Since the advent of the internet, technology’s biggest promise for education was democratisation. To what extent has this promise been fulfilled?
Fifteen or 20 years ago, access itself was the biggest challenge. A student living in a small town would have to leave home or spend a lot of money just to learn from a good teacher. The internet made quality education available to millions of students regardless of where they lived. So, from this perspective, the promise has been fulfilled around midway [50 per cent]. Today, a student in a village can attend the same class as someone in a metro, which is a remarkable shift. But access alone is no longer enough. Technology is democratising access, but learning is still largely one-size-fits-all.
What are some of the biggest problems of offline education or traditional edtech that AI edtech can effectively solve?
In a classroom, it’s impossible to know every moment where a student gets confused. Even online learning, while expanding access, often delivers the same content in the same way to everyone. While SMEs dedicated to doubt solving, extra mentorship, practice tests etc., are solving this problem, it is difficult to achieve the kind of scale we need. AI can help bridge that gap. It can understand where a student pauses, which mistakes keep repeating, how much time they’re spending on a topic, and intervene immediately. Right now, I don’t see AI replacing classrooms or teachers. I see it helping teachers give every student the kind of individual attention that time alone doesn’t always allow.
Can AI democratise education for nearly 250 million school-going children in India, who come from different regions, speak different languages and belong to diverse socio-economic strata?
For me, democratisation doesn’t mean giving everyone the same AI tool. It means ensuring every child, irrespective of where they come from, has an equal opportunity to learn well.
Our classrooms are incredibly diverse. Students learn in different languages, come from varied backgrounds, and have very different levels of exposure to technology. If we provide AI solutions that are built on the assumption that every learner has a premium to pay and have uninterrupted high-speed internet access, it will only benefit a small section of students.
The opportunity is to build AI that is affordable, understands Indian languages, works well even in low-resource environments and adapts to different learning contexts. That’s when technology truly becomes inclusive.
How do we ensure AI doesn’t widen educational inequality by disproportionately benefiting children who already have access to better schools, devices and the internet?
The answer lies in building products around the realities of our students. They should work across affordability, Indian languages, function well on affordable devices, use minimal data and remain accessible to families that may not have continuous connectivity.
If every child eventually has access to an AI tutor 24/7, how should schools redefine the role of classroom learning?
While AI can make learning more personalised, it can only replace the human relationships to some extent that shape a child’s education. This is especially true for the motivation and mentorship teachers provide and the relationships students build with each other.
What I see happening with AI is that classrooms would become even more interactive. Instead of spending most of the time delivering content, teachers can spend more time discussing ideas, solving problems, mentoring students and nurturing curiosity. If AI handles routine personalisation, teachers gain more time for what only teachers can do.
The real question every education company should ask is simple: Are students learning better because of what we’ve built?
Much of the AI conversation focuses on students. But isn’t India’s bigger bottleneck teacher capacity?
Whenever we talk about improving education, we should also ask how we’re supporting teachers. Teachers spend hours preparing lessons, evaluating answers, identifying learning gaps and answering repeated doubts. AI can be leveraged to reduce the burden of this repetitive work.
Imagine if a teacher walks into class already knowing which concepts most students struggled with the previous day, which questions caused confusion and who needs extra attention. That changes how effectively the teacher can use classroom time.
What risks do you see if India’s education ecosystem becomes dependent on foreign AI models?
AI learns from the data and models it is trained on. Education is deeply connected to language, culture, curriculum and local context. If we depend entirely on models built for different educational systems, we risk creating AI tools and solutions that don’t fully understand our students or reflect our classrooms. India has one of the world’s largest student populations and incredible linguistic diversity. We should actively build AI that understands our curricula, our languages and the way our students learn. It is necessary to ensure that the technology shaping the education of millions of children reflects their own context and realities.
The previous generation of edtech companies chased rapid growth, aggressive sales and high valuations, sometimes at the cost of educational outcomes. What are the biggest lessons AI edtech should learn from that period?
Education is built on trust, and trust is earned over years but can be lost very quickly. While technology, funding and growth are all important, they should always remain enablers, not the purpose. The real question every education company should ask is simple: are students learning better because of what we’ve built? If we focus on improving outcomes, students stay, parents trust us and growth follows naturally. If we focus only on growth, we eventually lose sight of why we started. That lesson remains just as relevant in the AI era.
Would you rather see AI integrated into government schools through public infrastructure or delivered by private start-ups competing in the market?
Government institutions have the scale to reach every child. Private companies often have the agility to innovate quickly. The best outcomes usually come when both work together with the student at the centre. Public systems can provide reach and continuity, while private innovation can continuously improve the quality of learning experiences. If our common goal is improving learning outcomes, a partnership is likely far more beneficial.
MORE FROM THIS ISSUE
If you were investing in an AI education start-up today, what learning outcomes would you insist on measuring before revenue growth?
The first question I’d ask is whether their support is actually helping students learn better.
Are they understanding concepts more deeply? Are they making fewer repeated mistakes? Are they staying engaged over time? Can they solve problems independently instead of depending on AI for answers? I’d also want to know whether the product works equally well for students across different languages, regions and backgrounds. Education is about changing learning outcomes, not just increasing screen time. If AI genuinely helps students learn better, sustainable growth will follow.
(This story appeared in Outlook magazine’s August 3 issue, 'The AI Divide', which focuses on how India's AI education ambitions are colliding with the reality of inadequate digital infrastructure, undertrained teachers and AI tools that are not built around Indian students' cultural context)
