OpenAI shares early data on teen ChatGPT use

OpenAI's 7 October update reports substantial use of teen learning tools. The figures describe engagement, not demonstrated gains in learning.

OpenAI has shared an early snapshot of how teenagers use its learning-focused ChatGPT experience. Its 7 October update reports substantial activity in study tools, while leaving an important distinction unresolved: using a learning feature is not the same as demonstrating that learning improved.

What the company measured

In an unspecified week, the company says nearly 1.2 million teens used Learning Visualizations and more than 180,000 used Study Mode. It reports average teen use below 15 minutes daily. OpenAI also reports about 2.7 million additional learning-related messages, on average, among groups given the teen experience compared with groups not yet given access.

The announcement does not provide the sample size, observation period or a sufficiently detailed denominator for that comparison. It therefore cannot establish a per-student improvement, a representative estimate for all teenagers, or a causal effect on achievement.

An update, not a first-ever finding

Nor is this OpenAI's first published teen-use statistic. On 16 July, it said nearly nine in ten teen ChatGPT users used it for learning, information, skill-building or productivity in a week. The October release is a later update about a more specific product experience.

Age, safeguards and privacy are separate questions

ChatGPT for Teens was announced on 18 August. OpenAI said accounts estimated to belong to someone under 18, or with a stated age of 13–17, would enter the experience automatically. Its design includes guided study, homework prompts, quizzes and age-appropriate safeguards. That describes the intended product behavior; it is not an independent audit of how reliably each protection works.

The age category also requires care. OpenAI's help documentation says age prediction can use account and usage signals and can make mistakes. An adult incorrectly placed in the teen experience can review available verification options. According to the same documentation, Persona handles ID or selfie verification; OpenAI says it does not receive those documents, although it receives age-related information needed to determine the account experience. The company says Persona deletes uploaded verification documents within seven days. These are the providers' stated practices, not a privacy audit conducted for this article.

Parental controls address a different boundary. Linked parents can manage selected settings and schedules, but OpenAI says linking does not expose a teenager's chat transcripts or real-time activity. Limited safety notifications are not continuous monitoring. Data-use settings also deserve attention: families should review whether conversations can be used to improve models, rather than assuming that a teen account automatically switches that use off.

Learning needs a different test

Independent research explains why engagement and achievement should be assessed separately. A 2024 Penn/Wharton experiment involving nearly 1,000 high-school mathematics students in Turkey compared unrestricted AI assistance, a teacher-informed tutor and traditional resources. The university's account reported better assisted practice but worse subsequent unaided test performance for the unrestricted group; the guarded tutor avoided that decline. That study did not evaluate the current ChatGPT for Teens product, and it should not be treated as a verdict on it.

What to watch next

For families and educators, the useful next question is what a learner can explain, check and apply without the assistant. Future disclosures would be more informative with clearly defined cohorts, dated observation windows, independently assessed learning outcomes and results for different groups. Usage averages can guide that investigation, but should not substitute for it. A productive AI study habit still needs room for the student's own reasoning and for support from teachers, advisers and family.

*Reporting note: Prepared with AI assistance and checked against the linked sources. Company usage figures are attributed claims; no private teen conversations or personal data were accessed for this report.*

Sources

- [Helping teens learn, plan, and shape the future of AI](https://openai.com/index/teens-learn-and-plan/) — OpenAI, 2026-10-07 - [Why teens deserve access to safe AI](https://openai.com/index/why-teens-deserve-access-safe-ai/) — OpenAI, 2026-07-16 - [Introducing ChatGPT for Teens: Built for learning, backed by protections](https://openai.com/index/chatgpt-for-teens/) — OpenAI, 2026-08-18 - [Age prediction in ChatGPT](https://help.openai.com/en/articles/12652064-age-prediction-in-chatgpt) — OpenAI Help Center, checked 18 October 2026 - [Managing parental controls in ChatGPT](https://help.openai.com/en/articles/12315553-managing-parental-controls-in-chatgpt) — OpenAI Help Center, checked 18 October 2026 - [Without Guardrails, Generative AI Can Harm Education](https://knowledge.wharton.upenn.edu/article/without-guardrails-generative-ai-can-harm-education/) — Knowledge at Wharton, University of Pennsylvania, 2024-08-27