Family AI Agreement: House Rules Kids and Parents Can Write Together

A family AI agreement is a short, written set of house rules for how your child may use chatbots and other AI tools. It covers what is usually okay, what needs a parent first, homework boundaries, private information, quiet hours, and what to do when something feels wrong.

Verbal rules fade. A fridge poster does not. Families who write the rules together also give children a clear path to ask for help without fear of punishment.

You can build a printable agreement in about two minutes with our free family AI agreement tool. It asks four or five short questions, keeps answers in your browser, and offers a fridge poster plus a fuller document you can revisit in three months. No account is required, and we do not ask for a child’s real name.

This guide explains why a written agreement helps, what to include, how rules change by age, and how the free tool maps to real family decisions. It extends the short agreement idea in our ChatGPT safety guide. It does not replace school policy, device settings, or product age rules.

Why a written family AI agreement beats verbal rules

Children already live with house rules for phones, games, and bedtime. AI feels newer, so parents often leave it as a spoken warning: “Don’t cheat,” “Don’t share personal stuff,” “Tell me if something weird happens.”

Spoken rules are easy to forget and easy to argue about later. A written family AI agreement does three practical things:

It makes the line visible. Everyone can point to the same page when homework feels fuzzy or a new app appears.
It invites the child into the decision. Agreements made together are easier to follow than rules handed down after a scare.
It pairs with settings. Quiet hours on a poster only work if quiet hours also exist in the app or on the device.

Common Sense Media’s long-standing Family Media Agreement taught many families to put screen expectations on paper. AI needs its own layer: homework honesty, private data typed into chat boxes, image and video uploads, and whether a chatbot is treated as a secret friend. Common Sense’s later research on teens and AI companions also encourages families to address companion-style use inside a media plan, not only after a problem appears.

Australia’s eSafety Commissioner has reported that many children who use AI assistants share personal or sensitive information, and that some encounter inappropriate interactions. A calm household agreement will not remove every risk. It does give children a rehearsed next step: stop, tell a trusted adult, and expect to be heard.

What to put in your family AI agreement

You do not need a long contract. You need clear language everyone understands. Our free tool builds the document from the same decisions most families actually make:

Decision	What you choose	Why it matters
Age band	6 to 8, 9 to 11, 12 to 14, or 15 to 17	Wording and independence change with age
Usual uses	Homework, curiosity, stories, chat, coding, games, languages, music, feelings, writing, and more	Names what is normally okay
Everything else	Ask a parent first, or not allowed for now	Closes the “new app / new feature” loophole
Images and video	Ask first (photos, real people, child’s face in video)	Upload and generation risks are easy to miss
Homework	Learning help, hints only, ask first, or not for schoolwork	Protects learning and school honesty
Quiet hours	None yet, every night, school nights, or a custom window	Keeps bedtime as bedtime
Parent visibility	Parent can review chats, sit together, or talk about use without reading chats	Visibility without secrecy

The printed result also includes child-facing promises, parent promises (listen first, no punishment for showing something uncomfortable), privacy lines, and an “if something feels wrong” clause. Those pieces matter as much as the “not allowed” list.

A short sample agreement you can say out loud

Use this as a starting conversation. Then generate the age-aware printable version with the tool so the wording matches your child’s age band.

AI can help us learn, create, and explore. It is not a person, and it can be wrong even when it sounds sure. We do not type addresses, passwords, codes, or other people’s private details. We use AI during agreed hours. Schoolwork follows the teacher’s rules: AI may help us think, but it does not do the work we hand in. If something feels weird, scary, secret, or too grown-up, we stop and tell a trusted adult. We will not be in trouble for asking for help. We will look at these rules again in about three months.

For families where a parent can review conversations, add:

A parent can read the chats. That is not a secret. We talk about what we see. Review is for safety and learning, not for catching you out.

That second line matches the tool’s visibility option “A parent can review conversations.” If your family prefers sitting together, or talking about AI use without reading chats, say that instead. Honesty beats hidden surveillance.

Age-banded tips for under-13s and teens

Under 13

Prefer adult-led or parent-managed tools. ChatGPT is not intended for independent use under 13. Do not create a false-age account or share an adult login.
Keep sessions short and nearby. “We sit together when AI is on” is often the right visibility choice.
Favor curiosity, stories, and simple practice over open-ended private chat.
Treat feelings support carefully: AI is not a counsellor. A grown-up should stay nearby.
Make a separate agreement for each younger sibling when ages differ.

For choosing apps with parent controls and age fit, use the safe AI checklist.

Teens (about 13 to 17)

Expect more independence, and keep homework honesty non-negotiable.
Name companion risk in plain words: AI is not a secret friend, therapist, or replacement for real people.
If the teen uses ChatGPT, link parental controls and read what they do and do not provide. Linked parents can manage selected settings and may receive limited safety notifications, but they cannot read the teen’s conversations. Either person can unlink. Details live in our ChatGPT safety guide and the Pip vs ChatGPT for Teens comparison.
Decide visibility together. “We talk about AI use” can be the right teen option when the product does not offer parent chat review. The agreement should still say when the teen must stop and involve an adult.

