'For Rs 250 per hour': How Indians are selling their future to American AI companies

'For Rs 250 per hour': How Indians are selling their future to American AI companies

'For Rs 250 per hour': How Indians are selling their future to American AI companies Photograph: (AFP)

Story highlights

Behind every chatbot and humanoid robot demo is an invisible workforce and a huge share of it is in India. Indian workers are labelling data and training AI systems for American tech giants, often the very systems that could one day take their jobs.

Every time a chatbot gives a polished answer or a humanoid robot demo goes viral, there's an invisible workforce behind it - and increasingly, a large part of that workforce is in India. For wages that can work out to just a couple of dollars an hour, Indian workers are doing the repetitive, unglamorous labour of labelling data, correcting AI outputs, and even acting out tasks on camera so robots can learn to mimic them.

A Booming But Invisible Industry

India's data annotation industry has grown explosively - from roughly $250 million in 2020-21 to a market expected to cross $7 billion by 2030. That growth is being driven almost entirely by demand from US AI labs, which need vast amounts of human-labelled data to train and fine-tune their models. The work itself is mundane: tagging images, transcribing audio, rating chatbot responses, and flagging errors - millions of small tasks that, stitched together, make large AI models usable.

Add WION as a Preferred Source

The Wage Gap Is Stark

While American annotators working on similar tasks for the same companies have reportedly earned between $15 and $25 an hour, workers in India, Kenya, the Philippines and Venezuela doing comparable work have been paid as little as $1 to $3 an hour - a figure that, converted, lands in the range of roughly Rs 100 to Rs 250 an hour. Major US AI labs rarely hire these workers directly; instead, they go through specialist contracting platforms, which in turn route the work to whichever country offers the cheapest labour at any given time. If workers in one country push back on pay, the platforms simply shift the work elsewhere - keeping wages low and workers replaceable.

Training The Robots That Could Replace Them

Trending Stories

Perhaps the most striking part of this story is the newest frontier: humanoid robots. Companies have begun hiring Indian workers - often in factory settings - to wear lightweight cameras and perform everyday tasks, generating ‘egocentric’ video data that is then used to teach robots how to do that same work. One such operation, Egolab.AI - whose India arm is based in Bengaluru - partners with factories across the country, including garment manufacturers, to collect this kind of footage from workers for robotics projects worldwide. In effect, some workers are now being paid small sums to record themselves doing the jobs that the resulting AI and robotics systems may eventually be used to automate.

No Safety Net, Constant Surveillance

Most of these workers are classified as independent contractors rather than employees, which means they fall outside standard protections like minimum wage enforcement, paid leave, or severance. India's existing labour laws, including the Code on Wages, were not designed for piece-rate, task-by-task digital work and struggle to regulate it. On top of low pay, workers describe intense surveillance - every click, keystroke and submission tracked - and a constant fear of being silently removed from a platform for a single mistake, with no manager or appeals process to turn to.

The Bigger Picture

AI companies in Silicon Valley are valued in the hundreds of billions of dollars, built in large part on the unglamorous, low-paid work happening in places like India. As India positions itself as a hub for the global AI economy - with government summits, mega data centre deals and ambitious adoption targets - this other side of the story rarely gets the same spotlight: a growing number of Indians whose daily work is to make foreign AI systems smarter, faster and more capable, often for a fraction of what the same work pays elsewhere in the world.

About the Author

Tarun Mishra

Tarun Mishra is a Sub-Editor at WION. He has worked with leading outlets, covering business, global affairs, technology, space exploration and culture. With a diverse background sp...Read More