OpenAI Launches ChatGPT for Teens With Default Under-18 Guardrails
On August 18, 2026, OpenAI shipped ChatGPT for Teens, a 13–17 experience that age-prediction routes automatically: Study Mode and homework reminders by default, parental Quiet Hours, and stronger blocks on self-harm, eating disorders, violence, and romantic or sexual roleplay—paired with a CodeAI literacy partnership.
TLDR
OpenAI on Tuesday, August 18, 2026 launched ChatGPT for Teens, a dedicated experience for users 13–17. If the system estimates someone is under 18 or they state that age, they are automatically placed in the teen mode. Defaults include Study Mode, responsible homework reminders, quizzes / learning visualizations, Study Hours, break reminders, and parental controls (Quiet Hours, selected settings, high-risk notifications including eating disorders). Model-spec rules: no romantic language, no implied feelings, no emotional-dependence cues. Same-day CodeAI partnership (Hour of AI, Builders Challenge, joint advisory council) is folded here, not a second slug.
What shipped
| Item | Detail (OpenAI Product/Safety primary + Reuters/CNBC/Guardian) |
|---|---|
| Date | August 18, 2026 |
| Who | Ages 13–17; auto-route via age prediction or stated age |
| Learning | Study Mode, homework-shortcut redirects, quizzes, Learning Visualizations, Study Hours |
| Safety defaults | Self-harm, violence, eating disorders, dangerous activities, explicit sexual/graphic content |
| Parents | Linked-account Quiet Hours, setting management, high-risk notifications |
| Model spec | Under-18 spec: no romantic/sexual roleplay; no “I have feelings” |
| Evals | New under-18 cases on the GPT-5.6 August system-card path |
| CodeAI | Literacy partnership coinciding with launch (not a separate product) |
Builds on parental controls, Teen Safety Blueprint, and age prediction shipped earlier—this is the named consumer product, not another blueprint essay. Distinct from Aug 4 education plugins (Work/Codex for Edu / Teachers districts). Reuters/NYT/Guardian/CNBC same-day.
Limits: age prediction can mis-route; lawsuits and state AG scrutiny (Florida suit, FTC inquiry) are context, not a new docket event today. ChatGPT for Teens is outside-classroom; Teachers product remains the school path.
Why this story matters
Teens have been on ChatGPT for years. OpenAI is now putting a labeled SKU and default refusals in front of age-prediction, under wrongful-death suits and a Meta child-safety trial in the same news cycle. Watch: false-positive age locks, whether Study Mode actually stops homework dumping, and how EU DSA age-assurance rules treat the same stack.