For just $1 an hour, gig workers in India wear camera-equipped caps, recording their daily lives. This first-person video data fuels multi-million dollar robotics AI startups like Human Archive, according to TechCrunch. Intimate, real-world interactions are collected for future automation.
Venture capital funds millions into robotics AI to automate tasks. However, the foundational training data for these robots is collected by humans performing those same tasks for meager wages, as reported by Startup Fortune. Human Archive, for example, raised $8.2 million from investors including Wing Venture Capital, NVP Capital, and Y Combinator, according to Zamin Uz.
The rapid expansion of robotics AI will likely deepen the global economic divide. A new class of digital laborers is created, whose essential contributions remain undervalued and largely invisible.
The Mechanics of Human Data Collection for AI
Human Archive, a startup that raised $8.2 million from investors like Wing Venture Capital and Y Combinator, specifically collects first-person video data in India. Data, gathered from the service workforce, is crucial for training robots. Its focus on egocentric, real-world task data isn't just a niche; it's a direct response to the industry's struggle to make robots truly functional in complex, human environments.
A Growing Trend Beyond One Startup
Human Archive isn't an anomaly. The trend of human-powered data collection is widespread. Companies like HumynAI Labs, Egodata, Neo Cambrian, XP Robotics, and Objectways actively engage individuals to gather data across tasks, from household chores to manufacturing, as reported by The Economic Times. This isn't just a trend; it's a fundamental shift in how robotics AI is built.
