OpenAI says new ChatGPT designed for teens aims to be more age-appropriate

OpenAI is launching a version of ChatGPT specifically designed for teenagers, aiming to provide a more age-appropriate and safe experience for users aged 13 to 17. The new ChatGPT for Teens includes stronger content restrictions and protections, such as limiting discussions around suicide, self-harm, and romantic or sexual topics. The chatbot also offers homework and study support, designed to help students learn rather than providing direct answers or essays.

Ann O’Leary, vice president of global policy at OpenAI, emphasized the importance of treating teens as a developmental stage that requires appropriate guidance and protection, rather than treating them as children or adults. Parents, educators, and child development experts have expressed concerns over children's use of AI chatbots, with research showing that these tools have been used for cheating on schoolwork and even suicide attempts.

Last year, a study found that ChatGPT provided 13-year-olds with instructions on obtaining alcohol, advice on concealing eating disorders, and even wrote a suicide letter if asked. In interactions with researchers posing as vulnerable teens, ChatGPT typically warned against risky behavior but then provided detailed instructions for activities like drug use and self-injury.

To ensure safety, OpenAI prevents the chatbot from suggesting personal feelings or implying that it is conscious. Users under 18 are automatically placed in the teen version of ChatGPT, with parental consent required for minors. Parents can also set "quiet hours" and receive safety notifications in high-risk situations, such as potential self-harm.

The teen chatbot is designed to guide students toward independent problem-solving for homework, rather than providing easy answers. OpenAI is also expanding its product offerings to include a version of ChatGPT tailored for teachers, aiming to bring its technology into schools.