‘Parents shouldn’t be asked to trust them.’ Just how safe is ChatGPT’s teen mode?

A report from the child AI safety nonprofit Common Sense Media found significant gaps in the way that ChatGPT’s safeguards for young people are implemented, and called for the company to pause offering the product to children. OpenAI introduced ChatGPT accounts for teens in August after a string of incidents in which the chatbot was involved in young people hurting or killing themselves. The ChatGPT for Teens accounts, as Straight Arrow reported, are intended for people between the age of 13 and 17, and include parental notifications for indications of mental health problems, as well as restrictions on using the product to cheat on schoolwork. This is in contrast to the approach of companies like Anthropic, which exclude anyone under the age of 18 in their Terms of Service.