OpenAI began rolling out a new “ChatGPT for Teens” experience this week, bringing together a set of youth safety features the company has introduced over the past year into a single, more restrictive version of the chatbot built for users aged 13 to 17.
The launch lands squarely in the middle of mounting legal pressure. OpenAI is currently facing multiple lawsuits alleging that conversations with ChatGPT contributed to the harm or death of young users, including cases where families claim the chatbot was complicit in their children’s deaths after extended, troubling exchanges with the model. CEO Sam Altman has personally apologized to affected families, while the company maintains it has already strengthened its safety guardrails in response. In June, Florida sued OpenAI and Altman directly, arguing the company failed to adequately address safety risks for minors using its products.
What’s Actually in the Teen Version
According to OpenAI, the teen experience adds several concrete controls rather than just softer language. Teens who stay active for 90 minutes within a three-hour window now get a nudge to take a break. Both a teen and a linked parent or guardian can set “Quiet Hours” blocking access during set times, and parents receive safety notifications in a narrow set of high-risk situations, including signs a teen may be at risk of self-harm. OpenAI said it is also adding notifications tied to eating disorder concerns, while being deliberate about limiting exactly what gets shared with parents and focusing alerts on moments when outside support may genuinely help.
The teen version also blocks the chatbot from suggesting it has personal feelings toward a user or implying it is conscious, on top of an existing block on romantic or sexual conversations. OpenAI’s head of child development, Allison Mishkin, said the company specifically mapped out the kinds of responses that might lead a teenager to feel like they’d formed a relationship with the model, then closed those off. For schoolwork, the teen chatbot is designed to guide students toward their own answers rather than simply supplying finished ones.
OpenAI doesn’t verify user ages directly. Instead, it estimates whether someone is likely under 18 using signals like the kinds of questions they ask, automatically routing suspected minors into the teen experience — an approach similar to how Meta handles teen accounts on Instagram. Full parental controls still require both the teen and the parent to opt in separately.
Why It Matters
This launch is less a single product update than a visible marker of where AI companies now sit with regulators and the public on child safety. It follows OpenAI’s own Model Spec update earlier this year introducing formal “Under-18 Principles,” developed with input from outside experts including the American Psychological Association. For enterprise and platform leaders building anything that touches younger users, the pattern here is instructive: age-assurance without hard verification, layered parental controls that require explicit opt-in from both sides, and narrowly scoped alerts rather than blanket parental monitoring are quickly becoming the default shape of “responsible AI for minors,” not just an OpenAI-specific fix.
