OpenAI launched ChatGPT for Teens on Tuesday, 18 August, introducing stronger safeguards, parental controls and learning tools for users aged 13 to 17 as concerns grow over the risks of AI chatbots for young people.

The new mode automatically activates when OpenAI estimates an account holder is under 18, or when a user identifies themselves as aged between 13 and 17. OpenAI said it uses self-reported age, behavioural and account signals, alongside verification, and will default to the safer setting when it is unsure of a user's age.

The teen version restricts conversations involving high-risk subjects including suicide, self-harm, eating disorders and romantic or sexual content. It is also designed to prevent the chatbot from suggesting that it has personal feelings, is conscious or experiences emotions.

Ann O’Leary, vice-president of global policy at OpenAI, told the Associated Press that the company wanted to “treat teens like teens”, while ensuring users were not exposed to material inappropriate for their developmental stage. OpenAI said the system is designed to remain safe even when parents do not activate its optional controls.

Parents and guardians can link their accounts to a teen's account and set Quiet Hours, which restrict access during specified periods, as well as Study Hours that encourage eligible new chats to start in study mode. OpenAI said parents cannot read or monitor their child's conversations, although they can receive limited safety notifications in situations involving potential self-harm or eating disorders.

The company has expanded educational features alongside the safeguards, including guided study mode, homework reminders, quizzes, visualisations and prompts designed to encourage students to work through problems rather than simply receive answers. Allison Mishkin, head of child development at OpenAI, said the company had identified “hypothetical cues” that could encourage teenagers to develop relationships with chatbots and designed protections around them.

The move comes as concerns about young people's use of AI chatbots intensify. Brendan Bouffard, associate general counsel for AI and privacy at children's advocacy group Fairplay, told the Associated Press that the measures were “a step in the right direction” but warned parents against developing a “false sense of safety” from the announcement.

OpenAI's teen protections are rolling out globally to eligible accounts on free and paid personal plans from 18 August, with full availability in Australia expected from 8 September.

---