OpenAI Puts a College Planner Inside ChatGPT for Teens — and Funds the Critics It Wants Feedback From

OpenAI says it is adding a College Planner to ChatGPT for Teens, the default experience applied to accounts the company identifies as belonging to users under 18, alongside new flashcard and quiz tools and multi-page note scanning. The company also published its first batch of usage statistics for the teen product and announced multi-year support for two outside organizations: College Advising Corps and the Digital Wellness Lab at Boston Children's Hospital. The numbers OpenAI chose to share are flattering, and the methodology behind them was not disclosed.

For anyone building consumer AI products, the announcement is worth reading twice. It is partly a feature release, partly a public positioning exercise on teen safety at a moment when chatbot use by minors is drawing scrutiny, and partly an attempt to structure outside input in a way OpenAI itself funds.

What OpenAI says teens are doing

OpenAI reports that ChatGPT now has 1.2 billion users overall. On the teen product specifically, it says that as the experience rolled out globally, teens with access sent about 2.7 million more learning-related messages on average than teens who did not yet have it.

That figure is the hardest one to interpret. OpenAI does not say what population the average is taken across, over what time window, how "learning-related" was classified, or whether the comparison group was matched on anything beyond access timing. A staggered rollout can produce a usable natural experiment, but the company published a headline number rather than a method. Treat it as directional at best.

Other figures are more concrete. In one week, OpenAI says nearly 1.2 million teens used Learning Visualizations for math and science concepts, and more than 180,000 used Study Mode for step-by-step support. The most-selected conversation starters were prompts asking ChatGPT to question the student's ideas, give feedback on a draft, and check evidence — all framed around critique rather than answer generation, which is the design story OpenAI wants to tell.

On time spent, OpenAI says teens average under 15 minutes a day, that fewer than 2% spend more than three consecutive hours in the product, and that in almost half of conversations where a break reminder fired, the teen took a break or ended the conversation within five minutes. Among the heaviest sessions — those running past three hours — the company says over 80% included at least one learning-related prompt.

Two caveats. First, "took a break or ended the conversation" does not tell us how long the break lasted or whether the user came straight back; roughly half of reminders apparently did not change behavior within five minutes. Second, OpenAI contrasts its under-15-minute average with a publicly reported five-hour daily average for teen social media. That comparison flatters the product but compares categories that are not alike: a task-oriented assistant and an infinite feed serve different jobs, and low session time is partly a function of what the tool is for.

College Planner: a logistics tool, deliberately scoped

OpenAI says hundreds of thousands of US teens already use ChatGPT weekly for college planning — researching schools, tracking deadlines, and understanding financial aid. College Planner is meant to consolidate that into a single structured plan covering application requirements, deadlines, tasks and financial-aid steps for the schools on a student's list, with a timeline view and the ability to ask follow-up questions in context.

The scoping is notable. OpenAI explicitly frames the tool as handling the administrative layer "without taking over the personal work" of deciding where to apply. That is the right line to draw, both because essay and school-choice automation invites obvious blowback from admissions offices, and because it leaves room for the human advisers OpenAI is simultaneously funding. Whether the product holds that line in practice — a student can always just ask ChatGPT what schools to apply to — is a separate question the announcement does not address.

The initial version is US-only, aimed at grades 10–12 and students planning on four-year colleges. OpenAI says it intends to expand to more countries and to two-year colleges, technical colleges and trade schools over time, which is where the equity argument gets more credible; four-year four-year planning is already the best-served segment. No launch date, no availability tier and no detail on whether College Planner sits behind a paid plan were given.

The financial-aid features — tracking aid deadlines, explaining requirements, surfacing scholarships and fee waivers — are the highest-value and highest-risk part. Aid rules are complex, jurisdiction-specific and change annually. An assistant that confidently misstates a FAFSA deadline or a waiver eligibility rule does real damage to the students least able to absorb it. OpenAI did not describe accuracy testing, data sourcing or update cadence for this information.

Funding the advisers and the advisory board

OpenAI says it is supporting College Advising Corps, a nonprofit that places recent graduates as advisers in under-resourced high schools, to expand into new states and launch an AI fellowship program where advisers test and share AI practices. CEO Ekaterina Struett is quoted saying the organization sees promise in College Planner as a preparation tool for adviser conversations. The size of OpenAI's contribution was not disclosed.

Separately, OpenAI is supporting the Digital Wellness Lab at Boston Children's Hospital and its Student Advisory Council for three years. For the 2026–27 school year the Lab will convene roughly 22 students across two tracks: about 12 aged 14–18 on digital wellbeing projects, and about 10 aged 16–20 focused on AI chatbots and emerging technology. OpenAI says it expects to seek feedback on teen safety defaults, parental controls, notifications and AI literacy, and commits to bringing recommendations into product decisions and reporting back on what changed.

The council is described as independent, and the Lab's leaders are quoted saying young people see what works and what fails in ways adults do not. Still, an advisory body funded by the company it advises has a structural tension, and roughly 22 students is a small sample for a product with a global teen user base. The accountability mechanism that matters is the promised report-back: whether OpenAI publishes recommendations it declined to adopt, not just the ones it shipped. The announcement does not say it will.

Study tools: small, practical, overdue

The least headline-grabbing additions may get the most use. OpenAI says students told it that photographing notes page by page took hours, so the iOS app now supports continuous multi-photo capture that stitches pages into a single PDF for upload. Android support is in progress. Flashcards can be generated from uploaded notes or a chosen topic, saved to the user's Library, marked as known or missed, shuffled, and attached to a scheduled practice task. Quizzes can now be built from uploaded notes, with better detection of when a user wants one.

These are table-stakes features in the study-app category, and their arrival inside ChatGPT puts direct pressure on standalone flashcard and quiz tools that have treated AI generation as their differentiator. The scheduled-task hook — ChatGPT prompting a student to review a deck — is the piece that turns a one-off generation into a retention loop.

The question the post does not answer

OpenAI published this two days after a post on building advertising for AI products. The teen announcement says nothing about whether advertising will appear in ChatGPT for Teens, how teen accounts are treated in ad targeting, or whether under-18 data is excluded. For a release built around trust, that silence is conspicuous. The post also does not explain how accounts are identified as under 18, which determines whether any of these protections reach the teens who most need them.

What to watch

• Whether OpenAI publishes methodology for the 2.7-million-message claim, including the comparison group and how learning-related messages were classified.
• Launch timing, pricing tier and geographic expansion for College Planner, plus any accuracy guarantees on financial-aid information.
• Whether the Digital Wellness Lab council's recommendations are published in full, including ones OpenAI chooses not to implement.
• How ChatGPT's advertising plans intersect with under-18 accounts — a point this announcement left unaddressed.
• Responses from admissions offices and school counselors, who will decide whether an AI-managed application timeline is a support tool or a problem.