The future of AI in India is not a single story — it is at least three stories happening at the same time. There is a woman in Tamil Nadu wearing a GoPro on her forehead, recording herself washing dishes so a robot in San Francisco can one day do the same. There is a software engineer in Bangalore watching her company's decade-long dominance quietly erode. And there is a government declaring that India will build its own large language models and lead the world. All three stories are true. None of them is complete on its own.
02:15
Workers at a Karur textile factory wearing GoPro cameras on their foreheads to record egocentric training data for robotics companies
Watch at 02:15 →
India is already the second-largest AI workforce in the world. But size and power are not the same thing. The central question hanging over the country's AI moment is whether India will be a creator of AI or, once again, the world's most efficient and affordable supplier of human labor to build someone else's technology.
What Is the Future of AI in India?
India's AI future is being pulled in two directions simultaneously. On one side, there is enormous optimism: a young population, massive engineering talent, strong English-language skills, and an existing technology services industry worth over $330 billion in annual exports. On the other side, there is a hard reality — India does not produce AI chips, has not built a globally competitive large language model, and private sector R&D investment has lagged far behind the United States and China.
Experts describe India's current position as a scramble. The country is experimenting with everything — building its own LLMs, positioning itself as a data provider, expanding global capability centers — but has not yet found a decisive edge in any single category. What India does have, in almost unlimited supply, is people. Whether that becomes an advantage or a vulnerability is the defining question of this decade.
08:40
Dharni, a 28-year-old teacher turned data worker, explains how she earns 1,000 rupees for 3 hours of recorded household activity
Watch at 08:40 →
Will AI Take Away Jobs in India?
The short answer is: yes, for some jobs, and significantly so. The longer answer is more complicated.
India's IT services industry — the engine that built Bangalore, funded middle-class families across the country, and made India the back office of the world — is facing its most serious threat in 25 years. The work this industry built its reputation on: routine software development, HR automation, accounting systems, testing, and maintenance — is precisely the kind of high-volume, well-defined task that AI does best. As one analyst put it, any profession where the task is well defined will be done by AI.
Companies are already reporting that with automation, 60% of certain workflows can be handled without human workers. You don't need 100 people — you need 40. That is not a theoretical future. That is happening now.
14:22
Poorani, assistant deputy manager at a data annotation firm, describes managing AI training projects across Southeast Asia, Europe, and North America
Watch at 14:22 →
But the picture is not uniformly bleak. The same AI wave that threatens existing jobs is also generating entirely new categories of work. Data annotation, model training, reinforcement learning from human feedback, AI quality assurance — these are growing industries, and India is positioned to capture a large share of them. The question is whether these new jobs will be better jobs, or simply a new form of low-wage digital labor.
What Is Data Annotation Work and Who Does It in India?
Data annotation is the foundational labor of AI. Every piece of information fed into a machine learning model needs to be labeled, explained, and verified by a human being so the computer can learn from it. Self-driving cars need humans to identify every truck, pedestrian, traffic light, and billboard in thousands of hours of footage. AI assistants need humans to rate responses, correct errors, and teach models what good answers look like.
India has had a large data annotation industry for over 15 years. Companies like NextWave recognized early that 60% of India's engineering colleges — and therefore 60% of its graduates — are concentrated in small towns, not major metros. These are first-generation graduates whose parents are farmers, daily wage laborers, and tailors. They took loans to get educated. And they needed jobs that did not require moving to Delhi, Mumbai, or Bangalore.
Data annotation delivered exactly that. Workers in towns like Malasamudram perform annotation tasks with the precision of a factory floor — absolute accuracy, every time, for every piece of data that passes through them. For many women in particular, these jobs have been life-changing: stable income, work close to home, and financial independence in communities that historically denied women both.
What Is Egocentric Data and Why Are Indian Workers Wearing Cameras?
In Karur, a small town in Tamil Nadu, workers at a textile factory are doing something that looks startling at first glance. They are wearing GoPro cameras and Meta smart glasses strapped to their foreheads while they stitch, fold, and iron fabric. They are not being surveilled by their employer. They are being paid to train robots.
