In most Indian classrooms, the biggest equity gap is not who shows up, but what happens after they sit down. AI-powered personalized learning is quietly changing that equation: one adaptive question, one tailored hint, one “I finally get it!” at a time. In this piece, I want to explore how these systems, when thoughtfully designed and implemented, can act as a great equaliser in Indian schools, and what it takes to ensure they reduce, rather than reinforce, class disparities.

For decades, the default model in our schools has been simple: one teacher, one textbook, one pace. The top third of the class is often bored; the bottom third is overwhelmed; the middle muddles through. AI‑driven personalised learning offers a fundamentally different promise: many paths through the same curriculum, calibrated in real time to each learner’s current level.

At the heart of most adaptive systems is a loop:

Think of it as a patient, always‑awake teaching assistant that keeps asking, “Is this too easy? Too hard? Just right?” and reshaping the next step accordingly. Over time, this loop acts like a personalised learning path, filling gaps that may have been accumulating silently for years.

In India’s context large class sizes, wide learning diversity, limited individual attention this kind of granular responsiveness is not a luxury. It is a potential lifeline, especially for students in under‑resourced schools who may be the first in their families to complete formal education.

The idea that every child deserves the “just right” level of challenge is not new. What is new is our ability to deliver it at scale. Research on AI‑driven adaptive learning in Indian public schools shows this clearly: when instruction is tailored to each child’s actual level, engagement rises, learning gaps shrink, and overall proficiency improves.

A large‑scale analysis of AI‑enabled platforms in state school systems serving over 4 lakh students reports:

A separate review of adaptive programs deployed in Indian languages highlights that, in some states, government‑school students have improved faster than private‑school peers in basic skills between 2022 and 2024. ASER 2024, for instance, notes a sharper recovery in reading among government‑school children over this period, and commentators link part of this “catch‑up” to targeted remediation and the gradual spread of adaptive tools in local languages.

Why does this matter for class disparities?

In other words, the same algorithm that adjusts a question for a child in an elite urban school can also do so for a child in a rural government school if we design for equity, not just efficiency.

One of the most intriguing developments in India’s digital education story is not a shiny new app, but a feature hidden in plain sight on DIKSHA national platform: AskDIKSHA.

DIKSHA, India’s national digital infrastructure for school education already hosts curriculum‑aligned resources in dozens of Indian languages and links printed textbooks to digital content via QR codes. Over the last few years, it has begun integrating AI‑based features designed explicitly for personalised adaptive learning:

Government documents explicitly recognise these features as enablers of personalised adaptive learning on top of the national platform.

Imagine a Class 7 student in a small government school:

For her, the textbook is no longer a static object. It is a doorway into an interactive space where she can ask questions without fear of judgement, rewind explanations as many times as needed, and receive practice tailored to her pace.

Teachers I’ve spoken to in government schools that use DIKSHA describe scenes where a usually quiet student suddenly becomes animated: “Madam, I asked the book a question and it answered!” Their role shifts from being the sole explainer to becoming a guide, helping students formulate better queries, interpret responses, and connect them back to local experiences.

Of course, the magic is not in the AI alone. It is in the way the platform is woven into classroom practice: shared device routines, group tasks around questions, teacher dashboards that highlight which concepts are confusing many students at once. When implemented thoughtfully, AskDIKSHA becomes a co‑teacher; when used superficially, it risks becoming just another novelty.

If AskDIKSHA represents AI inside the textbook, initiatives like AI Samrat (and similar AI literacy projects) represent AI as a subject of learning in its own right.

AI Samrat is designed as a large-scale AI literacy effort for government and affordable private schools across India, with a curriculum framework and content library aimed at over 50 lakh students, teachers, and parents. Its mission is to ensure that children especially in non‑elite schools do not encounter AI only as a mysterious black box, but as something they can understand, question, and use responsibly.

Key features include:

By early 2026, AI Samrat‑style initiatives report reaching over 10 lakh learners across more than 7,200 schools in around 180 districts, with a mix of government and affordable private deployments. Many of these schools are exactly the ones that have historically been last in line for cutting‑edge content.

What does AI literacy have to do with class disparities?

In interviews, school leaders participating in such programs describe a subtle mindset shift: “Earlier, AI felt like something that happens in other countries, in big tech companies. Now our students are asking, ‘Can we build something for our village?’” That question, asked in a government school classroom, is perhaps the most radical equaliser of all.

To understand how AI can truly equalise, we need to look beyond features and into the grain of classroom life.

In one upper primary government school in a semi‑urban block, teachers have embraced AskDIKSHA as part of their regular timetable. The school has a modest ICT room and a handful of tablets allocated under digital initiatives.

A typical Class 6 math period might unfold like this:

Over a few months, the teacher notices a pattern: children who used to hide at the back are now the ones eager to share what “the book” told them. The AI’s neutral, patient responses seem to lower the emotional stakes of not knowing. The teacher uses the platform’s insights to identify common misconceptions and adjust her next lesson.

This is not a futuristic, one‑device‑per‑child scenario. It is messy, noisy, shared‑device learning. But precisely because of that, it offers a glimpse of how AI can support equity when embedded in a design that values collaboration, voice, and reflection.

In another district, a cluster of government schools has adopted an AI literacy program similar in spirit to AI Samrat. Rather than treating AI as a separate “tech” subject for older students, they introduce it gently from upper primary level.

In a Class 7 social science lesson:

In language classes, students are asked to write short stories from the perspective of an algorithm: “I am the app that decides which videos you see…” These are then recorded as audio, sometimes with simple AI‑generated images.

