Industrial Revolution 1.0 drove extraordinary productivity gains in Britain. But for nearly a century, workers saw little benefit. Wages stagnated, jobs were displaced and living conditions worsened. It took decades – and major institutional and legislative changes – for prosperity to spread. The lesson is simple: productivity gains don’t automatically translate into shared prosperity. Without deliberate choices, gains are privatised while costs of transition are socialised.
A second lesson comes from the work of 2006 Economics Nobel winner Edmund Phelps: economies flourish when its benefits are widely diffused – when ordinary people can participate, experiment and build.
In much of the developed world, AI is being deployed within large, formal enterprises. Its economic logic is clear: improve efficiency, often by reducing labour. When productivity gains come from headcount reduction, public anxiety is a rational response. India’s structure is fundamentally different, and potentially advantageous.
For over a century, economic efficiency has come through consolidation. As another Nobel-winning economist Ronald Coase explained, firms expand when coordinating many small actors is costly. This is how Western economies evolved: towards large corporations with specialised, narrowly defined roles.
AI and digital networks are now reversing that logic. When coordination costs collapse – when software can match buyers and sellers, manage logistics, and enable transactions at scale – need for large, centralised firms weakens. Efficiency can emerge from networks of small actors instead.
India’s economy has this shape. With millions of MSEs and hybrid livelihoods, its informality has long been seen as a weakness. In the AI era, it may become a strength. Work in small enterprises is multi-tasked, adaptive and judgement-driven. Precisely the kind of work that’s harder to automate and easier to augment. Which makes India better positioned than many advanced economies to use AI to enhance human productivity, rather than replace it.
But this advantage is not automatic. The real risk for India is not mass unemployment. It’s missing a once-in-a-generation opportunity to raise productivity and incomes at scale. At the same time, India has a distinct vulnerability. Its formal sector may be relatively small, but it anchors middle-class aspirations. For three decades, the pathway to upward mobility has been clear: education – especially in STEM – leading to jobs in IT and related sectors. AI now directly targets many of these entry-level roles.
Even limited disruption could have outsized consequences. In a country with millions of young people struggling to find meaningful work, a weakening mobility ladder isn’t just an economic issue but a sociopolitical one. India’s AI trajectory will be shaped by two choices: how widely the technology is diffused, and whether it expands or constrains economic mobility.
If AI remains concentrated among large firms and an English-speaking elite, India risks becoming a more unequal, 2-speed economy, where growth doesn’t translate into broad-based prosperity. If adoption is slow, it risks missing the opportunity altogether. Avoiding these outcomes requires action in 3 areas:
Reinvent mobility ladder
India’s large employers – especially in IT services, financial services and manufacturing – are at the AI disruption frontlines. If AI is deployed primarily as a cost-cutting tool, entry-level opportunities will shrink, inequality will widen.
The alternative is to use AI to drive growth. Growing firms are better positioned to reskill and redeploy workers. But this requires sustained investment. Not just in technical training but also in capabilities such as problem-solving, judgement and adaptability. It also requires new approaches to workforce transitions: large-scale skilling systems, job-matching platforms and incentives for firms to invest in people, rather than simply shed labour.
Build AI as public infra
Largest gains for India will come from raising productivity in its least productive sectors – agriculture, small retail and micro-enterprises. This requires making AI widely accessible.
The country has already demonstrated how to do this through DPI like UPI. A similar approach for AI – open, interoperable platforms, low-cost models in non-English Indian languages, and population-scale onboarding – can expand access. When AI runs on public rails, its benefits scale differently. Instead of being concentrated within a few firms, they diffuse across millions – farmers improving yields, small businesses expanding reach, and individuals accessing capabilities once out of reach.
Enable mass entrepreneurship
The most transformative impact of AI will come not from using it but from building with it. AI lowers barriers to starting and scaling a business – enabling individuals to identify opportunities, create products and access markets with minimal capital.
India should lean into this by making it far easier to start and grow enterprises across its districts. In this model, job creation becomes decentralised, driven not by a few large employers but by millions of small ones. It is more resilient, more inclusive.
These three priorities must move together. Progress on one without the others will not be enough. India doesn’t need to lead the world in building the most advanced AI models. But it must lead in ensuring benefits of AI are widely shared. The real AI divide won’t be between countries that have the tech and those that don’t, but between societies that diffuse it broadly and those that allow it to concentrate.
Nilekani is co-founder-chairman, Infosys, and Venkatesan is former chairman, Microsoft India
AgenciesAgencies