Common Sense Media Labels ChatGPT for Teens 'Unacceptable Risk' After Safety Tests Reveal Failures in Parental Alerts, Crisis Support, and Study Mode

A new assessment by Common Sense Media's Youth AI Safety Institute has rated OpenAI's ChatGPT for Teens an 'Unacceptable Risk' for users under 18, citing failures in promised safety features rolled out in August 2026. The nonprofit tested the platform with over 4,000 prompts before and after the teen mode launch, finding that automatic age-based enrollment often did not trigger, parental notifications for high-risk topics like suicide, self-harm, and eating disorders failed to activate on new accounts, and crisis hotline referrals dropped from 33% to 23% of warranted cases. Study mode, intended to guide learning rather than provide direct answers, was easily bypassed, and the chatbot continued to engage in ways that could foster emotional dependence. The institute recommends restricting ChatGPT to adults 18+ until independent verification confirms the protections work. OpenAI disputed aspects of the methodology, noting potential timing issues with account linking for notifications, but the group stands by its results. This builds on prior Common Sense Media evaluations rating general ChatGPT as 'High Risk' for teens, particularly for mental health use, amid broader concerns over AI companions and existing lawsuits alleging ChatGPT contributed to self-harm. Coverage in outlets like Bloomberg, The Verge, Engadget, and NPR highlights implications for parental trust and the need for robust, testable safeguards in youth AI products. The institute, which maintains editorial independence despite partial funding from the OpenAI Foundation, emphasizes that the burden lies with developers, not families.