NEP at six: Is India's education system ready for the AI age?

This article is authored by Khushboo Awasthi and Mithun Mathew.

In Shirala, a taluka in Maharashtra's Sangli district, a group of government school teachers spent last summer learning something none of them had been trained for: how to help children investigate the problems around them, with Artificial Intelligence (AI) as one of the tools in their hands. Most of them had never written a line of code. Several were teaching in classrooms where the internet drops out by the afternoon.

Education

What has happened since is worth pausing on. Students in these schools are no longer asking what AI is. They are asking what problems around them are worth solving - the borewell that runs dry in March, the crops their parents grow, the neighbour who cannot read a government form, and how AI might become one of several tools to help. The district is now preparing a student challenge built around exactly those questions, and the programme is expanding to over 40 more schools.

Nothing about that classroom looks futuristic. There is no robotics lab, no gleaming hardware. What has changed is the posture of the room. Children have moved from waiting for answers to designing questions.

That, and not the equipment, is what AI readiness actually looks like.

NEP 2020 was remarkably prescient. It asked for critical thinking over rote learning, competencies over content, experiential and interdisciplinary classrooms, and technology woven into teaching. It gave India the right compass.

But technology has not waited for our implementation timelines. AI has grown faster than any reform cycle can move, and the honest question on this anniversary is not whether NEP got the vision right. It did. The question is whether our implementation plan and pace reflects the reality that has arrived since.

NITI Aayog's May 2026 report on school education is a useful mirror here. It documents a decade of real gains: expanded digital infrastructure, more computers, internet connectivity and smart classrooms across states, better inclusion, stronger foundations. It also asks the system to integrate AI for pedagogical innovation. And it is unsparing about what remains: digital access that is still deeply unequal between urban and rural schools, and children who complete years of schooling without mastering reading and arithmetic.

But statistics are only half the story. In the schools we walk through with the Shikshagraha movement, we see the other half — teachers making the most of limited resources, and the gap between having technology in classrooms and using it meaningfully. The infrastructure is being built. The real question is not what we build, but how we use it, and who gets to decide.

AI promises learning that is far more personalised than a crowded classroom can usually offer. Every child could have a tutor that understands their pace, their language, their interests and their gaps. It is a real possibility, and we should want it for every government school child, not only the privileged few.

But personalisation is not, by itself, transformation. A model can adapt the pace of a lesson; it cannot decide what is worth learning, or build the classroom culture in which a child is willing to be wrong in front of others. What turns AI into better learning is what has always turned resources into learning - sound curriculum, thoughtful pedagogy, and a teacher who knows when to step in and when to step back. AI amplifies good teaching. It does not substitute for it.

But history offers a warning. Social media promised connection and too often delivered isolation. Every major technological revolution has increased individual capability; very few have strengthened our collective capability. AI may become the most personalised technology humanity has ever built. If AI-powered learning slowly becomes solitary learning- each child alone with an intelligent machine- we may perfect instruction while weakening one of education’s most important purposes.

Because schools are not only where children learn mathematics or science. They are among the last public institutions where children from different castes, faiths and incomes learn to disagree respectfully, cooperate across differences, build trust, negotiate conflict and imagine a shared future. In an age where intelligence is becoming ever more individualised, schools become one of the few places that cultivate what AI cannot personalise: our capacity to belong to one another. This is not a soft benefit of schooling. It is democracy’s apprenticeship.

If NEP 2020 was the first horizon - the vision - the next phase must be the second horizon: specific, practical, implementable. Five things would change the trajectory.

Make AI literacy universal from middle school, not an elective for well-resourced schools and teach it the way NEP intended learning to work, across disciplines rather than in a silo. A class investigating why the village borewell runs dry in March is doing science, mathematics, geography, civics and language at once, with AI as one instrument among many. Prompt design and collaborative problem-solving belong inside that work, not in a separate computer period.

Redesign assessments, board examinations included, to reward reasoning, synthesis and application rather than recall; nothing else will matter if the exam still pays for reproduction.

Teach ethics from the first lesson, so children learn to use AI wisely as they learn to use it at all.

Support teachers as learning architects, not content experts; the teacher confident enough to say “let us find out together” now matters more than the one who knows every answer.

And build the shared scaffolding for equity - district AI innovation hubs, community maker spaces like MakerGhat, pooled infrastructure, so a child’s access to the future does not depend on her PIN code.

Progress should be measured not by devices procured or smart classrooms installed, but by teacher confidence, student problem-solving, equitable participation and the responsible use of AI. Infrastructure enables change; pedagogy determines whether change actually happens.

I want to return to Shirala, because that classroom tells us something hopeful. AI is not a threat to be managed. It is the most significant invitation in a century for education to reconsider its own purpose. AI makes people more capable, but capability is not agency. It can answer questions; it cannot decide which questions are worth asking. It can produce solutions; it cannot choose which problems deserve solving. Every generation inherits some uncertainty; this one inherits it at machine speed. That places a responsibility on education beyond employability: to help young people believe they can shape the future, not merely adapt to it.

So the defining question is no longer whether our children can outperform AI. They will not, and they should not have to. It is whether we use this moment to raise merely more capable individuals or more capable and more responsible citizens, able to work alongside these machines while strengthening what the machines cannot replicate: judgement, creativity, empathy, ethics, resilience and purpose. For decades, our system has produced knowledgeable people; the task now is to nurture responsible ones - agents of change.

That answer will not be written by technology. It will be written in our classrooms.

NEP 2020 gave us the foundation. The second horizon must ensure that every child in every public school leaves not merely employable in an AI economy, but equipped to shape society with wisdom, responsibility and agency.

The future is arriving faster than our policy can keep up. Our children are already running to meet it. The system owes them the responsibility of keeping up.

This article is authored by Khushboo Awasthi, co-founder, ShikshaLokam & Mantra4Change; chief curator, Shikshagraha & Mithun Mathew, co-founder and director partnerships, MakerGhat.