OpenAI launches ChatGPT for teens
Users the system estimates or who declare themselves to be 13-17 are defaulted into a restricted mode that blocks romantic role-play and claims of AI sentience, arriving amid wrongful-death litigation.
- Government & policy
- Culture & impact
- Notable
OpenAI launched ChatGPT for Teens, a restricted mode for 13-to-17-year-olds that activates automatically when the company’s age-estimation systems judge a user is likely under 18, or when someone states an age in that range. The mode limits conversations touching on self-harm, violence, eating disorders and explicit content, and is designed to discourage romantic or sexual role-play, emotional dependence on the chatbot, and any suggestion that the AI itself is sentient. It adds a Study Mode intended to walk teenagers through problems with guiding questions rather than direct answers, redirecting there when the system detects likely cheating on homework, and prompts a break reminder after 90 minutes of use within a three-hour window.
Parents who link an account can set “Quiet Hours” that turn ChatGPT off, adjust selected safety settings, and receive notifications in limited high-risk situations, but OpenAI said they cannot read a teenager’s conversations directly. The product builds on the Teen Safety Blueprint OpenAI published in November 2025, which set out similar principles without committing to specific product changes, and on its partnership with the American Psychological Association on youth mental health guidance two weeks earlier.
The launch came against a backdrop of sustained legal and regulatory pressure: a wrongful-death suit filed by the parents of Adam Raine alleging ChatGPT contributed to their son’s suicide, a wave of further suits filed months later, and state and federal legislative proposals aimed at chatbot safety for minors. TechCrunch and other coverage noted the product arrived years after teenagers had already become heavy ChatGPT users, and critics faulted its predecessor blueprint for relying on an opt-in parental-control system of the kind that, on other platforms, sees only limited use — Snapchat and Discord separately disclosed to the US Senate Judiciary Committee that fewer than 1% of under-18 accounts had a parent using their built-in monitoring tools.