OpenAI launched ChatGPT for Teens, a mode that switches on by itself once the company’s systems read a user as under 18 and then restricts high-risk conversations, including ones about self-harm, violence, eating disorders, and sexually explicit or graphic material. Users give an age at signup and the terms set a floor of 13, but plenty lie, so OpenAI says it reads more than 2,000 signals, including login times and account age, which it calls minimally invasive of privacy. The mode can alert adults who opted into parental controls, and if it decides a teenager is doing homework, it’s meant to ask questions back rather than hand over the answer. Families have sued OpenAI over what their children encountered on the chatbot, and Florida sued in June over how it described safety for minors.
