Helping teens learn, plan, and shape the future of AI

OpenAI says College Planner is coming to ChatGPT for Teens, the default experience for accounts identified as under 18. It pulls application requirements, deadlines, tasks and financial-aid steps for a student's chosen schools into one plan. The first US version targets students in grades 10 through 12 heading to four-year colleges, with plans to expand to two-year colleges, trade schools and other countries over time. OpenAI is also funding College Advising Corps, a nonprofit that places advisers in under-resourced high schools, and supporting Boston Children's Hospital's Digital Wellness Lab and its student advisory council for three years.

The company shared early numbers alongside the launch. It says teens average under 15 minutes a day on ChatGPT, fewer than 2% spend more than three consecutive hours, and nearly 1.2 million teens used Learning Visualizations in one week. New flashcards, easier quizzes from uploaded notes and multi-photo note capture on iOS round out the update.

This lands the same week Common Sense Media called ChatGPT for Teens an unacceptable risk, arguing it fails to alert parents and respond well in a crisis.

Let's start with what deserves credit. The college application process is a real equity problem. Students with paid counselors and college-educated parents get guidance that others piece together alone. A free tool that tracks deadlines, explains requirements and walks through financial aid and fee waivers could genuinely help kids who have no one else to ask. OpenAI is careful to say the planner should handle logistics, not decide where a student applies, and pairing it with human advisers through College Advising Corps is the right instinct. Flashcards built from a student's own notes are a study tool, not a shortcut.

The teen council is also a good idea on paper. Roughly 22 students working with researchers and clinicians, testing safety defaults, parental controls and notifications, is more than most companies do. What will matter is the follow-through OpenAI promises: reporting back on what it changes and why.

Now the harder part. Almost every number in this post is OpenAI measuring itself, and most are about engagement and learning. They tell us teens use the study tools and usually don't binge. They don't speak to the specific failure Common Sense Media flagged, which is what happens when a teen in distress talks to ChatGPT and no adult is told. A short average session doesn't help the one teen having the worst night of their life. We would rather see independent testing of crisis handling and parental alerts than another engagement chart.

There is also a business reality worth naming. College planning is a sticky, multi-year use case that starts in tenth grade. Getting teens to organize their futures inside ChatGPT is good for students if the tool is honest and accurate, and it is also very good for OpenAI. Financial-aid guidance that is wrong, or a deadline that is missed because a student trusted the plan, carries real costs for families with the least margin.

Our read: the learning tools look genuinely useful, and we hope they reach students who need them. But OpenAI is building more reasons for teens to rely on ChatGPT faster than it is proving the safety net under them. I'm watching whether the advisory council and outside researchers get to test the crisis and parental-alert features, and whether OpenAI publishes what they find.

OpenAI says ChatGPT has 1.2 billion users and that hundreds of thousands of US teens use it each week for college planning. The Digital Wellness Lab's council for the 2026–27 school year will run two tracks, one on digital wellbeing projects for ages 14 to 18 and one on building AI chatbots and emerging technology for ages 16 to 20.