ChatGPT's parental controls link a parent's account to a teen's account, ages 13 to 17. Both sides have to agree to the link. This page covers the setup, what you can set and what you cannot, as of October 2026.
Set them up
- Open ChatGPT, tap your profile, then Settings, then Parental controls.
- Send an invitation to your teen's account by email. A teen can also invite a parent from their own settings.
- Once the invitation is accepted, pick your teen under Family members.
- Set the controls listed below.
- Choose how you want to get safety alerts: email, text message or app notification.
OpenAI says alerts can take up to a few hours to become active after linking. A teen can be linked to one parent at a time.
What you can set
- Quiet hours. ChatGPT is blocked for your teen during the hours you choose. You can set more than one block.
- Study hours. New chats in those hours start in study mode. Access is not blocked.
- Voice mode, image generation and memory. Each can be switched off.
- Sensitive content. Reduced by default on a linked teen account.
- Model training. You can stop your teen's chats from being used to train OpenAI's models.
What you cannot do
- Read the chats. Parental controls do not show you your teen's conversations.
- Keep the link in place. Your teen can unlink at any time, and that turns the limits and alerts off.
- Link without consent. If your teen does not accept the invitation, there are no controls.
How the safety alerts work
When a message is flagged for self-harm, a person at OpenAI reviews it. If the concern is confirmed, OpenAI tries to reach the parent within a few hours. The alert says your teen may have mentioned self-harm. It does not quote the chat.
In October 2026 Common Sense Media reported that in its tests the alerts did not fire on newly linked accounts. OpenAI disputes this and says the testing may have ended before the controls were active. Until that is settled, do not rely on the alert alone.
If your kid is under 13
These controls do not apply. OpenAI does not allow ChatGPT accounts under 13.
Sources: OpenAI help center, "Parental controls in ChatGPT"; Bitdefender's guide of January 19, 2026; Common Sense Media's assessment as reported on October 7, 2026.