ChatGPT Now Has Parental Controls — Here's How They Work

OpenAI launched parental controls for ChatGPT users on its web version, according to announcements from the company on their blog post and on social media platform 'X'. The mobile version of the app will be released "soon" the company said.

The AI giant also posted some answers in the reply to their 'X' post to what they anticipated to be frequently asked questions that would be posed to them.

"We’ve worked closely with experts, advocacy groups, and policymakers to help inform our approach — we expect to refine and expand on these controls over time," the company said.

The company also posted a link to its resource page which broke down how the controls work, how they can be accessed and what parents can do with them.

Parental controls work via a link set up between a parent and their teenage child (children under 13 are not meant to use ChatGPT, according to OpenAI). Parents can access ChatGPT via the Parental controls setting on the web version of the app or enter the URL — chatgpt.com/parent-resources.

Teenagers can be invited via e-mail or text message and once they accept, parents can start using the controls. The child also has the option to unlink their accounts, but parents will be notified if the account is disconnected from the controls.

However, the teenagers will still have partial privacy as OpenAI stated that parents won't have access to their children's chats. If ChatGPT's system as well as its trained reviewers notice a safety risk, parents will be notified with only the relevant part of their child's chats being flagged to them.

Parents can receive the notification via e-mail, SMS, push notification or all three.

The controls will allow parents to restrict features such as access to image generation, graphic content and viral challenges. The amount of time that a child has access to the AI can also be limited, via the 'Quiet Hours' feature.

Parents can also remove their child's access to voice mode and disable the chatbot's access to transcripts to train itself. They can also turn off the 'memory' feature so it won't record or remember past chats.

It also has a built-in safety notification where it will alert the parent if it detects a "serious safety risk" for the child.