Column | AI Won't Take Most Of Our Jobs, But That May Not Be Reassuring

India is creating more of the latter kind of work. A recent study by Ashoka University finds that India’s non-farm sectors added around six crore workers between 2018 and 2025. However, they were mainly in elementary occupations such as construction labour, personal services, and retail and shop workers. Some high-skilled professional jobs grew, too, but the growth in the middle rung—jobs that could help workers move from basic work to professional careers—remained weak.

This may look like a revival of blue-collar jobs. But more work does not automatically mean better work. While real wages in high-skilled professional jobs rose by more than 20 per cent over this period, wage growth was more modest in low-skilled ones (8–10 per cent). In other words, the Indian labour market created jobs that still required little formal training, paid modestly and offered no clear upward career progression. So, while it created work, it was not what a younger, aspirational and more educated Indian workforce had hoped for.

Now add AI to this equation. A recent study by Vidya Mahambare, Vivek Jadhav and Vishnu Venugopalan estimates that only about one in four Indians work is in an occupation highly exposed to AI; only one in 10 face pressure from AI replacing their jobs. Most still work in jobs that are difficult to automate from a computer screen.

This gives India some protection from an immediate AI-led disruption. But this insulation can quickly become a low-income trap. Much of the debate asks whether AI will create more jobs or not. India should ask a more useful question: can AI help the economy produce more with its current workforce and can these workers receive a fair share of that gain?

Imagine a technician using AI to find a fault faster and service more machines. The firm earns more because each worker can do more. But if the technician remains an easily replaceable contractor, who is informally employed, most of that gain may go to the firm. Put differently, a job that is safe from automation can still trap workers in low pay and insecurity.

Reskilling Workers

How we reskill workers will decide which path India takes. Employers today say they cannot find job-ready workers, while jobseekers struggle to turn certificates into jobs. A 2025 Comptroller and Auditor General of India audit of the PM Kaushal Vikas Yojana, India’s flagship skills programme, found that only 40 per cent of certified candidates could be placed in jobs. The audit flagged no clear long-term strategy, weak quality controls, and training that did not match the sectoral and regional needs as key reasons for non-placement.

So, India does not simply need workers with better skills, it needs skills that employers can trust. AI will only make this problem harder. Job tasks are changing rapidly. Workers will have to reskill throughout their careers, and importantly, be able to credibly signal their skills to prospective employers. Another mass certification drive will not be enough.

One answer may be better apprenticeship programmes where workers can learn on-the-job. This creates a familiar problem: while one firm bears the cost to train a worker, another can simply poach that trained worker. A recent review of stakeholders by the NITI Aayog finds exactly this: firms, especially smaller ones, worry about supervision and compliance costs, low returns to training and the poaching of good apprentices. In other words, left to themselves, firms will train fewer people than the economy needs.

The other solution could be rethinking the entire paradigm of (re)skilling—by separating skill acquisition from their credible signalling to potential employers. AI can make the former more accessible, cheaper and adaptable. It can explain a circuit diagram, simulate a repair task and personalise each lesson, even in vernacular languages!

But learning something is not the same as proving you can do it. A chatbot cannot tell a client whether an electrician can wire a building safely. If the market is flooded with workers carrying certificates of uncertain quality, it may further depress wages and increase employers’ mistrust in them. This is where policy can matter. Governments need not deliver every course. Rather, they can play the role of making workers’ skills credible to employers. Workers should be able to carry a trusted record of their skills from one firm, job and city to another, and the government can create this infrastructure as opposed to focusing on the number of certificates being issued.

With fiscal budgets tightening, it may appear that we are at a fork in the road. On the one hand, policymakers face the task of creating enough work for millions of people today, while also investing in scientists, engineers and firms of tomorrow that push the technological frontier. While appearing to be a tradeoff, they are instead reinforcing requirements.

The electric-vehicle industry can create jobs only when mechanics can become diagnostic technicians. Sustainable sources of energy can be scaled only when manual installers become energy specialists. Modern factories need operators who can work with intelligent machines on complex tasks, not workers doing repetitive tasks on an assembly line.

That missing middle of technicians, supervisors and growing small firms is the bridge between mass employment and innovating at the frontier. Without new technology, millions may remain stuck in low-productivity work. Without workers who can service that technology, its gains will remain concentrated in a handful of firms, individuals and cities.

India should not aspire to become the world’s reserve army for jobs that machines cannot automate (yet). It should use this unique opportunity of a longer runway to raise the productivity and pay of work that still need humans. AI’s most valuable gift to India’s labour market is time. Squandering it would turn blue-collar resilience into another low-income trap.

(Views expressed are personal)

Gaurav Chiplunkar is Assistant Professor in the Global Economies and Markets group at the Darden School of Business, University of Virginia, US