Our children are learning with AI, safety must grow with them

A Class 12 student struggling with a science chapter can now ask AI to explain it in simpler words. A teenager preparing for an entrance exam can use it to understand where they went wrong in a practice question. A student in a smaller town can ask questions in their own language, at their own pace, without feeling embarrassed.

This is no longer a distant possibility. For many young people, AI is already becoming part of how they learn, revise, create and explore new skills. Global OpenAI data suggests that nine in ten teens use ChatGPT for learning and skill-building in a single week. In India, this trend is even more pronounced: users are twice as likely to turn to ChatGPT for learning and education compared to the global median.

For a country as large and diverse as India, this opens up enormous possibilities. Access to good teachers, quality learning material, language support and individual attention still varies widely. AI can help bridge some of these gaps by making learning more personal, more accessible and more interactive.

But with this opportunity comes an important responsibility: how do we make sure AI is safe for young people?

The question is not only whether teenagers can access AI. The real question is how that access shapes the way they learn, think and make decisions. How do we protect them from harmful or age-inappropriate content? How do we prevent unhealthy dependence? And how do we make sure AI helps children think better, instead of thinking on their behalf?

Teen safety, therefore, cannot only be about blocking harmful content. It must also be about helping young people use AI safely, critically and independently.

Safety must reflect how Indian families use technology
Many teenagers in India do not have a device of their own. Phones and computers are often shared between family members, and access can be even more limited for girls.

This matters because the age registered on an account may not always match the person using it. Age-estimation technology can add another layer of protection. ChatGPT uses behavioural and account-level signals to help identify under-18 users and automatically apply additional safeguards.

Keeping teens safe is a shared responsibility
AI companies cannot solve for teen safety alone. Parents, educators, and trusted adults have an important role in helping teens build agency: understanding that AI can be wrong, recognising unhealthy patterns of use, and knowing when to step away.

But many adults are themselves still learning how to use AI responsibly, even as they are expected to guide teenagers growing up with it. Teen safety, therefore, is not just a product question, but a societal one. Good design can help bridge this gap, giving families greater oversight without taking away teenagers’ independence.

Safety is also about learning
Safety is also about what AI encourages young people to do. Does it help them develop judgement, curiosity and independent thinking? If a student asks for the answer to a maths problem, for example, AI can simply provide it, or it can help them work through the problem themselves.

For younger users, we should prioritise the second approach. While adults may use AI to get things done faster, for teens, the focus should also be on learning. AI should encourage them to ask better questions, work through problems and arrive at answers themselves. The aim is to strengthen learning and independence, rather than make it easier to avoid the work of learning.

These principles should also shape how AI products work in practice. For example, parents can be given tools to shape how their teens use AI, including settings for study hours. During this time, AI could focus on guiding teens through problems step by step, rather than simply giving them answers.

India can help shape safer AI
India has an opportunity not just to adopt AI for young people, but to help shape what responsible AI for teens should look like.

With one of the world’s largest young populations, many languages, and very different ways of accessing technology, India’s experience can offer lessons for how AI is designed for young people around the world.

If AI is going to become part of how our children learn and grow, keeping them safe has to grow with it.
(The author is the Head of Strategy & Global Affairs, OpenAI India)