Homework rules that protect learning

Homework is where family AI agreements fail most often, because “help” and “do it for me” can look similar on a screen.

A practical family line, matching the tool’s options:

Yes, for learning help: AI may explain, hint, and quiz. It may not write the work that gets handed in.
Hints and quizzes only: No finished answers or paragraphs.
Only with a parent or teacher’s okay.
Not for schoolwork.

When in doubt, the teacher’s rule wins. Put that sentence on the poster.

For prompts, subject examples, and how to judge cheating versus learning support, see AI for Homework: How Kids Can Learn Without Cheating.

Privacy and oversharing

Children often treat a chatbot like a private notebook. It is closer to typing into a product that may store, process, or train on content, depending on the service and its settings.

Keep the agreement concrete:

No home address, live location, or school name on a document that does not need it.
No passwords, verification codes, or photos of other people without asking.
No health, money, or family secrets typed into a chatbot.
Ask a parent before uploading photos, screenshots, or generating pictures of real people.
Ask before video tools, especially anything that uses a child’s face.

If your family uses multiple AI apps, the privacy lines should travel with the child. The agreement is about behavior, not about one brand.

Companion use, feelings, and “secret friend” risk

Some children use AI for jokes and curiosity. Others start leaning on it for comfort, validation, or a sense of being uniquely understood. Companion-style products are designed to keep people talking. That design can blur the line between a tool and a relationship.

Your agreement can stay short and firm:

Friendly chat is okay only if it does not replace friends, sleep, or school.
Feelings talk needs a grown-up nearby. AI is not a counsellor.
AI is not a secret friend. No “don’t tell your parents” requests.
If content feels sexual, violent, manipulative, or like something to hide, stop and tell a trusted adult.

Parents promise something in return: you will listen before you react, and you will not punish a child for showing you something uncomfortable. Without that promise, children hide the moments you most need to see.

The APA’s advisory on AI and adolescent well-being and Common Sense’s companion research are useful background if you want a deeper parent briefing. For immediate household language, the short clauses above are enough.

When to revisit the agreement

Set a date. Our tool suggests looking again in about three months. Also reopen the conversation when:

A major new AI feature or app arrives in the house
School AI policy changes
Your child moves into a new age band
Homework stress rises or grades slip in ways that track with AI use
You notice secrecy, bedtime fights, or preference for AI over people

Updating the agreement with the child present teaches judgment. Ask what still feels fair, what is too loose, and which tools classmates are using.

Checklist: run the conversation this week

After the poster: when Pip is the natural next step

A written agreement is the conversation. Product settings are how some families keep the conversation honest on busy weeks.

Pip is parent-managed safe AI chat for kids on iOS and Android. Parents create the account and child profiles. Families can set quiet hours and topic limits, and parents can review conversations instead of borrowing an adult ChatGPT login for a child who should not have one. In the US App Store, Pip is $9.99. Pip is not affiliated with OpenAI, and it is not a ChatGPT clone for under-13s.

If you mainly need a fridge poster, stop at the free tool. If you want the same themes (quiet hours, boundaries, visibility) applied inside a child-focused product, Pip is built for that next step. For a side-by-side with ChatGPT for Teens, read Pip vs ChatGPT for Teens.

FAQ

What is a family AI agreement?

It is a short written set of household rules for AI chatbots and related tools. Typical topics include allowed uses, homework, privacy, quiet hours, parent visibility, and what to do when something feels wrong.

How is this different from a screen time or family media agreement?

A general media agreement covers devices, time, and online conduct. A family AI agreement adds AI-specific lines: homework honesty, what never to type into a chatbot, image and video uploads, and whether AI may act like a companion or secret friend.

Do we need a separate agreement for ChatGPT?

You need clear rules for every AI tool in the house. ChatGPT has its own age rules and parental controls, and linked parents still cannot read teen chats. Use this agreement for family behavior, then read Is ChatGPT safe for kids? for product-specific setup.

Does a written agreement replace parental controls?

No. The poster and the settings should match. Quiet hours on paper without quiet hours in the app are mostly decoration.

What if we have more than one child?

Make one agreement per age band. Siblings can share a fridge poster and still have different homework or independence rules.

Should parents read every chat?

Only if that is the honest family rule and the product supports it. The healthier options are review with the child’s knowledge, co-use, or regular conversations about AI use. Secret monitoring teaches children to hide.

Is your free tool saving what we type?

No. The family AI agreement tool generates the document in your browser. Printing or copying is up to you. It does not create an account from those answers.

How often should we update the rules?

About every three months is a solid default, plus any time school policy, tools, or your child’s maturity changes.