ChatGPT parental controls: how to set them up and what parents can (and can't) see

A step-by-step guide to ChatGPT's parental controls: linking your teen's account, which settings you can change, when you'll get safety alerts and how age prediction works. Plus tips for talking about it at home.

Hands holding a phone next to a laptop on a desk
Image: Firmbee en Pixabay

Lots of teens use ChatGPT for homework, questions or just to chat. If you’re a parent, you may be wondering what your kid sees in there and whether you can set any limits. You can: ChatGPT has had parental controls since 2025.

This guide uses OpenAI’s official help pages to explain how to turn them on, what you can change and, just as important, what you can’t see.

At a glance

Question Short answer
Can I read their chats? No. Not their history and not what they’re typing live
Can I limit when they use it? Yes, with “quiet hours”
Can I turn off voice or images? Yes, each one separately
Will I be told about serious problems? Yes, in high-risk cases such as self-harm, without seeing the conversation
Will my teen know? Yes. You both have to accept the link, and you’re notified if they remove it
Until what age? Until 18, when the account becomes an adult account

Your teen needs their own ChatGPT account, and you need yours. Then:

  1. Open ChatGPT (the web version is easiest) and go to Settings → Parental controls.
  2. Tap Add family member.
  3. Choose to invite them by phone or email and enter theirs.
  4. Your teen gets the invite and accepts it.

Your teen can also invite you. One parent can link several teens, but each teen can only be linked to one parent.

Either of you can unlink at any time. If your teen does it, you’ll be notified.

Step 2: choose the settings

Once linked, you can change these from your own account. Defaults are in brackets:

Setting What it does
Reduce sensitive content (on) Limits graphic content, dangerous viral challenges, romantic, sexual or violent roleplay and extreme beauty ideals
Quiet hours (off) Times when your teen can’t use ChatGPT, such as overnight or during school
Voice mode (on) Talking to ChatGPT out loud
Image generation (on) Creating and editing images
Memory (on where available) Letting ChatGPT remember details from earlier chats
Study mode (off) New chats start in study mode, which guides instead of handing over answers
Study hours (off) A schedule when study mode opens; it doesn’t block use
Improve the model for everyone (on) Whether their chats can be used to train the AI
Health (on where available) ChatGPT’s health section

There are also settings for work and coding features (Codex network access and cloud browser) that most families won’t need to touch.

Our suggested starting point: keep “reduce sensitive content” on, set quiet hours overnight, turn off “improve the model for everyone,” and think about whether your teen really needs memory and voice mode. If they use it for school, try study mode; there are more ideas in how to study with AI without cheating.

Step 3: understand safety alerts

Alerts switch on automatically once you link (they can take a few hours to start). You get them by email, text or push notification, only in specific cases, for example:

  • When OpenAI’s trained reviewers spot a serious self-harm risk.
  • Signs of disordered eating.
  • When the account is suspended or closed over violence.

OpenAI is clear that alerts don’t include the conversation and are not real-time monitoring. Think of it as an alarm for serious cases, not a camera.

Age prediction: what happens if ChatGPT thinks your kid is under 18

Even if someone enters a birth date when signing up, ChatGPT tries to guess whether the user is under 18 from signals like the topics they discuss, what time of day they use it and how old the account is.

  • If it thinks the user is a minor, it switches on the teen experience: less sensitive content, some features limited and no ads.
  • If it gets an adult wrong, the account isn’t blocked: the user is asked to verify their age through a third-party company (Persona), with a selfie, an ID or both, depending on the country. OpenAI says it doesn’t receive those images and that Persona deletes them within 7 days.

OpenAI admits age prediction isn’t perfect. In fact, the state of Oregon has just sued the company, claiming the system “frequently fails.” We cover it in Oregon and California sue OpenAI.

What the controls don’t do

  • They don’t replace talking at home. No filter is perfect, and some teens know how to get around them.
  • They don’t work if your teen uses another account or the app while logged out, or on other chatbots (Gemini, Meta AI, Character.AI…), which have their own rules.
  • They don’t control what your teen shares. Teach them what never to tell ChatGPT: home address, photos, passwords, school details.

How to talk about it with your teen

  1. Ask what they use it for, without judging. Often it’s homework or curiosity.
  2. Explain that AI makes mistakes and sometimes makes things up.
  3. Remind them it isn’t a friend or a therapist. If something is worrying them, they should talk to a person: you, a teacher, a doctor.
  4. Agree on the settings together. It works better than imposing them, and the link is always visible to them anyway.

If your teen or someone you know is struggling or having thoughts of suicide, please reach out: in the US, call or text 988; in the UK and Ireland, call Samaritans on 116 123; in an emergency, call your local emergency number.

This guide explains app features and isn’t a substitute for advice from a health or education professional.

Sources