OpenAI Is Rolling Out ChatGPT For Teens Globally With Stronger Protections

OpenAI is expanding its ChatGPT chatbot to include a version specifically tailored for teenagers, designated as ChatGPT for Teens. Aimed at users aged between 13 and 17, this new iteration of ChatGPT comes with enhanced safety features, parental control options, and additional study support to encourage learning. Key features of the new model include a Study Mode, which guides students through problems step-by-step instead of simply providing answers.

Additionally, a responsible homework reminder helps prevent students from attempting to shortcut their assignments. If the system identifies an attempt to bypass this mode, users are automatically redirected back to Study Mode. The model also incorporates quizzes and visual aids at the end of lessons to ensure comprehension. A new Study Hours feature allows teens and parents to designate specific periods when Study Mode is enabled by default.

Furthermore, OpenAI is collaborating with CodeAI to promote AI literacy among students and teachers. Once ChatGPT determines that a user is under 18, they are automatically placed into ChatGPT for Teens, which enforces a set of stronger safeguards and parental controls. These protections prevent minors from discussing sensitive topics such as self-harm, violence, and explicit sexual content.

Parents with linked teen accounts can set Quiet Hours to restrict access during certain times of the day and manage selected settings, receiving safety notifications in the process. To prevent romantic or sexualized interactions, ChatGPT for Teens will block such roleplays and avoid using romantic language or implying the bot has feelings or consciousness.

The chatbot also sends break reminders and alerts users before uploading potentially sensitive images. The safety measures for teens are based on research involving developmental science, expert input, and rigorous testing of AI guardrails to identify and address potential risks. OpenAI is sharing its research publicly, beginning with the evaluations conducted for users under 18.