A new frontier in artificial intelligence is emerging, one that requires the 'eyes' of human laborers to help machines navigate the physical world. In Indian textile factories and on the streets of American cities, workers are increasingly being outfitted with wearable cameras to record their every move. This 'first-person perspective' (POV) data has become the most valuable currency for companies building the next generation of humanoid robots.
Traditionally, robots were trained in computer-generated 3D environments, but these simulations often fail to capture the messy reality of physical labor. By harvesting real-world visual data, firms like Egolab AI are creating massive datasets that teach machines how to sew a sleeve or wash a dish. This shift marks a transition from 'digital AI' to 'physical intelligence,' where the target is the automation of manual labor.
In India's Gurugram garment hub, the reality for workers is far from a voluntary contribution to science. Employees report being forced to wear head-mounted cameras for entire shifts, often without their consent. Many describe the devices as uncomfortable, with batteries heating up near their temples, leading some to describe the technology as 'sucking our blood' while recording their most private conversations and movements.
This data is not just used for training; it is also being repurposed for invasive surveillance. Sample reports from data aggregators show detailed 'productivity analyses' that rank workers from best to worst, calculating exactly how much time is lost to chatting or idle breaks. This creates a double-edged sword: workers are simultaneously being squeezed for efficiency today while providing the blueprints for their replacement tomorrow.
The phenomenon is not limited to the Global South. In the United States, gig giants like DoorDash and Uber are incentivizing their delivery fleets to record domestic chores, such as folding laundry or making beds, for additional pay. By reframing labor as 'data labeling,' these platforms are tapping into a vast, low-cost pool of human intuition that roboticists at Tesla and Boston Dynamics desperately need.
As tech companies race toward 'General Purpose Robotics,' the ethical gap continues to widen. While startups claim these workers are 'contributing to progress,' the lack of collective data rights means that individual laborers have no share in the immense value they are generating. The global supply chain of AI is now being built on the literal movements of the very people who stand the most to lose from its success.