In a modest kitchen in Chennai, Nagireddy Sriramyachandra straps a smartphone to her head and films herself slicing mangoes. The 25-year-old housewife is not a social media creator but a foot soldier in a global artificial intelligence supply chain, earning around $2 an hour to produce footage that trains robots to mimic human movement. Her work is part of a vast, informal data-labelling industry that has turned India, the world’s most populous country, into a training ground for domestic automation. “Who else will give you 250 rupees an hour just for doing housework?” she asks, aware that the machines she is teaching may one day make such paid chores obsolete.

Viewed from New Delhi, this paradox is reshaping labour markets. Thousands of Indians are now performing and recording mundane tasks — folding laundry, sweeping floors, cooking — for technology companies racing to build general-purpose home robots. The pay, though meagre by Western standards, outstrips many local alternatives, drawing a hidden army of trainers who are effectively scripting their own displacement. Analysts in London note that this mirrors earlier waves of automation: low-wage workers in emerging economies are often the first to both enable and bear the brunt of technological disruption.

In China, the next chapter is already being written. In Beijing, 56-year-old cleaner Lin Meiqiong now works alongside a wheeled, camera-studded robot that can grip and move objects. Deployed by the platform 58.com, these human-robot hybrid teams are still novelties — more effective at generating data and customer curiosity than at independent labour. “Definitely, it’s different,” a client remarks. The Chinese experiment underscores a geographic division: India produces the foundational training data, China tests early physical deployments, and Western companies often develop the underlying algorithms — a tripartite structure that distributes risk and reward unevenly.

The ethical calculus grows more complex when considering the long arc of this development. Workers like Sriramyachandra are not only cheap data generators but also potential consumers — she muses about one day owning a robot herself. Yet as domestic robots inch from Mumbai test kitchens to Beijing apartments and eventually to London or New York homes, the millions employed in the global care and domestic work economy face an uncertain future. Forward-looking analysts warn that without policy frameworks to manage the transition, the current training boom may simply accelerate a race to the bottom, leaving those who taught the machines with little more than the memory of a modest hourly wage.