Thousands of Indian workers are using head-mounted cameras to record their daily tasks to train AI-powered humanoid robots [1].
This data collection represents a critical step for robotics companies seeking to automate complex human movements. However, it creates a paradox where workers provide the exact blueprints needed to automate their own livelihoods.
In Bangalore, factory employees are recording their motions, while in Chennai, garment workers and housewives are doing the same [1], [2], [3]. The footage, known as egocentric video, captures a first-person perspective of tasks ranging from industrial assembly to household chores, such as washing dishes and folding towels [3].
Videos of this practice first went viral in April 2026 [2]. The process requires large amounts of motion data to ensure robots can replicate human dexterity and precision in real-world environments [1], [2].
For those involved, the prospect of automation brings immediate financial anxiety. Lalita, a garment worker in Chennai, expressed concern over the future of her employment as the technology evolves.
"Who is going to pay us when we’re replaced by robots?" Lalita said [1].
Companies are targeting these specific roles because the robots are intended to perform the very jobs the humans are currently filming [1], [2]. The scale of the operation spans multiple sectors, blending industrial labor in Bangalore with domestic work in Chennai [1], [3].
“"Who is going to pay us when we’re replaced by robots?"”
The use of 'egocentric' data collection marks a shift toward high-fidelity imitation learning in robotics. By leveraging low-cost labor in India to map human movement, AI firms can accelerate the deployment of humanoid robots. This creates a precarious economic cycle where the workforce is actively subsidizing the development of the technology that will likely render their specific skill sets obsolete.