This is called egocentric data collection. Companies building humanoid robots need to teach those robots how to perform everyday physical tasks — folding clothes, washing dishes, cleaning a bathroom, picking up objects. The most effective way to gather this training data is to record humans doing those exact tasks from a first-person, eye-level perspective. That is the egocentric view.
One data collection company operating in India explained the scale of demand: clients are expecting 500 million hours of this kind of footage — in a single day. Teachers, homemakers, factory workers, and gig workers across India are being recruited to film their daily lives for platforms that will use that footage to train the next generation of household robots. Workers receive additional pay — roughly 10,000 rupees a month on top of regular wages in some cases — but many are only dimly aware that the footage they record today may eventually train a robot to do the job they currently hold.
How Is AI Threatening India's $340 Billion IT Industry?
India's information technology services sector is the country's single largest export — generating between $330 and $340 billion annually. It built Bangalore. It created the Indian middle class as we know it today. For 25 years, the pitch was simple and devastatingly effective: India could solve your tech problems faster, better, and at 20 cents on the dollar compared to Western alternatives.
That model is now under serious pressure. Large language models can now perform a significant portion of routine software development, testing, and documentation — the exact work that formed the backbone of the Indian IT services business. Companies no longer need to hire a team of engineers to build internal software tools; they can use AI to do it in days.
Critically, Indian IT giants were slow to invest in AI. They were profitable, dominant, and not looking for disruption. As one observer noted bluntly: they were not the vanguard of bringing AI to India — and that was their problem. Private sector R&D investment in India has lagged significantly behind global peers, partly because companies lacked confidence that research investments would pay off at scale.
The result is that an industry born from one wave of globalization — the outsourcing boom of the 1990s — is now facing an existential challenge from the next wave: AI-driven automation of the very services it built its empire on.
Is India Being Exploited in the Global AI Supply Chain?
This is the uncomfortable question that runs beneath every optimistic statistic. India's role in the global AI supply chain — providing data annotation, egocentric training footage, human feedback for model improvement — looks, at its worst, like a new version of an old story: cheap labor building wealth for companies based in California and Beijing, with limited value retained in India itself.
Critics argue that this supply chain is extractive by design. The knowledge embedded in India's languages, behaviors, and cultural practices is being harvested to train AI systems that companies in the West are monetizing in the billions. Workers generating that data are paid per task, with no equity, no intellectual property rights, and no share of the upside.
There are also darker scenarios. In some geographies, workers are reportedly asked to wear tracking cameras by employers with no additional compensation, and the data collected can be used to monitor and discipline the very workers who generated it. The line between opportunity and exploitation is thin — and not always visible to the people on the ground.
The more optimistic view holds that if a price is agreed upon and someone is willing to work, it is a fair trade. But as one commentator asked: are we data working our way out of employment, into a world where the technology companies walk away with the lion's share of the value?
Can India Actually Compete in the Global AI Race?
India has genuine assets: the world's second-largest AI workforce, millions of English-speaking engineers, a proven ability to absorb and deploy technology at scale, and a domestic market of 1.5 billion people generating data in dozens of languages and contexts that no other country can replicate.
But it also has real gaps. It does not manufacture AI chips. It has not produced a globally competitive foundational model. Investors are pulling money out of Indian markets at record pace, redirecting it toward Taiwan and South Korea — countries that make the hardware the AI race runs on. The resources being deployed in the US and China dwarf anything India has committed to AI development.
The most honest assessment may be this: India has the talent. What it has not yet demonstrated is the institutional will, the private sector investment, or the strategic clarity to turn that talent into genuine AI leadership rather than just a more sophisticated version of the same back-office role it has played for the last 25 years. If India is to be a developed nation by 2047, AI is not optional. But becoming the world's AI factory is not the same as winning the AI race. The difference between those two outcomes is the most important policy question India faces right now.
