AI in India: Is it a prosperity engine or mobility killer?

Every major technological revolution has followed a similar arc. First comes the breakthrough. Then the disruption. Then, if institutions evolve wisely, shared prosperity. If they do not, inequality hardens and social stress rises. Artificial intelligence will be no different.

Technology creates adaptive crises before shared prosperity: The first Industrial Revolution made Britain rich, but not quickly, and not for everyone. Early factories were brutal. Wages stagnated for decades even as output soared. It took unionization, social safety nets and mass public education before productivity gains translated into broadly rising incomes.

The lesson is clear: technology does not automatically produce prosperity. Institutions determine who benefits.

If AI is deployed primarily to replace human labour and compress costs, productivity may rise while inequality widens. Erik Brynjolfsson calls this the “Turing Trap”—using AI to automate what humans already do rather than augmenting what humans could do better.

AI is not primarily a technology challenge. It is an adaptation challenge.

Diffusion determines destiny: India has lived through two digital revolutions, and the contrast is instructive. In the 1990s and early 2000s, India rode the personal computer (PC) and internet wave. We built global IT champions like Infosys and TCS. A world-class export sector emerged. But PC penetration remained narrow. Broadband access was limited. The gains accrued largely to an urban, English-speaking elite.

That was narrow diffusion. After 2011, mobile internet, cheap data and digital public infrastructure changed the equation. UPI scaled to hundreds of millions of users. Digital payments reached street vendors and small merchants. Inclusion deepened. The digital economy’s share of GDP surged.

The PC amplified the skilled elite. The mobile-plus-UPI stack amplified the masses.

AI now stands at the same fork in the road. If it follows the PC path, it will enrich the top 10%. If it follows the digital public infrastructure path, it will raise productivity across the nation. The critical question is not whether India builds frontier models. It is whether AI capability reaches every farmer, health worker, MSME and student.

Broad diffusion drives mass flourishing. Narrow diffusion concentrates gains.

Technology is amoral—it amplifies society: Technology does not determine outcomes; it magnifies underlying institutions and values.

Radio in the 1930s strengthened democratic trust in the United States under Franklin D. Roosevelt. In Nazi Germany, it amplified propaganda and hate. Computerization in the United States coincided with rising wage inequality. In Japan, lifetime employment norms and retraining cushioned disruption.Same technology. Different institutions. Different outcomes.

AI will amplify India as it is. We are entrepreneurial and digitally ambitious. We are also deeply unequal, with a median income near ₹10,000–12,000 per month and 85% of employment informal. Our challenge is not high wages being displaced by automation. It is low productivity and stalled mobility.

In the West, AI is debated as a threat to high-paying knowledge jobs. In India, the deeper risk is something subtler: the removal of ladders.

For three decades, IT services and back-office roles provided upward mobility for first-generation graduates. Generative AI directly targets routine coding, documentation and processing tasks. If entry-level roles shrink before reskilling expands, mobility narrows. Not mass unemployment—but quiet aspiration collapse.

That would be destabilizing in a young nation.

Adoption always looks slow—until it isn’t: Electricity was invented in the late 19th century but transformed productivity decades later. Factories had to be redesigned. Management practices had to change. Organizational models had to evolve.

Personal computers entered enterprises in the late 1980s as stand-alone productivity tools. Only after networking, ERP systems and process reengineering did productivity surge in the late 1990s. Economist Robert Solow famously quipped that the computer age was visible everywhere except in productivity statistics—until it suddenly was.

Today, many companies are stuck in AI pilots that do not scale. We are likely in the “trough of disillusionment.” It is tempting to become either despondent (“AI is overhyped”) or sanguine (“This will take decades”).

Both are mistakes. Firms are still treating AI as a powerful tool, not an organizational redesign challenge. When workflows, incentives and management models are rebuilt around AI, adoption will not be linear. It will follow an S-curve—slow, then sudden.

This quiet period is an adaptation window. If we use it well, we can shape the outcome.

India’s choice—defending the frontier or raising the floor: For developed economies, the AI debate is about protecting existing prosperity. For India, it is about creating prosperity before our demographic window closes.

We need AI in agriculture to raise yields and incomes. AI embedded in logistics for small traders. Vernacular, voice-first tools for micro-entrepreneurs. AI-enabled public services that reduce leakage and friction. Productivity gains must reach the informal majority.

The central risk for India is not automation shock. It is mobility collapse combined with narrow diffusion. We can imagine three paths.

In a ‘dual-speed India,’ elite firms adopt AI aggressively while MSMEs and informal workers lag. GDP grows, but inequality widens and aspirations stall.

In a ‘stagnation trap,’ AI remains confined to pilots and prestige projects. Growth slows quietly. No crisis, just drift.

Or we can pursue ‘Viksit Bharat’: broad diffusion, mass AI fluency, MSME adoption, and guardrails that protect trust. Productivity rises across the base. Growth becomes inclusive and self-reinforcing.

Here’s what must be done. For CEOs, AI must move beyond cost-cutting. The real question is where value migrates. Organizations must be redesigned, not merely optimized. Augmentation should precede automation. Reskilling must be real, not symbolic.

For the government, ensure universal access, vernacular capability, MSME adoption and guardrails against misuse. Sovereignty is not about owning every chip. It is about ensuring 1.4 billion people can productively use AI.

History warns us that powerful technologies create adaptive crises before shared prosperity. But it also shows that societies that diffuse capability broadly and redesign institutions wisely can bend technology toward flourishing.

AI will not determine India’s future. Our institutional choices will.

The author is a chair of the Global Energy Alliance for People and Planet and a former chairman of Microsoft India.