The unveiling of the ChatGPT for Teens program at the Taj Lands End in Bandra came during their Open Dialogue on Youth Safety and AI event at which actors Soha Ali Khan and Tisca Chopra were also present, and during which a companion piece called Parents’ Guide to Learning with ChatGPT was announced.
Also read: Windows 11 laptop battery draining fast? Quick fixes that will help
The key number behind all of this: India has the biggest student body using ChatGPT out of all of OpenAI’s students across the globe. According to Pragya Misra, Head of Strategy & Global Affairs for OpenAI in India, the magnitude of this was stated onstage.”One of the things we realized is that learning is actually the use case that ChatGPT is most useful [for],” Misra said, describing how that insight shaped everything from Study Mode to today’s teen-specific rollout.
ChatGPT for Teens is not an option but the default mode for users. OpenAI systems attempt to gauge age based on certain signals such as handwriting in images, phrasing of prompts and even patterns of usage (the way a teenager uses ChatGPT after an hour of school versus an adult doing the same). If OpenAI identifies a user between the ages of 13 and 17 or identifies them as explicitly identifying themselves as being in that range, the teen experience kicks in automatically. Otherwise, OpenAI takes the safer route and defaults to the teen mode rather than the adult one.
Among these features are Study Mode where questions are answered with prompts in a Socratic manner. The Responsible Homework Reminder will remind a user to do work rather than simply copy-paste answers from ChatGPT. There are quizzes and visualizations of learning processes created with the help of teachers and child psychologists. And Study Hours where parents/teens can lock their ChatGPT session in study mode during specified hours.
Guardrails become even more robust automatically in the case of self-harm, violence, eating disorders, and sexual content, in addition to specific design considerations that ensure that the language remains parent-like and not peer-like nor romantic-like, as per Misra, to steer clear of the emotional dependency which OpenAI has noticed from teenagers on AI.
Parental Controls enable a parent account linked to the child’s to manage some of these guardrails through Quiet Hours, sensitivity level control, and Study Hours activation. While parents cannot monitor the content of their teenager’s chats with AI, they can be notified under certain very high-risk circumstances, and also separately, when OpenAI decides to suspend the teenager’s account due to a violation of policy.
Also read: Anthropic CEO Dario Amodei on AI’s trust problem: Why skepticism isn’t going away
In its demonstration of Codex, an AI agent specialized for coding tasks, OpenAI reserved Codex a separate slide in the presentation entitled “Building with Codex.” The pitch: the information about how kids spend time at school is dispersed across the school portal, email, calendar, app for sports activities, reminders about payments, and WhatsApp; all those pieces of information get updated at various hours of the day. The concept illustrated in the slide: the user (parent) asks Codex what should be paid attention to, and it collects all those scattered pieces of information (the worksheet is due tomorrow, you need to sign the permission slip, and football practice is now scheduled at 5:30 PM) into one place, the illustrated “Family Dashboard” with tabs for academics, calendar, activities, messages, forms and fees, and wellbeing. OpenAI used explicitly clear framing regarding parents’ control: “Codex reads, organizes, and suggests actions” with the “review and approve” process when, according to the slide, “you stay in control.”
It may be pointed out that the slide is a concept illustration with a mockup but not a real demo or a screenshot from the shipped version; also, the concept is outside the feature list of ChatGPT for Teens.
This new announcement is not a singular launch but is actually part of an overall rollout process that has been planned by OpenAI for the last year. Study Mode was officially launched in August 2025 as OpenAI invited educators and child psychologists to develop the model according to the way it should react to the students and not just give answers. The ChatGPT for teens announced today can actually be considered as the enhanced version of the same concept.
As India is home to the largest number of students using ChatGPT from OpenAI, the new features, such as age prediction, study mode as default, and notifications for parents based on policy violations, are all targeted towards this audience. In her personal testimony, Misra revealed that she had always been a student who was “always looking for shortcuts.” It now remains to be seen whether the age prediction feature, which depends on handwriting and usage behavior, works effectively with this group of students.
Also read: We are designing monitors according to how Indians actually work: BenQ’s Rajeev Singh