Will AI Take Your Job? 197 Million Workers and Who Really Bears the Risk
Every few weeks a new number lands and the headlines reach for the same word: threat. This month the number is a big one. A new study estimates that around 197 million Indian workers hold jobs exposed to AI automation. It is the kind of figure designed to make you anxious, and anxiety is a reasonable first response. But as a sociologist, I want to ask a different question than the headlines do. Not simply “will AI take jobs?” — but “whose jobs, and who decides who bears the cost?” Because the answer to that question tells you far more about the society we are building than any single statistic.
This is not a techno-panic piece, and it is not a cheerful “AI will create more than it destroys” piece either. It is an attempt to read the automation debate the way it deserves to be read: as a question about power, class and who absorbs the shock of change.
What the numbers actually say
Two reports this month frame the debate. A study from the Great Lakes Institute of Management estimates that roughly 197 million Indian workers are in roles heavy in routine tasks that could be exposed to AI automation, with a much smaller group — around 46 million — in jobs where AI mainly boosts productivity (ThePrint, 2026). Importantly, the authors are careful: this does not mean 197 million jobs will vanish. It means that is the scale of the workforce that could be affected.
From the other side of the world, a McKinsey analysis of the United States argues something subtler than mass unemployment. It expects AI and automation to reshuffle work rather than simply destroy it — millions of workers may need to change occupations entirely, but growth elsewhere could offset the losses. Their phrase for the coming decade is that the challenge is “mobility, not scarcity” (Fortune, 2026). Hold onto that phrase; we will come back to it, because it is doing more work than it appears to.
“Exposed” is not the same as “doomed”
Before anyone panics, a crucial distinction. When a study says a job is “exposed” or “vulnerable” to automation, it means the job contains tasks a machine could, in principle, do. It does not mean the job will disappear, or disappear soon. Technology adoption is slow, uneven, shaped by cost, regulation, culture and resistance. Jobs are also bundles of tasks, and automating some tasks often reshapes a role rather than erasing it. So treat the big scary number as a map of pressure, not a death list. The sociologically interesting question is what happens to people as that pressure builds.
The real story: automation follows the fault lines we already have
Here is the pattern that the headline numbers tend to bury. Automation risk is not spread evenly across society like rain. It concentrates. The jobs most exposed are those built on routine, repetitive tasks — much office and administrative support, data entry, some retail and transport work — and these are disproportionately lower-paid and already-precarious roles. In other words, the technology does not fall on an undifferentiated “workforce.” It tracks the lines of class, income, gender and region that already divide who has security and who does not.
This is why sociologists have long argued that technological change is never purely technical. Decades ago, Harry Braverman described how industrial technology was used not just to increase output but to deskill work and shift control away from workers. The tools are new; the dynamic is familiar. AI becomes one more mechanism through which risk is pushed downward — toward the people least cushioned to absorb it.
India’s informal twist
Most of the loudest automation commentary comes from the Global North, and it quietly assumes a world of formal jobs, contracts and benefits. India’s labour market looks very different. The overwhelming majority of Indian workers are in the informal sector — without written contracts, pensions or protections. That changes the automation story in two directions at once. On one hand, informal and rural work may be slower to automate. On the other, when disruption does come, informal workers have almost no institutional cushion to fall back on: no redundancy pay, no retraining entitlement, no safety net. The same shock that is a career setback for a salaried professional can be a livelihood collapse for an informal worker. A serious conversation about AI and Indian jobs has to start there, not with Silicon Valley’s assumptions.
The quiet politics of “reskilling”
Now back to that phrase — “mobility, not scarcity.” It sounds reassuring, even optimistic. But listen to what it does. It relocates the problem. If the challenge is mobility, then the task is for workers to move — to retrain, to upskill, to adapt. The responsibility quietly shifts onto the individual. “Just reskill” can become a way of saying: the disruption is coming, and managing it is your problem now.
I am not against reskilling — of course people should have the chance to learn and adapt. But we should notice when a structural change is reframed as a personal to-do list. The 55-year-old data-entry clerk, the small-town clerk with caring responsibilities, the informal worker with no savings — “mobility” is not equally available to them. Treating adaptation as purely a matter of individual effort ignores the structural barriers that determine who can move and who gets left behind. That is a political choice dressed up as common sense.
What a fairer response would look like
If automation concentrates risk on those least able to bear it, then justice lies in redistributing that risk — not leaving it where it falls. That points beyond reskilling slogans toward the harder, less fashionable work: strengthening social protection so that transition does not mean destitution; labour policy that recognises informal workers; genuine, funded retraining rather than motivational posters; and a public conversation that treats displaced workers as citizens owed support, not as units who failed to adapt. These are the questions AI is really asking us — and they are social and political questions, not technical ones. On how AI is reshaping research and knowledge work specifically, see my piece on how AI is transforming social research, and on using it responsibly, can I use ChatGPT for my thesis.
Frequently asked questions
How many Indian jobs are at risk from AI?
A 2026 Great Lakes Institute study estimates around 197 million workers are in routine-heavy roles exposed to automation — but “exposed” means potentially affected, not certain to disappear.
Will AI cause mass unemployment?
Most serious analyses expect AI to reshape and relocate work rather than cause a simple net collapse. The main challenge is helping people move into new jobs — though the disruption is very unevenly shared.
Which jobs are most at risk from automation?
Routine, repetitive roles — much office and administrative support, data entry, some retail and transport. These are often lower-paid and already-precarious, which is why automation tends to widen inequality.
Is reskilling the answer to AI job loss?
It helps, but alone it shifts the burden onto individuals. A fair response also needs social protection, labour policy and funded transition support.
Why does automation affect some groups more than others?
Because exposure tracks existing divides of class, income, gender and region. The people in the most routine, least protected jobs face the most risk and have the fewest cushions.
Final thought
The next time a 197-million-sized number flashes across your screen, resist the two easy reactions — blind panic and breezy optimism. Ask the sociologist’s question instead: who carries this risk, and who gets to decide? AI will not simply “take jobs” in some neutral, weatherlike way. It will be deployed, governed and absorbed within a society that already decides whose security matters. Whether this wave widens our divides or is managed with fairness is not a question for the technology. It is a question for us.
References
Mahambare, V., Jadhav, V., & Venugopalan, V. (2026). AI, Automation and India’s Workforce: Who Gains, Who’s Exposed, and How Policy Can Shape the Future (Working Paper). Centre for Finance and Economics Research, Great Lakes Institute of Management, Chennai. https://greatlakes.edu.in/chennai/assets/greatlakes/images/GLCFER-WP5.pdf
Khatoon, A. (2026, October 11). Nearly 197 million Indian workers in jobs vulnerable to AI automation, finds Great Lakes Institute study. ThePrint. https://theprint.in/feature/nearly-197-million-indian-workers-in-jobs-vulnerable-to-ai-automation-finds-great-lakes-institute-study/3068436/
Gioino, C. (2026, October 3). McKinsey: AI will create more jobs than it kills — after destroying 11 million. Fortune. https://fortune.com/2026/10/03/ai-job-displacement-creation-new-careers-11-million-americans-workforce-mobility/
McKinsey Global Institute. (2026). Workforce in Motion: Skills and Pathways to Future Jobs in the United States. https://www.mckinsey.com/mgi/our-research/workforce-in-motion-skills-and-pathways-to-future-jobs-in-the-united-states
Braverman, H. (1974). Labor and Monopoly Capital: The Degradation of Work in the Twentieth Century. Monthly Review Press.