OpenAI rolls out ChatGPT for Teens after a wrongful-death lawsuit

OpenAI began rolling out ChatGPT for Teens globally on August 18, a separate version of the chatbot that automatically becomes the default experience for any user the system predicts is under 18, or who has said so directly, The Verge reported. The company expects the global rollout to finish within two weeks.

The restrictions are built around tighter relational boundaries and parental visibility. Under the teen version, the model isn't supposed to call itself a friend, claim to have personal feelings, or suggest anything resembling sentience, and it delivers break reminders more frequently than the standard product. Parents can set Quiet Hours, and teens can too. Customization is limited to wallpapers rather than the fuller personalization standard ChatGPT users get, and the system is designed to notify parents when it detects policy violations such as violent threats or signs of an eating disorder — content OpenAI staff review before a parent is alerted, with a target notification window of about one hour.

“If we predict you're under 18, or you've told us so, this becomes your default experience.”

— Lauren Jonas, OpenAI's head of youth and families

The teen version also leans harder into two features OpenAI had already been building out over the past year: Study Mode, which uses a Socratic approach that pushes users toward their own answer instead of simply supplying one, and a new companion feature called Study Hours that turns Study Mode on automatically during set times of day. OpenAI says visual explanations built into the product now cover more than 300 topics, up from 70 at Study Mode's launch in March, and the system is designed to redirect students who appear to be trying to skip their homework outright. The company ran randomized controlled trials it says showed positive gains in performance from Study Mode, though it declined to share the specific numbers.

• Global rollout began August 18, expected to complete within about two weeks
• Auto-enrollment: any user OpenAI predicts is under 18 defaults to the teen version
• Model restricted from claiming friendship, feelings, or sentience with teen users
• Parental notifications for detected violent threats or eating-disorder signals, targeted within about an hour
• Visual topic explanations expanded from 70 at March launch to 300-plus now

The launch follows a wrongful-death lawsuit filed against OpenAI over 16-year-old Adam Raine, which the company cited last September when it first announced it was building a system to automatically identify teen users and restrict what they could do with the product. That case sits alongside a broader wave of legal and regulatory pressure on AI and social platforms over minors specifically: Meta was separately ordered to pay an additional $567 million in a New Mexico child-safety case earlier this month, a reminder that the platforms actually shaping how kids use AI day to day are under scrutiny well beyond any one company's chatbot. The rollout also lands next to a separate, less adversarial data point on the same question: a recent look at what kids themselves say they think about AI found most of them aren't actually asking for deeper emotional engagement with chatbots in the first place, which makes OpenAI's decision to strip out exactly that kind of relational language for teen accounts look less like a guess and more like a response to what the research already suggested.

This article is for informational purposes only and does not constitute investment advice.