India Added 104 Million Jobs Since 2019, Mostly Farming and Low-Skill Work. AI Threatens What's Left In The Middle

India's government can point to a real number: employment among working-age Indians rose from 47% in 2018-19 to 56.2% in 2025, adding roughly 104 million workers, according to Economic Times contributors Kanika Mahajan, Anisha Sharma and Mansi Wadhwa. That sounds like a jobs boom.

It isn't one, at least not the kind India needs.

More than 40 million of those new workers went into agriculture, almost all of them self-employed, the Economic Times piece reports. That's backwards. Countries that are developing normally move workers OUT of low-productivity farming, not into it. Services added about 30 million jobs, but the authors say that mix includes both good formal work and low-paid informal gigs.

The jobs that didn't show up are the ones in the middle: clerks, technicians, installers, repair techs, production-support roles. Those are the jobs that historically let someone with a basic education climb into stable, decent-paying work. According to the Economic Times analysis, wage growth over the past seven years was strongest at the bottom of the skill ladder, moderate at the top, and weakest in the middle.

Mining and construction labor grew by about 14.8 million workers. Cleaners, helpers and food-prep assistants grew by 47 lakh (roughly 4.7 million), per the same report. Meanwhile real wages for technicians and associate professionals, including machine operators and lab assistants, fell about 20%, even as top professionals saw real median wage growth of 22.6%.

Education isn't paying off like it used to, either. Compared to workers who never finished 8th grade, the wage premium for finishing 12th grade dropped 9 percentage points between 2018-19 and 2025, the Economic Times contributors found. That's the backdrop for the student protests over exam leaks and scarce jobs that have been rattling parts of India.

Goldman Sachs says AI won't be the main event, but it will hit specific jobs hard

Goldman Sachs Chief India Economist Santanu Sengupta told Bloomberg Television, in comments reported by both Business Standard and Business Today on August 14, that India is less exposed to AI-driven job losses than many other countries. His reasoning: construction and retail trade together make up about 40% of India's workforce, and those are physical, mechanical jobs that AI doesn't touch much yet.

The risk, Sengupta said, sits mostly in services: postal and telecom work, and IT services, especially call centers. Finance, healthcare, education and business services, by contrast, could see AI boost productivity rather than cut jobs.

Goldman's own modeling, cited by Business Standard, estimates that sequencing AI adoption correctly could add 0.4 percentage points to India's overall productivity over a 10-year period. Sengupta said if the rollout is gradual, "the productivity benefits will outweigh the potential job losses" over a five-year window.

India's government isn't dismissing the risk either. The Economic Survey 2025-2026 found that roughly 26% of Indian employment is exposed to AI, though the survey notes exposure can mean displacement OR productivity gains, depending on the job, according to NDTV Profit. AI-related job postings in South Asia more than doubled between January 2023 and March 2025, NDTV Profit reported, citing the government survey, with demand for AI skills growing faster than demand for non-AI roles.

The disconnect: physical labor insulation doesn't fix the middle-skill collapse

The jobs AI threatens most, per Sengupta, are exactly the kind of mid-skill service and IT roles that the Economic Times data says are already the missing piece in India's labor market. If AI eats further into call centers and routine IT services, it's not creating a new problem. It's deepening an existing one—a labor market that already skipped the middle rung of the ladder.

A U.S. parallel worth noting

Fox News reported on a similar clash playing out in American politics, where Sen. Bernie Sanders, I-Vt., has pushed for government intervention on AI job losses, including a virtual press call from Reps. Greg Casar, Valerie Foushee and Sara Jacobs proposing an AI tax to fund an employment program. Sanders said in April, "How the hell do you go forward and throw millions of people out on the street without planning what's going to happen?"

But Fox News noted the U.S. labor market hasn't followed that script: total non-farm payrolls have grown every year since AI chatbots launched in late 2022, and unemployment has stayed below 4.5%. Richard Stern of Advancing American Freedom told Fox the debate reflects "the difference between free market and central planning."

Breitbart's John Carney offered a blunter, industry-specific warning on coding jobs specifically, arguing the U.S. "overproduced" computer science graduates and that AI, offshoring to India and Malaysia, and H-1B visa hiring are combining to push unemployment up among young coders. Carney said displaced coders may have to "learn to weld" instead, echoing the old "learn to code" advice given to displaced factory workers.

None of this settles whether AI is a net job creator or destroyer in either country. What the data shows, concretely, is that India's labor market has a structural problem: a hollowed-out middle that predates the current AI wave. Goldman Sachs' own sector-by-sector breakdown puts AI's sharpest edge right where that hole already exists. Whether India's government can move fast enough on the "targeted AI skilling" the Economic Times contributors call for, before the next cohort of Class 12 graduates hits a market with no obvious rung on the ladder, remains unclear.