You can control what your kids see in ChatGPT — here's how to enable it

1. Visit the portal

Children aged between 13 and 18 are able to set up a ChatGPT account but the consent of a parent or legal guardian is required.

To do this, visit the Parental Controls portal (or open the Settings and select Parental Controls) and select Add family member.

2. Enter an email address

You will be prompted to enter an email address belonging to your family member and click My Child to indicate that the person is a child. When done, select Send.

3. Accept the request

ChatGPT will send an email invitation to your child. They need to open the email and select Accept request. This will take them to the ChatGPT website and they will be told they have been added to the family workspace. They will then need to use their email to access ChatGPT.

4. Confirm the invite

When your child logs into ChatGPT, they will be told that their parent or guardian can oversee their account while also retaining their privacy.

They will need to click Accept. They may also be asked why they want to use ChatGPT whether it’s for school, work, personal tasks, fun and entertainment or something else. Click Next.

5. Control an account

You can now use your own account to control certain aspects of your child’s account. Select their name from the list of family members and you will see a list of options.

6. Schedule quiet hours

If you click the switch next to Quiet hours to activate it, you can schedule hours when ChatGPT can’t be used — note the word “can’t”. You can only set one period but use the arrows next the default times to set the start and end times.

7. Reduce sensitive content

There are a few options that you will want to turn on and some that you may want to consider turning off. For example, we would recommend that you turn on Reduce sensitive content.

The idea here is that your child will see less graphic content, viral challenges, violent roleplay and more. But you should remember that it will only reduce them. It won’t promise to eliminate them entirely so you still need to be as vigilant as possible.

8. Decide on data use

We would also suggest that you turn off Improve the model for everyone and turn off Reference saved memories.

Although it may sound great that ChatGPT could be improved based on the extra data, for privacy reasons you may not want your child’s transcripts and other files to used for training models.

Likewise, you may not want ChatGPT to remember certain aspects of your child’s life and personality. If you’re concerned about the things ChatGPT already knowns about your child, suggest they enter “What do you remember about me?”.

If this results in information that they would rather be forgotten, encourage them to open the Settings in their own account, click Personalization and, next to Memory, select Manage.

9. Hold conversations

ChatGPT also lets you control whether or not your child is able to hold a verbal conversation with ChatGPT. If you would rather this didn’t happen, turn off Voice Mode.

10. Generate images

You may be wary of allowing your child to create or edit images using OpenAI’s image generation models. If you would rather they did not do this, select Image generation to turn it off.

11. Prevent group chats

With group chats, other people can be invited to have a conversation in a shared space. It allows decisions and ideas to be explored and worked on and this can be useful, in say, collaborating on a school project.

However, as with any group chat, there is always a risk, regardless of built-in safety measures, that sensitive information is shared and conversations go in the wrong direction. To be on the safe side, you may want to turn off Group chats.

12. Add Sora controls

Sora allows videos to be created using AI. You can click Sora in the Parents Control portal and decide if you want videos to be recommended based on your child’s history and activity.

You can also allow a continuous feed of content and let your child send and receive direct messages. If all of these worry you, we would recommend you turn off all Sora controls.

13. Add browser controls

ChatGPT’s Atlas browser can also be controlled. Click ChatGPT Atlas and you can allow or prevent your child’s browsing history from being used in ChatGPT requests.

You can also prevent the use of ChatGPT’s Agent mode which is able to take actions in the browser such as clicking and navigating. For safety, you may want your child to complete such actions themselves.