By Julie Jargon
Sept. 29, 2025
In an effort to safeguard the ChatGPT experience for teens, OpenAI is now offering parental controls. The new feature will start rolling out to all ChatGPT users on Monday. Parents must have their own ChatGPT accounts to use the controls.
iStock image
In addition to the new parental controls, detailed below, parents will receive emergency notifications in some cases where their teens—or adult children who agree to link accounts—ask ChatGPT about suicide or self harm. The prompts will be reviewed by a worker before a notification goes to parents via a chosen means, such as text or email. OpenAI will provide the time of day and other relevant context, but not the transcript of the chat.
The new ChatGPT parental controls include additional safeguards, time restrictions and other options. Photo: OpenAI
The updates come after a California family last month sued the company over concerns that their teenage son’s suicide was related to troubling conversations with ChatGPT. There have also been many reported instances of people experiencing delusions and psychosis—including one that ended in a murder-suicide—after engaging with the chatbot. OpenAI said it would update its model to better support people during times of mental distress.
The new parental controls can be used by families with dependents of any age, though kids are supposed to be 13 or older to use the app. Parents can invite their kids to link accounts or teens can invite their parents. Unlike some other platforms’ settings, these stay active until they are turned off, regardless of how old the dependent is.
“It’s not age-gated. You could invite your 35-year-old child,” said Lauren Jonas, OpenAI’s head of youth well-being.
The parental controls can be found in the ChatGPT app’s settings. Parents can invite their teens, or teens can invite their parents. Photo: OpenAI
• Family invites. You can find the option in settings. After tapping Parental Controls, you’ll be prompted to invite a family member. To proceed, the invited party must accept. Parents can add as many teens as they wish, and invites can be sent via email or phone.
• Settings and options. Parents can choose whether to allow their kids to hear AI voice responses or to allow ChatGPT to create or edit images that kids might request.
They can turn off “saved memories,” which the bot uses to remember details from prior chats. In our reporting, that feature has been shown to exacerbate the bot’s tendency to “hallucinate,” or invent false information.
Parents can also set specific hours that their kids can’t use ChatGPT. Another option prevents teens’ ChatGPT conversations from being used to train OpenAI’s models.
Kids can’t turn off any of the settings that parents turn on.
• Default protections. For teens who agree to the parental controls, the restricted content experience will be on by default. That means sensitive content such as viral challenges, extreme beauty ideals and sexual, romantic or violent role-play won’t show up in ChatGPT’s answers.
(News Corp, owner of The Wall Street Journal, has a content-licensing partnership with OpenAI.)
Write to Julie Jargon at Julie.Jargon@wsj.com
This Wall Street Journal article was legally licensed by AdvisorStream.
Dow Jones & Company, Inc.
