AI Rules for My Kids & Teens: The Family AI Agreement

Most family AI agreements I've seen are two pages long. They get signed once, go in a drawer, and nobody thinks about them again.

I don't need my kids to sign a contract. I need them to remember a few rules when they're on their own, and I need a way to check those rules that doesn't turn into a argument or an "oh I forgot".

So this is our family AI agreement, the one we actually use at home, in four steps: the conversation, the rules, the follow-up, and what happens when someone breaks a rule. There's a one-page printable at the end if you want something for the fridge.

Natalie Gibson

The short version: A family AI agreement that sticks comes down to four steps: have a 10-minute conversation, agree on seven short rules (private stays private, try it first, check it twice, say when AI helped, weird? tell us, ask before new apps, people before AI), check in once a week, and respond calmly when a rule gets broken.

Step 1: Have the conversation

A family AI agreement lands better when kids help make it. This takes about 10 minutes, and it works best somewhere relaxed. The car is perfect: nobody has to make eye contact, and nobody can leave.

Start by listening. Ask before you explain anything:

• "What do you use AI for?"
• "Which apps you use have AI in them?" Many kids don't think of Snapchat, Instagram or games as AI.
• "Has AI ever told you something that turned out to be wrong?"
• "Has anything an AI said ever felt weird?"

You'll probably learn something.

Then say why. Keep it short and positive. Something like:

AI can help you learn and make really cool stuff. But it can get things wrong, and some apps collect a lot about you. So we're going to agree on a few rules, like we have for the car or the internet. They're for all of us, me too.

Go through the rules together. Read each one out loud and ask your kid to say what it means in their own words. If they can't, reword it until they can.

Let them add one. Ask, "What rule would you add?" Kids follow rules they helped write, and they're often stricter than you'd expect.

End with the promise. This is the most important sentence of the whole conversation:

If anything with AI ever feels weird, you can tell me, and you will never be in trouble for telling me.

With teens, flip it. Ask them to draft the rules first, then compare notes. Teens push back on rules handed down, but they'll defend rules they wrote. Be upfront about what you can see in their apps and why.

Step 2: Agree on the 7 rules

These seven rules are the heart of our family AI agreement. Each one is five words or fewer, so a kid can say it from memory. That's the whole test: if your kid can say all seven back to you by dinner, they're working.

1. Private stays private. Never type your full name, address, school, phone number, passwords or photos of other people into an AI tool. For teens: add photos of friends, screenshots of group chats, and anything you'd hate to see posted.
2. Try it first. Have a go before asking AI. Use it to get unstuck, not to skip the thinking. For teens: agree which subjects AI can help with, based on their school's rules. A tool that tutors with questions, like the AI tutor, fits this rule.
3. Check it twice. AI can sound completely sure and still be wrong. Anything important gets checked with a second source or a person. For teens: health, news and money questions always get checked.
4. Say when AI helped. If AI helped with schoolwork, say so when the teacher asks or the rules require it. For teens: this includes college essays, applications and anything posted as their own work.
5. Weird? Tell us. If an AI says something upsetting, asks for personal information, or tells you to keep a secret, tell a parent. For teens: name a backup adult they can go to first if they'd rather.
6. Ask before new apps. New AI apps, and AI features inside apps you already have, need a parent's OK. For teens: switch from "ask first" to "tell us within a day."
7. People before AI. AI is a tool, not a friend. Feelings, worries and big decisions go to real people first. For teens: AI chatbots that act like a boyfriend, girlfriend or best friend are off-limits.

Plus one of your own

The rule your kids added. Ours is they are not allowed to use each others HeyOtto account.

Let settings do some of the work

Rules go with your kid to every app, including ones you haven't heard of. Settings catch what slips through. Here's what to look for in any AI app, and how we built HeyOtto to handle each one. Those settings live on the safety page and in the parent dashboard.

Step 3: Follow up without nagging

A family AI agreement only works if it stays in use, and rules fade when nobody mentions them. You don't need monitoring software or a weekly lecture, just a couple of habits.

The weekly 5-minute AI show-and-tell. Once a week, your kid shows you one thing they did with AI: a question, a project, or an answer that turned out to be wrong. It feels like sharing, not an inspection, and it quietly checks three of the rules at once. [Natalie: when and where this happens in your house.]

A one-minute check, one rule at a time. Pick one rule a week and ask in passing:

Keep the rules where you'll see them. Stick the printable on the fridge, or make it the lock screen on a shared tablet.

Revisit them at the start of each school year, when someone gets a new device or app, and whenever a rule keeps getting broken. For teens, each review is a chance to hand over more freedom where they've earned it.

Step 4: What happens if your kids break the rules

They will break one eventually. How you react decides whether they keep telling you things, so it's worth deciding now, before it happens.

First, keep the promise. If your kid told you themselves, the telling is never punished, even if they broke a rule along the way. You can still deal with the rule, just not by punishing the honesty.

Then, three steps every time:

1. Find out what happened. Ask what they were trying to do and what the AI said or made. Stay calm; you're gathering facts.
2. Fix any safety issue first. Delete shared information, change a password, report something, or talk to the school if needed.
3. Pick a consequence that matches the rule. Taking the phone for a week teaches less than a consequence connected to what went wrong.

For teens, freedom is the consequence. A broken rule moves them back a step, like from "tell us within a day" to "ask first," for a set time. They get it back by keeping the rule, and they know exactly how.

If the same rule keeps breaking, it probably needs rewriting, not stricter enforcement. Take it back to the conversation in Step 1.

Print your family AI agreement

Just kidding... we do not have one for you to print. Have fun with your family, create your own.

You don't have to get it perfect

You don't need to understand every AI app to raise kids who use AI well. Start with the conversation, keep the rules short, check in once a week, and stay calm when something goes wrong. Your family AI agreement will change as your kids grow. The promise that they can always tell you shouldn't.

If you're looking for an AI your kids can use that's built around rules like these, see how HeyOtto works for families.