OpenAI said on Aug. 19 it will launch “ChatGPT for Teens,” which applies age-appropriate learning and safety features by default for teenage users.

The OpenAI system automatically switches users to the teen environment if it estimates they are under 18 or if they state their age is 13 to 17.

A key feature of ChatGPT for Teens is stronger safety guardrails than the general user environment. It applies model- and product-level protections in high-risk areas such as self-harm, violence, eating disorders, risky behavior and explicit sexual or violent content to reduce the likelihood that teens are exposed to content inappropriate for their developmental stage.

It also restricts ChatGPT from expressing romantic feelings to teens or encouraging emotional dependence, and limits it from speaking as if it has emotions or consciousness. It provides notifications recommending a break after a certain amount of use.

Learning features have also been strengthened for teens. If a user tries to get an assignment answer directly, it steers them to Study Mode so they participate in step-by-step solution processes, and it provides quizzes and learning visualization features. It also introduces Study Hours, which turns on Study Mode by default at times set by teens or parents.

For teen accounts connected to parents, parents can set Quiet Hours and manage some settings, and in limited high-risk situations, notifications are sent to parents.

OpenAI explained that it reflected in its product design a principle that AI should support learning outside the classroom, while teachers should lead how AI is used in educational settings.