Teachers report that this approach does more than tick an AI‑literacy box. It amplifies foundational skills reading, writing, critical thinking while making abstract AI concepts concrete and emotionally resonant.

For students from disadvantaged backgrounds, who may be navigating powerful algorithmic systems without adult guidance, this kind of critical literacy is a shield. It helps them ask, “Who built this? Who benefits? Where is my data going?” questions that are central to equal participation in a digital society.

Technology by itself never guarantees equity. It can just as easily widen gaps if deployed without attention to context, power, and pedagogy. The difference between AI as a great equaliser and AI as a great divider often comes down to design.

From the cases, conversations, and research, I see several design principles that matter:

In many low‑resource schools, students share devices, connectivity is patchy, and teachers juggle multiple grades. Successful AI‑enabled programs do not assume one‑to‑one access or uninterrupted internet. They:

This learner‑first design makes it more likely that the poorest child in the class will actually benefit, not just the most digitally privileged.

The most powerful AI‑supported lessons I have observed are suffused with joy: students arguing with the system’s answer, laughing when it misinterprets a question, challenging each other to “beat” a practice level.

Design elements that foster this include:

When joy and agency are central, AI stops feeling like a surveillance tool and starts feeling like a playground for thinking.

Every successful deployment I have encountered treats AI as a tool for teachers, not a replacement of teachers. This shows up in:

When teachers feel empowered, they use AI to differentiate homework, form flexible groups, and target remediation. When they feel sidelined, they either resist the tool or use it minimally, often reinforcing existing hierarchies.

Given India’s linguistic diversity and the persistent digital divide, AI systems that support multiple languages and modalities (text, audio, video) are inherently more equalising.

This matters because class disparities often map onto language and literacy disparities. A tool that understands a child’s own language is already doing a small act of justice.

Whenever we talk about AI in public education, especially in a country as large as India, one question looms: can we afford it? The more provocative counter‑question is: can we afford not to do it?

Analyses of personalised adaptive learning interventions in India, including AI‑driven platforms in government schools, consistently rank them among the more cost‑effective education reforms. This is because:

One evaluation of a state‑wide adaptive initiative noted average learning improvements of over 40% in certain subjects as students progressed from foundational concepts to grade‑level proficiency, at a per‑student cost significantly lower than adding extra teaching hours.

Of course, these numbers mask substantial upfront costs: content alignment, platform development, teacher training, and maintenance. But those are largely one‑time or periodic investments.

The alternative of continuing with a one‑pace, one‑path system is not costless. It manifests as:

From a public finance perspective, investing in AI‑enabled personalisation that reduces learning poverty and narrows class gaps can yield long‑term returns in the form of a more skilled workforce, lower remedial costs in higher education, and reduced social inequality.

The research is clear: AI is not a magic bullet. Its cost‑effectiveness hinges on:

When these conditions are met, AI‑driven personalised learning is not just affordable; it is strategically compelling.

No article in this series feels complete without listening to the voices of those living this transformation: principals, teachers, students, and parents. Their reflections keep the conversation grounded.

A principal of a government secondary school in a district piloting adaptive tools framed it this way: “We once thought of smart classrooms as a fancy room with a projector. Now I think of ‘smartness’ as an invisible network called AskDIKSHA, our adaptive practice, the AI literacy sessions. It is like electricity: you don’t talk about it once it works; you just design your school around it.”

He emphasised that the real shift came when the school stopped treating AI as a separate project and started weaving it into the daily timetable: “Every child, at least three times a week, should touch a system that responds to them individually. That’s my rule now.”

A language teacher in a rural upper primary school described her journey with an AI literacy module: “At first I was scared, I thought I needed to know more than the children. But the training said: think of yourself as a designer. Use the stories, the prompts, the tools to create experiences. That changed everything.”

She began using simple AI‑supported exercises: students dictated short essays into a device, then edited the transcriptions, discussing what the AI misunderstood and why. “We turned errors into puzzles,” she laughed. “For once, the machine was wrong and the children were right.”

In every school I visited that had embraced AI‑driven tools meaningfully, I met at least one student who said some version of: “I didn’t know this thing I was doing had a name.”

A Class 8 boy in a government school, who regularly tinkered with mobile settings and helped neighbours install apps, discovered the AI literacy club: “When they explained what algorithms are, I thought, ‘Oh, so that’s what I have been feeling.’ I always wondered why my videos are different from my cousin’s. Now I can talk about it.”

For him, AI stopped being an invisible force and became a subject of inquiry. That shift from being acted upon to being able to analyse and act is at the heart of AI as an equaliser.

Earlier articles in this series have explored curriculum lag, design‑led learning, and the role of storytelling and choice in making classrooms more alive. This piece sits at that intersection, but with a new lens: AI as a design material in the hands of educators.

The patterns repeat:

But when we start with design principles joy, choice, authenticity, reflection and weave AI into them, something different emerges:

In that sense, AI is not the hero of the story; it is a powerful supporting character. The protagonists remain the same: thoughtful adults, curious children, and a public system willing to reimagine itself.

If we are serious about AI as a great equaliser in Indian schools, a few commitments feel non‑negotiable.

If I had to choose one image for the future we are working towards, it would not be a pyramid with a few at the top and many at the bottom. It would be a circle: a classroom where each child’s path around the circle is different in detail but equal in dignity.

AI alone cannot draw that circle. But it can help us see, with unprecedented clarity, where each child stands, what they need next, and how far they have come. It can free teachers from some of the mechanical burdens so they can invest more in the human work of teaching. It can carry high‑quality explanations into remote corners where no private tutor will ever go.