A new pipeline of AI development is emerging from an unexpected source: gig workers in Nigeria and other developing nations training humanoid robots from their homes. Medical student Zeus, working from his Lagos apartment, represents a growing workforce recording themselves performing everyday movements—raising hands, walking, gesturing—which AI companies then use to train robot locomotion and manipulation systems. These workers, compensated at rates far below Silicon Valley standards, form the invisible foundation of robot development that companies publicize as breakthroughs. The arrangement exploits geographic wage disparities, allowing AI firms to access affordable, flexible labor while workers gain supplementary income.
This trend raises significant policy and ethical questions about labor standards in AI development. Unlike traditional gig work platforms that at least maintain theoretical oversight, robot training occurs in distributed home environments with minimal regulation or worker protections. There's no guaranteed minimum wage adjusted for local economies, no safety standards for workers repeatedly performing physical movements, and no clarity on how long footage remains stored or how it might be used beyond stated purposes. As these workers contribute directly to technologies worth billions, they remain largely anonymous and powerless in negotiations with multinational corporations.
The phenomenon underscores a critical gap in AI policy frameworks. While regulators focus on algorithmic bias and model safety, they've largely overlooked the labor practices fueling AI advancement. As humanoid robots become increasingly capable and commercially viable, governments must address whether current gig work regulations adequately protect workers training frontier AI systems, and whether companies should implement stronger labor standards and transparency measures regardless of geographic location.