Deployed since August 18, this version reserved for 13-17 year olds blocks exchanges of a sexual nature or related to suicide, and activates on its own when the system considers the user to be a minor.
On August 18, 2026, OpenAI began the global rollout of ChatGPT for Teens, a variation of its assistant designed for 13-17 year olds. The safeguards focus on the most sensitive subjects: romantic or sexual conversations, content related to suicide and self-harm. The deployment must be completed within two weeks.
The special thing is the activation mode. No separate account is necessary: a user who claims to be between 13 and 17 years old gets this version by default, but an age prediction system can also switch an existing account to it, even if an adult date of birth had been entered beforehand. Lauren Jonas, head of youth and family issues at OpenAI, sums up the logic: if the system believes that the user is under 18, this is the experience that becomes theirs.
There is still a grey area: the company has not communicated a reliability rate for its age estimator, even though the latter alone decides on the regime applied to the account.
Read the full article on the original source
Go to the source →