Every parent's instinct with a new technology is to reach for a ban or a lecture. Both fail with AI for the same reason: the tool is everywhere, and a rule your child had no part in making is a rule they'll route around. What works is a short, shared agreement — a handful of plain rules your child helped write, posted where everyone can see them. Here's one you can use tonight.
Copy the template below, tweak it together, and stick it on the fridge.
- A shared agreement beats a ban. Kids follow rules they helped write; they route around rules imposed on them.
- Cover five things: where/when, what-for, the honesty line, privacy, and the 'tell me' rule.
- Use the copy-and-paste template below — then sign it together and post it where it's visible.
- Revisit it as your child grows; the agreement should get lighter as trust is earned.
Why a family AI agreement beats a ban
Banning AI does three unhelpful things: it pushes use underground, it removes your chance to teach good habits, and it frames AI as forbidden fruit. An agreement does the opposite. It keeps the behaviour visible, turns you into a guide rather than a police officer, and — because your child co-writes it — gives them ownership. Ownership is what turns a rule into a habit.
The magic ingredient is co-authorship. Ask your child which rules seem fair and why. You'll get better rules and far more buy-in than anything you hand down.
The rules that actually work
- Where & when. Shared spaces, agreed hours — not overnight, not alone in the bedroom for long stretches.
- What it's for. Learning, creating and help with tasks — not companionship, not secrets, not as a stand-in for a friend.
- The honesty line. AI can coach, explain and check. It can't do the work you're meant to do. Copying its answer is the same as copying anyone's — see our guide to homework and honesty.
- Privacy. No full names, school, address, phone, passwords or photos in a prompt. Ever.
- Check before you trust. AI can be confidently wrong; important facts get verified — the habit behind spotting fakes.
- The 'tell me' rule. Anything scary, weird or upsetting = come to me, no punishment.
The free family AI agreement (copy & paste)
Here's a ready-made version. Edit the bits in brackets together, then both sign it.
Our Family AI Agreement
We agree that AI is a helpful tool when we use it well and safely. In our home:
1. I use AI in [shared spaces], during [agreed times], and not overnight.
2. I use AI to learn, create and get help — not as a friend, a secret-keeper or a therapist.
3. I keep myself private: no name, school, address, phone, passwords or photos go into any AI.
4. I stay honest: AI can help me understand and check my work, but it doesn't do my work for me.
5. I stay a sceptic: I check important facts, because AI can sound sure and still be wrong.
6. I speak up: if anything feels scary, strange or upsetting, I tell [parent/guardian] and I won't be in trouble.
7. Grown-ups keep their side too: [we set AI up safely, we're happy to help, and we'll answer questions honestly].
Signed: ______________ (child) ______________ (parent) Date: __________
Want the setup side that goes with rule 7? Follow our step-by-step guide to deciding on and setting up ChatGPT for your child.
How to make it stick
- Sign it together and post it where it's seen — the fridge, the study wall.
- Model it. Narrate your own AI use out loud: "I'm going to double-check this, because it might be wrong." Kids copy what you do, not what you say.
- Review, don't relitigate. Once a month, ask what's working. Adjust one thing rather than rewriting everything.
- Reward the right moves. Praise a caught mistake or an honest "I used AI to plan, then wrote it myself."
Adjusting it by age
| Age | What changes |
|---|---|
| 10–12 | More together-time; shorter sessions; parent operates most tools; heavier emphasis on 'tell me' and privacy. |
| 13–15 | More independence; keep the honesty and privacy rules firm; move from supervising to spot-checking. |
| 16+ | Mostly self-managed; the agreement becomes a light touchstone, not a leash; focus shifts to using AI well for study and creating. |
The direction of travel is always the same: as trust is earned, the agreement gets lighter. For the bigger conversation around all this, see how to talk to your kids about AI.
An agreement sets the boundaries; a course builds the judgement behind them. AI Fundamentals for Kids teaches 10–16 year olds how AI works, how to use it honestly, and how to stay safe — so the rules make sense from the inside. Self-paced, no coding.
Explore the course →Ages 10–16 · No coding · Parent-guided · 7-day money-back guaranteeFrequently asked questions
What rules should kids follow when using AI?
Five essentials: use it in shared spaces at agreed times; use it to learn and create (not for companionship or secrets); stay honest (AI helps, it doesn't do the work); keep personal details out of prompts; and tell a parent if anything feels off. The template above turns these into a signed agreement.
Is it better to ban AI or set rules?
Set rules. Bans push AI use underground, remove your chance to teach good habits, and make it more tempting. A short agreement your child helps write keeps use visible and teaches judgement — which is what actually protects them long term.
At what age should kids start following AI rules?
As soon as they start using AI — typically around 10, and earlier if they use any AI-powered app. Younger children need simpler rules and more supervision; the agreement grows up with them.
How do I get my child to actually follow the agreement?
Co-write it, sign it together, post it visibly, model the behaviour yourself, and review it monthly rather than lecturing. Buy-in comes from ownership and from seeing you follow the same rules.
Can I use this AI agreement for a classroom?
Yes — teachers can adapt the template for a class or club. If you'd like a group or school arrangement for the course too, contact us and we'll help.