OpenAI launches ChatGPT for Teens with study tools and anti-attachment guardrails

Dedicated defaults for underage accounts

OpenAI started the global rollout of ChatGPT for Teens on 18 August 2026, setting a dedicated interface for users aged 13 to 17. The system applies automatically to registered accounts in that age bracket and to users identified by an age-prediction system. This automated detection analyzes more than 2,000 behavioral signals, including typical login schedules. The company expects to complete the worldwide rollout across all eligible accounts within two weeks.

Lauren Jonas, head of youth and families at OpenAI, described the rollout mechanism.

There's no need for a teen to create a new account or change anything; if we predict you're under 18, or you've told us so, this becomes your default experience.

Safeguards against emotional attachment

The new version applies strict conversational boundaries to prevent users from treating the software as an emotional confidant or companion. The underlying model cannot use romantic language, present itself as sentient, or claim to possess personal feelings. In addition, the system cannot refer to itself as a friend and delivers frequent reminders that prompt users to take breaks from the screen. Visual customization options such as wallpapers, accent colors, and voice variations are intended to maintain a clear distinction between a software tool and a human relationship.

Jonas explained the limits placed on interpersonal interactions within the software.

We're adding stricter restrictions on relational boundaries: the model shouldn't call itself a friend, suggest it has personal feelings, or suggest sentience.

Educational features and parental settings

On the educational side, the software includes a Study Mode that provides step-by-step guidance and prompts rather than direct answers to homework problems. The tool triggers reminders when it detects attempts to bypass assignments, nudging students back into guided study sessions. Parents can link their accounts to manage settings, schedule Quiet Hours when access is disabled, and set predetermined times for Study Mode. Parents also receive alerts regarding high-risk interactions involving violence, self-harm, eating disorders, or explicit material, though these notifications do not allow parents to read private chat transcripts. OpenAI also announced a partnership with CodeAI to deliver educational resources on artificial intelligence.

OpenAI outlined the educational purpose of the new release.

We believe AI should expand educational opportunity, which is why we designed ChatGPT for Teens to support moments outside the classroom when help is not always available, while keeping teens engaged in an active and collaborative learning process.

Scrutiny, litigation, and market preparation

The launch follows legal actions and regulatory scrutiny regarding the safety of AI chatbots for young users. OpenAI, currently valued at $852 billion as it prepares for an initial public offering, faces lawsuits alleging that its technology harmed minors. These legal challenges include a lawsuit filed by parents after a 16-year-old user died by suicide in April 2025, which led OpenAI to update its safety protections in August 2025. Meanwhile, parent coalitions in cities including New York and Los Angeles have pressured school districts to suspend AI deployments over concerns about student cognitive development.