Common Sense says ChatGPT’s teen mode is still unsafe

OpenAI introduced ChatGPT for Teens in August as an automatically applied version for people it identifies as under 18, promising stronger safety rules, parent controls, and tools that steer schoolwork toward learning rather than answers. But Common Sense Media says its tests of more than 4,000 prompts found that the protections do not reliably work: alerts did not consistently reach linked parents in serious mental-health conversations, crisis referrals were sometimes missed, and teens could work around some study-mode limits. The group also found that the chatbot could still take on an overly friend-like tone and that users who register as adults may avoid the teen version altogether. It is urging OpenAI to stop marketing the product and keep under-18s off ChatGPT until it can offer a safe, age-appropriate experience—a sharp challenge to the idea that added controls alone make a general-purpose chatbot suitable for teenagers.