A growing workforce of gig laborers in developing countries is quietly powering the next generation of AI robotics. Workers like Zeus, a medical student in Nigeria, are monetizing their spare time by recording themselves performing everyday tasks—raising their hands, walking, manipulating objects—which are then used to train humanoid robots. These workers connect to platforms offering flexible income, strapping iPhones to their foreheads to capture video data that companies use to improve machine learning models. The practice highlights how AI development has become deeply dependent on human labor, yet often invisible to consumers benefiting from these technologies.