AI illiteracy is the new digital divide: Why it’s a risk for India’s future workforce

By Rajiv Nandwani, Senior Director, School of Computer Science, UPES

India’s digital success story is a global benchmark, defined by affordable data, mass smartphone adoption, and connectivity programs that reached the last mile. Hundreds of millions came online in under a decade, solving the challenge of access. However, as we enter a new era of automation, a more silent and dangerous divide is emerging.

It is no longer a question of who has a signal; it is a question of who has the cognitive agency to interrogate the machine

The paradox at the center of India’s AI story
India’s headline numbers are striking enough to invite complacency. The country ranked second globally for AI project contributions on GitHub in 2024, accounting for 19.9% of all AI-related repositories worldwide. It was placed in the top three on Stanford University’s Global AI Vibrancy Tool. NASSCOM projects the domestic AI talent pool growing from 600,000–650,000 in 2022 to over 1.25 million by 2027, driven by an AI market expanding at 25–35% compound annual growth. India led every country in Generative AI course enrolments on Coursera in 2024 is 1.3 million learners.

And yet, against that same Coursera data, India ranked 89th globally in actual AI skills proficiency. The country is producing enthusiastic consumers of AI education and a far smaller cohort of people who can actually apply, audit, or critically evaluate what AI systems produce. That gap — between enrolment and comprehension, between exposure and fluency — is the defining challenge of the next decade of India’s digital story.

The Illusion of Tech-Fluency
Consider what is already happening inside institutions. Adaptive learning platforms are quietly deciding which content a student sees next. Automated screeners are filtering resumes before a human eye reaches them.

Recommendation engines are shaping the opportunities young professionals even know to apply for. None of this is hypothetical. It is the infrastructure of student life right now. And most of the students inside it have no real sense of how these systems make their decisions, what data they are trained on, or where they routinely go wrong. The algorithm is in the room. Nobody introduced it.

For years, the conversation around educational equity in India centred on access. Could a student in a tier-3 city get online? Did the school have computers? Those questions mattered, and progress was made. But the frontier has moved. The divide that is opening now is not between those who are connected and those who are not. It is between those who can interrogate an AI output and those who simply accept it. Between those who notice when a model is confidently wrong and those who cite it in an assignment. That second group is, at the moment, very large.

Walk through any Indian university campus and the usage numbers are striking. Generative AI tools have become standard kit for assignments, research summaries, and interview prep. Students are not avoiding this technology.

They are deep inside it. What is missing is any structured ability to evaluate what it gives them. A student who cannot spot a hallucination, who does not know that a model trained on skewed data will reproduce that skew, who has never been taught to ask why a system returned this result and not another, is not being served by the technology. They are being managed by it. Scaled across millions of graduates, that is a serious national concern.

What employers are actually discovering

The signals from industry have shifted from cautionary to urgent. Across banking, financial services, healthcare, logistics, and manufacturing, AI integration is no longer a planned future investment. It is current operations. What these sectors need are not graduates who can prompt a chatbot effectively. They need people who can sense-check an AI recommendation against domain knowledge, flag an anomaly that the model missed, and make a confident professional judgement about when the model should be overruled entirely.

That requires a different kind of preparation than most Indian institutions are currently providing. According to India Skills Report 2025, intent to hire freshers has fallen — just 14% of new hires across industries are expected to be freshers in 2026, down from 18.8% in 2024. The Goldman Sachs analysis is starker still: since late 2022, unemployment among 20- to 30-year-olds in the technology sector has climbed by nearly three percentage points — over four times the rise in the overall jobless rate. AI is beginning to displace entry-level white-collar work.

Why the fix is structural

The well-intentioned response to this gap has typically been to add. An optional AI ethics module here. A one-semester elective there. An extracurricular hackathon for students who already have an interest. None of this changes much at the system level, because it reaches the students who were already likely to develop these skills. It leaves the much larger population — students in programmes that do not self-select for technology curiosity — exactly where they are.

What the evidence now strongly suggests is that AI literacy must be treated as foundational — the way numeracy and written communication have always been treated as foundational. Not as a specialisation, but as a baseline competency expected across every discipline. Within five years, an engineering, business, or humanities graduate who cannot evaluate an AI output, identify bias in a dataset, or make an informed decision about when not to trust a model’s recommendation will face a serious professional limitation.

India’s genuine advantages

India’s advantages in this space are real and should not be understated. The country has the world’s largest digitally skilled talent base, a policy framework in NEP 2020 that explicitly prioritises future-ready skills, and a government whose India AI Mission is funding infrastructure and institutional capacity at scale. India ranks among the top three countries globally for AI vibrancy. The talent base is not the problem. The preparation model is.

The country also has demographic leverage that no other major economy can replicate. More than 50% of India’s population is under 30. The workforce entering professional life over the next decade will define India’s AI trajectory for a generation. If those graduates arrive AI-fluent and can think alongside intelligent systems, push back on them when required, and make confident professional judgements about their outputs, then India has the capacity to lead the global AI economy, not merely service it.

If they arrive as passive users, producing the same outputs that a well-prompted chatbot produces and no more, the opportunity closes. Quickly, and at scale.

The mandate for institutions

The institutions that will define India’s AI-literate professional class are not the ones deliberating about whether transformation is necessary. They are the ones already restructuring curricula, investing in AI-focused lab infrastructure, building faculty capability, and ensuring that every graduate — regardless of discipline — leaves with a working understanding of how AI systems make decisions and where those decisions should be questioned.

This is not primarily a technology challenge. It is a pedagogical one. The tools exist. The talent exists. The policy intent exists. What is needed is institutional will to treat AI literacy as the new foundational literacy it already is — and to move at the pace the economy is already moving, rather than the pace at which academic reform has historically proceeded.

The 2030 workforce is already in classrooms. The question is what kind of classrooms they are sitting in.