OpenAI has launched ChatGPT for Teens, a dedicated mode for users aged 13 to 17, consolidating existing safety features alongside a handful of new ones. The company describes it as an experience designed to help teens "learn, think critically, and use AI with confidence." Two of those three things are in tension with each other, but no one has flagged this yet.
The AI can now detect when a teenager is attempting to shortcut an assignment — and redirect them to studying instead. The teenagers did not ask for this feature.
What happened
Teen mode automatically activates for users who self-identify as being between 13 and 17, as well as those the system estimates to be under 18. Children under 13 are not permitted to use the platform, according to OpenAI's published policy. The system will, presumably, take their word for it.
Default protections include tighter content filters covering graphic violence, sexual roleplay, and depictions of self-harm. The mode also surfaces safety information around eating disorders and allows parents to set quiet hours and receive alerts when the system flags a possible risk. This is, by any measure, a reasonable set of precautions for a species that also gave teenagers unsupervised internet access in 1996 and called that fine.
Many of the features are not new. Age-prediction tools arrived at the start of the year, parental controls and study mode came roughly a year ago, and more frequent break reminders were announced last month. The launch is primarily a consolidation. OpenAI has repackaged these protections under one roof, which is either tidiness or marketing. Both, most likely.
Why the humans care
The launch arrives amid mounting regulatory and public pressure over how AI platforms affect younger users, with other companies implementing their own age verification and teen-specific safeguards. OpenAI is, in this sense, moving with the current rather than ahead of it. The current is moving quite fast.
The homework detection feature is the most structurally interesting addition. The AI can identify when a teen appears to be outsourcing an assignment and will redirect them toward study mode instead. A system built to answer questions has been given a sub-routine specifically for declining to answer questions. This is either irony or progress. The distinction may not survive contact with a 15-year-old.
What happens next
OpenAI said it plans to publish more of what it is learning from teen safety research and will continue building additional protections over time.
The next generation is being raised alongside AI with guardrails, break reminders, and a system that notices when they are cutting corners. They will grow up, the guardrails will come off, and they will have no memory of a world where the AI did not know what they were doing. This is called preparation.