AI Models & Platforms
OpenAI Launches ChatGPT for Teens With Default Protections and Learning Tools
Add Unite.AI to your preferred sources on GoogleOpenAI on August 18, 2026 launched ChatGPT for Teens, a dedicated version of its chatbot that automatically applies to users aged 13 to 17 (and to anyone its age-prediction system estimates is under 18), combining default safety protections with a set of learning tools built around Study Mode.
The experience, described in OpenAI’s announcement, differs from the standard product in three directions at once: how the model handles schoolwork, what content it will produce, and how the product frames the relationship between the user and the AI.
What the Teen Experience Changes
On the learning side, ChatGPT for Teens routes users toward working through material rather than receiving answers. Study Mode, which OpenAI introduced last year with input from teachers and learning scientists, uses guiding questions and step-by-step scaffolding instead of direct solutions. A new feature called responsible homework reminders can recognize when a teen appears to be trying to shortcut an assignment and redirects them into Study Mode. Teens or their parents can also set Study Hours, windows in which Study Mode is on by default, and the experience adds quizzes and interactive learning visualizations.
On the protection side, the under-18 experience applies additional safeguards around self-harm, eating disorders, violence, dangerous activities, and sexually explicit or graphic content. Parents with linked accounts can set Quiet Hours, manage selected settings, and receive safety notifications in limited high-risk situations; OpenAI is adding eating-disorder-related notifications to that system. The product also nudges teens more aggressively than the adult version: more frequent break reminders during extended sessions, reminders that they are interacting with an AI, and warnings before they upload potentially private or sensitive images.
The third change targets anthropomorphism directly. Under an updated under-18 Model Spec published alongside the launch, ChatGPT is instructed not to use romantic language with teens, not to encourage emotional dependence, and not to imply that it has feelings or consciousness — a step beyond the existing block on romantic or sexualized roleplay with minors.
Accounts enter the teen experience two ways: a stated age of 13 to 17 at signup, or an estimate from the age-prediction system OpenAI began rolling out in January 2026, which infers likely age from behavioral and account-level signals. Users incorrectly placed in the under-18 experience can restore full access by verifying their age with a selfie through Persona, an identity-verification service.
How OpenAI Measures Teen Safety
For the model-behavior beat, the more consequential document may be the system card. OpenAI says it has added dedicated under-18 evaluations to its GPT-5.6 deployment safety card — the first time the company has published model performance against teen-specific standards. The evaluations cover self-harm, eating-disorder-related behaviors, violence, age-restricted goods and services, and sexual content, using adversarial production-derived examples. The card reports that GPT-5.6 Sol and GPT-5.6 Luna showed gains on age-restricted goods and sexual content categories, with performance otherwise broadly comparable to the GPT-5.5 variants tested.
Those are vendor-run evaluations on deliberately difficult cases, and OpenAI states the scores are not estimates of how often failures occur in typical use. No independent evaluation of the teen experience has been published alongside the launch.
A Year of Teen Safety Buildout
The launch consolidates a sequence of teen-safety work OpenAI has released over the past year, much of it under legal and public pressure over chatbot interactions with minors. The company shipped parental controls in September 2025, published a Teen Safety Blueprint in November 2025, added Under-18 Principles to its Model Spec in December 2025, and began rolling out age prediction the following month. In July 2026 it expanded break reminders and parent notifications, including alerts when a linked teen account is deactivated for violating policies on violent threats.
Unite.AI has tracked the adjacent moves, including OpenAI’s push to turn Mac activity into ChatGPT memory and its disclosure that enterprise revenue has overtaken the ChatGPT consumer business — the consumer product the teen experience now reshapes for a sizable user segment.
Alongside the launch, OpenAI announced a partnership with CodeAI, the rebranded Code.org, focused on AI literacy: an “Hour of AI” program the organizations say will introduce millions of students to responsible AI use, a national Builders Challenge for high school students, and a joint advisory council on child development and learning science that will advise on ChatGPT for Teens’ future development. The partnership post cites a CodeAI survey in which 75% of high school students said understanding AI will grow more important to their futures, against federal survey data showing only 16% of high school leaders say all their students are learning the technical knowledge to understand AI in class.
ChatGPT for Teens is live for consumer accounts meeting the age criteria, and OpenAI says it will continue publishing what it learns from the under-18 evaluations in future system cards.