AI Platforms
How to keep your child safe using ChatGPT
7 min read · Updated
ChatGPT is for ages 13 and up, with a parent’s permission until 18, and accounts it identifies as under 18 get its teen experience automatically. To keep your child safe, link their account to yours through parental controls, switch off what they do not need, and agree rules on schoolwork and personal information. The controls do not show you their chats, so the conversation does most of the work.
Most children who use ChatGPT use it for ordinary things: explaining a topic from class, checking a math method, drafting an email, answering a question they were curious about. Keeping them safe is less about locking it down and more about a sensible setup, a few clear agreements, and staying in the conversation. Every other app’s minimum age is in AI chatbot age limits.
This guide is the practical checklist. For the full detail of every setting, see ChatGPT parental controls.
1. Check your child is old enough
OpenAI’s Terms of Use say users must be at least 13, or the minimum age required in their country, and that anyone under 18 must have a parent or legal guardian’s permission. If your child is under 13, ChatGPT is not meant for them, and the safest setup is not to have an account at all.
For 13- to 17-year-olds, ChatGPT automatically switches eligible accounts into its ChatGPT for Teens experience when the age on the account, a verified age or OpenAI’s age prediction indicates the user is under 18. That experience adds content safeguards, study tools and reminders to take breaks. It is a good baseline, not a complete setup.
2. Set up their own account, then link it
- Let your child create their own account with their real date of birth, rather than using yours. OpenAI’s terms do not allow sharing an account, and parental controls work by linking two separate accounts.
- Link the account to yours through parental controls. Either of you can send the invitation, and your child accepts it from their own account.
- Do it together, at the same screen. Settings that were agreed are far less likely to be worked around than settings that were a surprise.
Be clear with yourself about what linking does. It lets you change settings and receive narrow safety notices. It does not let you see their chats or history.
3. Review the settings together
Once linked, these are the settings most families look at first:
- Quiet hours, for example overnight and during school hours if that fits your family.
- Improve the model for everyone, which you can turn off so their conversations are not used to train OpenAI’s models.
- Reference saved memories, if you would rather ChatGPT did not carry personal details from one chat to the next.
- Image generation and Voice Mode, which you can turn off if your child does not need them.
- Study mode and study hours, which make new chats start with guided questions instead of finished answers.
- Reduce sensitive content, which is on by default and worth leaving on.
Revisit the settings every few months. What suits a 13-year-old rarely suits a 16-year-old, and loosening a setting together is a good way to show trust is being earned.
4. Agree what never goes into a chat
Children often share more than they mean to. A conversation with ChatGPT is a message sent to a company, not a private thought. A short, specific list works better than a general warning:
- Full name together with school, address or phone number.
- Passwords, account codes or anything used to log in.
- Photos of documents or homework sheets with names on them.
- Screenshots of group chats or other people’s messages.
- A friend’s private situation that the friend has not chosen to share.
AI privacy for families explains what happens to what they type and which settings matter most.
5. Set a clear rule for schoolwork
Schoolwork causes more real trouble than anything else, mostly because the line between help and doing the work is blurry. A rule that names examples works best:
- Fine: explain this, quiz me, check my reasoning, tell me what my essay is missing.
- Ask first: rewording my own sentences, making a first outline, translating my own work.
- Not fine: write it for me so I can hand it in.
Check the school’s own policy first, so your rule and theirs agree. Teach one habit above all: ChatGPT can sound confident and still be wrong, so anything that matters, such as a fact, a quote or a citation, gets checked somewhere else.
6. Talk about it, often and lightly
The goal of the first conversation is not to set rules. It is to find out how your child actually uses it, so that later conversations are easy. Curiosity works better than suspicion.
Openers that tend to get a real answer:
- What is ChatGPT actually good at? Show me something you used it for this week.
- Has it ever told you something you knew was wrong?
- What does your teacher say you can use it for?
- Is there anything you would never type into it?
- If something it said ever bothered you, would you tell me?
How to talk to your child about AI has more on tone and timing.
Signs something may deserve attention
None of these means something is wrong. Each is a reason for a calm conversation, not a conclusion:
- Using ChatGPT late at night, well outside the hours you agreed.
- Talking about it as a friend or the main place they go for advice.
- Unusual secrecy about one AI app in particular, or a second account you did not know about.
- Schoolwork that no longer sounds like them, or that they cannot explain when asked.
- Seeming upset, withdrawn or unsettled after time spent with an AI chatbot.
If a conversation raises something about their wellbeing, the next step is support from people: you, their school, their doctor or another qualified professional. A settings change is not the answer to a worry about how your child is feeling.
What parental controls will and won’t tell you
With accounts linked, OpenAI may contact you in a small number of serious situations, such as when trained reviewers identify a serious self-harm concern. Those safety notices do not include the conversation. Independent testing has found they do not always arrive; Is ChatGPT for Teens safe? covers what it found. On an ordinary week you will hear nothing, and the controls only cover ChatGPT. Parental controls across AI services explains how other AI apps compare.
November 17, 2026: Oregon sued OpenAI, alleging among other things that ChatGPT’s age checks frequently fail and that ChatGPT for Teens did not tell parents when a young user asked about suicide or self-harm. These are allegations, not court findings, but they are a good reason not to treat any setting as the whole answer. Oregon’s lawsuit against OpenAI sets out what parents can do.
No tool replaces the conversation
Every setting on this page can be sidestepped with a second account, a friend’s laptop or a different AI app. What carries over to all of them is your child’s own judgment and their willingness to tell you when something feels off. The settings buy time and set expectations. The relationship does the rest.
Where TeenCue fits
Some parents want more everyday context than ChatGPT’s controls give, especially when a child uses more than one AI service. On your child’s computer, in Chrome, Firefox, Edge or Brave, TeenCue records what a chat on a supported AI website was about: a category, a topic and a concern rating. Parents never see what was typed, on a computer it does not block anything, and it is visible in their browser rather than hidden. On a teen’s iPhone or iPad, the TeenCue app blocks AI apps and other browsers, so AI chats happen in Safari. It sits alongside ChatGPT’s own parental controls rather than replacing them. How TeenCue works has the detail.
Common questions
Is ChatGPT safe for a 13-year-old?
Thirteen is the minimum age in OpenAI’s Terms of Use, unless your country sets a higher one, and accounts identified as under 18 get extra content safeguards automatically. Whether it suits your child depends on the setup and the agreements around it, especially about schoolwork and personal information.
Can I see my child’s ChatGPT history?
Not through parental controls. OpenAI states that linking accounts does not give a parent access to conversations or chat history. Some families agree to look at examples together, which works best when it is agreed in advance.
Should my child use ChatGPT for homework?
For explanations, practice questions and feedback on their own work, it can help. Agree a clear rule that matches the school’s policy, and ask them to explain their reasoning out loud now and then.
What should I do if I am worried about something my child asked ChatGPT?
Start with a calm, open conversation rather than a confrontation. If it concerns their safety or wellbeing, involve a doctor, school counselor or other professional, and in an emergency call 988 in the US or your local emergency number.
Sources
- OpenAI Terms of Use, effective 1 January 2026: Registration and access (checked 8 October 2026)
- OpenAI Help Center: ChatGPT for Teens (checked 8 October 2026)
- OpenAI Help Center: Managing parental controls in ChatGPT (checked 8 October 2026)
- OpenAI Help Center: Age prediction in ChatGPT (checked 8 October 2026)