OpenAI Says Teens Spend Less Than 15 Minutes a Day on ChatGPT

OpenAI has released its first report on how teenagers are using ChatGPT, offering a glimpse into the role the AI chatbot is beginning to play in the lives of younger users. According to the company, teens spend less than 15 minutes a day on ChatGPT on average, while fewer than 2% use the service for more than three consecutive hours.

The figures arrive at an important moment for OpenAI. ChatGPT has become a widely used tool for learning, writing, research and everyday questions, but the rapid adoption of conversational AI has also created difficult questions about how these systems should behave around younger users. OpenAI introduced a dedicated ChatGPT for Teens experience in August with additional safeguards, parental controls and learning-focused features.

OpenAI Says Teen ChatGPT Use Looks Different From Social Media

OpenAI is emphasizing the relatively short amount of time teenagers spend with ChatGPT compared with traditional social-media platforms. The company says the average teen spends under 15 minutes a day with the chatbot, and fewer than 2% spend more than three consecutive hours using it.

The company also highlighted what happens when its break reminders appear. In nearly half of teen conversations where a break reminder was shown, the user either stopped using ChatGPT or took a break within five minutes. OpenAI presents the figures as evidence that its efforts to encourage healthier patterns of AI use can influence user behavior.

There is an important distinction here, however. ChatGPT is not designed around the same continuous feed model as many social platforms. OpenAI argues that its product is primarily being used for learning, creating and solving problems rather than passive scrolling. That difference is part of the company's argument that time spent with ChatGPT should not automatically be interpreted in the same way as time spent on social media.

Learning Is Becoming a Major Part of Teen Use

OpenAI's report also points to education as one of the strongest use cases for its teen-focused experience. The company says teens with access to ChatGPT for Teens sent about 2.7 million more learning-related messages on average than comparable teens who had not yet received access.

The numbers around specific tools are also notable. OpenAI says nearly 1.2 million teens used Learning Visualizations in one week, while more than 180,000 used Study Mode for step-by-step learning support. The company says the most common conversation starters include asking for help developing ideas, receiving feedback on drafts and checking evidence.

That direction is significant because it shows how OpenAI wants ChatGPT to fit into education. Rather than positioning the chatbot simply as a machine that produces answers, the company is increasingly presenting it as a learning assistant that can help students understand difficult subjects, organize information and work through problems.

OpenAI is also adding features aimed at longer-term academic planning. Its latest announcement says a College Planner is coming to ChatGPT for Teens, initially focusing on U.S. students in grades 10–12 and helping them organize application requirements, deadlines and financial-aid steps.

Safety Concerns Have Not Disappeared

The relatively low average usage figure does not settle the larger debate around teenagers and AI. Safety researchers and child-focused technology organizations continue to question whether existing safeguards are strong enough for younger users.

Common Sense Media's Youth AI Safety Institute published a separate assessment on the same day as OpenAI's usage report and rated ChatGPT for Teens an “Unacceptable Risk.” The organization said it tested more than 4,000 prompts before and after the launch of the teen experience and argued that some of OpenAI's advertised protections did not consistently perform as expected.

Among the issues raised were parental notifications, break reminders and the way the chatbot handles conversations involving sensitive situations. Common Sense said some protections worked, including restrictions around explicit sexual role-play, but argued that other safeguards were less reliable than parents might expect.

OpenAI disputes the assessment. The company has said that much of Common Sense Media's testing may have taken place before parental-control features were fully activated, arguing that the organization's methodology does not accurately represent how the current teen safeguards operate.

OpenAI Is Trying to Change How ChatGPT Behaves Around Teens

The teen experience is more than a separate collection of educational features. OpenAI says accounts identified as belonging to users under 18 are automatically placed into an age-appropriate experience, with additional protections designed around sensitive content and healthy use.

The company says those protections include parental controls, Quiet Hours, break reminders, sensitive-image warnings and additional safeguards intended to reduce exposure to inappropriate or harmful material. It has also updated its model behavior guidelines so that ChatGPT should avoid encouraging emotional dependence or presenting itself as a human relationship for younger users.

That approach reflects a broader shift in the AI industry. As chatbots become more capable and conversational, the challenge is no longer simply preventing obviously inappropriate answers. Developers also have to consider how frequently people use these systems, how they interpret the chatbot's responses and whether AI becomes a substitute for appropriate real-world support.

The Bigger Question Is Not Just Screen Time

OpenAI's 15-minute figure is useful, but it only answers one part of the question.

A teenager could spend a short amount of time asking ChatGPT for help with a difficult homework problem and receive genuine educational value. Another user could spend the same amount of time seeking advice on a complicated personal issue. The duration is identical, but the potential impact is very different.

That is why the emerging debate around AI safety is becoming more complicated than traditional screen-time discussions. Researchers, parents and technology companies increasingly have to look at what young people are doing with AI, how the system responds and what safeguards exist around those interactions, rather than focusing only on the number of minutes spent inside an application.

OpenAI itself appears to recognize that distinction. Its latest teen-focused features emphasize learning, critical thinking, college planning and structured study, while the company continues to develop safeguards designed specifically for younger users.

A Critical Test for OpenAI's Teen Strategy

The next phase will be less about whether teenagers use ChatGPT and more about whether OpenAI can demonstrate that its safeguards work consistently at scale.

The company now has a large and growing population of younger users interacting with increasingly capable AI systems. That creates an unusual balancing act: restricting the technology too heavily could reduce its educational value, while insufficient safeguards could leave younger users exposed to risks that are still not fully understood.

OpenAI's latest data suggests that most teens are not spending hours continuously chatting with ChatGPT. But the criticism from independent safety researchers shows why usage statistics alone cannot settle the argument.

For OpenAI, the real test will be whether it can combine the educational potential of ChatGPT with protections that parents, educators and safety researchers can independently trust. As AI becomes a more familiar part of everyday education, that balance may become one of the company's most important long-term challenges.

Sources

• OpenAI — “Helping teens learn, plan, and shape the future of AI”
• OpenAI — “Introducing ChatGPT for Teens: Built for learning, backed by protections”
• Reuters — “OpenAI says teens use ChatGPT under 15 minutes a day as worries over risks grow”
• Common Sense Media / Youth AI Safety Institute — “ChatGPT for Teens Risk Assessment”
• Associated Press — “Some guardrails on ChatGPT for Teens don't work as promised, watchdog group says”
• TechCrunch — Coverage of ChatGPT for Teens safety testing