How do I set up parental controls on ChatGPT?
Link your ChatGPT account to your teen's by sending an invite from Settings, then Parental controls. Once your teen (13-17) accepts, you can set quiet hours, turn off voice mode, memory and image generation, keep extra content protections on, and opt out of model training. You cannot read your teen's chats.
ChatGPT parental controls work by linking a parent's account to a teen's account. OpenAI launched them in September 2025, and its help article on parental controls is the place to check for the current list of settings, because they change. The controls are for teens aged 13 to 17. Children under 13 are not allowed to use ChatGPT at all (see what age kids can use ChatGPT).
Step by step
- Both of you need your own account. The controls link two separate accounts. Don't share a login with your teen.
- Send the invite. In ChatGPT, open Settings, choose Parental controls, and invite your teen. A teen can also invite a parent.
- Your teen accepts. The link only works once they agree, which is a good moment to talk about why you are doing it.
- Choose the settings together. Go through each one with your teen rather than configuring it in secret.
What you can control
| Setting | What it does |
|---|---|
| Quiet hours | Blocks ChatGPT on the teen's account during times you choose, such as school nights after 9pm. |
| Voice mode | Turn it off if you'd rather they type, or keep it on for a child who learns better by talking. |
| Memory | Stops ChatGPT remembering details across conversations. |
| Image generation | Turn off the ability to create images. |
| Content protections | Linked teen accounts get extra limits on graphic, sexual and risky content. Leave these on. |
| Model training | Opt out so your teen's conversations aren't used to train OpenAI's models. |
What the controls don't do
- You can't read their conversations. OpenAI designed the controls this way on purpose. If you want visibility, agree that your teen will show you chats when you ask, or sit together for homework sessions.
- Safety notifications are rare and limited. If OpenAI's systems and reviewers detect signs that a teen may be at serious risk of self-harm, the parent may be notified with only the information needed to help, not the transcript.
- They only cover the linked account. A teen can still use ChatGPT logged out or open a second account. OpenAI's age prediction adds teen protections to accounts that look like they belong to a minor, but it is not perfect.
- Either of you can unlink. The controls rely on trust as much as on settings.
Make a family agreement too
Settings work best alongside a few clear rules. A simple version for a 13 to 15-year-old:
- AI can explain a topic, quiz me, or check my reasoning. It doesn't write my homework.
- I never share my full name, address, school, passwords or photos.
- If ChatGPT says something upsetting or strange, I tell a parent, and I won't be in trouble for it.
- No AI in my bedroom after quiet hours start.
Revisit it every few months. What fits a 13-year-old will feel too tight at 16.
For children under 13
ChatGPT's controls aren't designed for younger children, so the answer there is a different product. Look for a children's service with age-adapted answers and parent visibility. Askie, for example, is built for ages 4-15, lets parents review conversation transcripts and artwork, and keeps settings behind a parent-only PIN. Our guide to parental controls for AI apps lists what to look for in any app.
A safe AI your child can talk to
Askie is a voice-first AI companion built for children, with answers adapted to your child's age and parental controls you set.