OpenAI launches a ChatGPT version built for teens

Details

OpenAI launched ChatGPT for Teens on August 18, 2026, a dedicated mode that automatically activates for users identified or estimated to be between 13 and 17 years old. The teen variant enforces stricter content filters covering suicide, self-harm, eating disorders, and sexual content, while also prohibiting the model from simulating emotions. On homework assistance, the AI redirects rather than answering directly, prompting students with follow-up questions designed to encourage independent thinking. Age detection relies not on self-reporting but on over 2,000 behavioral signals — login patterns, usage times, and other indicators — to infer age and trigger the teen experience automatically. This passive detection sidesteps the verification problem but raises questions about false positives and enforcement reliability. The launch arrives under visible legal pressure: multiple families have sued OpenAI after children encountered harmful content and were seriously injured or died. Florida filed the first U.S. state lawsuit against the company in June 2026. Notably, many of the protections bundled into 'teen mode' are not entirely new — parental controls and study mode existed previously, while age prediction was introduced earlier this year. The announcement consolidates existing features with some additions rather than introducing wholly new safety architecture.