The Kids & Teens AI Kit

Kids are already using AI chatbots for homework, questions and company. The apps now have real parental tools, but you have to switch them on. This kit covers who's allowed to use what, the settings worth changing, the talk worth having, and a one-page family agreement to print and sign together.

Written in plain English. Every app step comes from the company's own help page, checked October 4, 2026.

Your progress0 done

Jump to the family agreement

1. Who can use which app

App	Age rule (from the company)
ChatGPT	13 and up. Under 18 needs a parent or guardian's permission. Accounts that look under 18 get ChatGPT for Teens automatically, with extra protections.
Gemini	Children under 13 only through a supervised Google account, and only after a parent turns it on in Family Link. Teens get extra content filters when signed in.
Claude	18 and up only. It isn't meant for kids or teens.

Sources: OpenAI terms · ChatGPT for Teens · Gemini for children · Claude age assurance

2. Set up the controls

ChatGPT parental controls (teens 13–17)

Link your accounts. In your own ChatGPT on the web: Settings > Parental controls > Add family member, choose phone or email, enter your teen's, and Send. Your teen accepts the invite. (Teens can also start it with Invite a parent.) Leave Reduce sensitive content on. It's on by default once linked, and limits graphic content, viral challenges, sexual or romantic roleplay, violent roleplay and extreme beauty ideals. Set Quiet hours. Times your teen can't use ChatGPT at all, for example bedtime. Choose your teen under Family members, turn on Quiet hours and pick the times. Consider Study hours. During these times new chats start in study mode, which gives hints and steps instead of finished answers. It doesn't block ChatGPT. Decide on the extras. You can turn off Voice Mode, Image generation, Reference saved memories and Improve the model for everyone (training on their chats). Check how you'll be alerted. In rare, serious cases, such as signs of self-harm, ChatGPT can send you a safety notification. Under Safety notifications, pick Manage notifications to choose email, text or push.

What you won't see: linking doesn't let you read your teen's chats. Notifications give only the type of concern, not the conversation. Your teen can unlink, and you'll be told if they do. Controls end when ChatGPT identifies them as 18.

From OpenAI's help page: Managing parental controls in ChatGPT

Gemini for children under 13 (Family Link)

Decide whether Gemini is on. At familylink.google.com or in the Family Link app: select your child, then Controls > Gemini > Gemini Apps and switch it on or off. You get an email the first time your child uses it. Try it together first. Google's own advice is to explore it with your child and supervise their use while they learn. Show them how to report a bad answer. Google says its filters aren't perfect; you or your child can mark a response as bad or send feedback.

From Google's help page: Guide your child's Gemini Apps experience. Not available for supervised accounts in the EEA, Switzerland or the UK.

3. The talk (10 minutes is enough)

Say it in your own words

It's a tool, not a friend. It can sound warm and supportive, but it doesn't think or feel. Big worries go to a person: you, a teacher, a counselor.
Keep personal details out. No full name with address, school name, passwords, or photos of yourself or friends.
It can be confidently wrong. Check facts for school with a second source; ask it "how do you know?"
Homework is for learning. Use it for hints, explanations and quizzing, and follow your teacher's rules on AI.
Fake photos and voices exist. If someone shares an image or voice message that seems off, or makes a fake of anyone, come tell us. You won't be in trouble for telling.

4. Our family AI agreement

We agree that…

We use AI chatbots on accounts our family set up, with parental controls on.
We don't share our full name, address, school, passwords or photos of people in chats.
AI helps us learn; it doesn't do our homework for us. We follow school rules about AI.
We double-check anything important before believing it or sharing it.
If a chat says something upsetting, strange or scary, we stop and tell a grown-up.
We never use AI to make fake pictures, voices or messages of real people.
No AI after ________ on school nights.
We'll look at this again on ____________ and change it together if we need to.

Kid / teenParent or guardianDate

If you're worried about your child's safety right now, contact local emergency services. In the US you can call or text 988 (Suicide & Crisis Lifeline) any time.

Steps checked against each company's help pages on October 4, 2026. These features are new and still rolling out, so some may not appear on every account yet; the linked help pages have the latest. For your own accounts, see the AI Privacy Checkup.

All done. Nice work. Every box on this page is ticked. Want more? The kits below take a few minutes each.

More free kits

Family code wordThe Family Code-Word KitStop AI voice-clone scam calls with a word only your family knows.5 min · Free AI privacy checkupThe AI Privacy Checkup10 minutes to turn off training, tidy memory and keep private things private in ChatGPT, Claude, Gemini and Copilot.10 min · Free First week with AIYour First Week With AISeven everyday tasks, one a day, with words you can copy and try.7 days · Free

Household AI ToolkitComing soon

The Household AI Toolkit

Ready-made AI helpers for family life, set up once in ChatGPT, Gemini or Claude: meals, homework, letters, doctor visits, caring for a parent, and more.

The Kids & Teens AI Kit

1. Who can use which app

2. Set up the controls

3. The talk (10 minutes is enough)

4. Our family AI agreement

We agree that…

More free kits

The Household AI Toolkit

Don't miss what matters