missions. The child learns with an adult rather than being handed an empty chat box and a very large responsibility.

Explore AI Explorers