In India, gig workers earn as little as $1 per hour. Their task: wear camera-equipped caps and sensor devices to collect intimate first-person video data inside homes and workplaces. This data trains the next generation of physical AI robots. Human Archive, a startup founded by UC Berkeley and Stanford researchers, recently secured $8.2 million from investors like Wing Venture Capital, NVP Capital, and Y Combinator, according to TechCrunch and Zamin Uz.
Yet, as millions flow into advanced AI and robotics, the essential data collection for these technologies often relies on low-wage human labor. This creates a stark ethical disconnect. The rapid expansion of physical AI will likely intensify the demand for human-generated datasets, risking deeper labor exploitation and privacy breaches unless robust ethical guidelines and regulations take hold.
The $1-an-Hour Human Data Pipeline
Human Archive's model is simple: pay workers $1 per hour to collect first-person video data in India, training robots for physical AI, as reported by TechCrunch and Zamin Uz. This approach grants AI companies cheap access to vast real-world data, but at a significant human cost and with inherent privacy risks for both workers and those recorded.
Not an Isolated Incident: The Pronto Controversy
This isn't an isolated case. Bengaluru-based startup Pronto also piloted a program to record its staff providing home services, capturing videos of tasks like cleaning and meal preparation in users' homes for AI training, according to Forbes India. Such initiatives reveal a troubling industry trend: the push for AI advancement often overlooks the ethical implications of data collection.
India's Gig Economy: A Global AI Training Ground
India's gig economy has become a global hub for generating these 'first-person' datasets. Workers record human actions and movements across diverse environments: factories, warehouses, kitchens, and private households, according to The Federal. This widespread reliance on low-wage gig workers for sensitive data, especially in private spaces, poses a systemic ethical challenge at the core of physical AI development.
The Future of Human-Powered AI
Companies like Human Archive and Pronto prioritize speed, building AI models on a foundation of ethically questionable, low-wage data collection, according to TechCrunch and Forbes India. This approach invites significant future liabilities concerning privacy and labor exploitation. The millions from investors like Wing Venture Capital and Y Combinator inadvertently normalize the invasion of private spaces and the exploitation of vulnerable workers. Urgent scrutiny of the entire AI supply chain is demanded.
The 'Pronto controversy' is a stark reminder: the pursuit of advanced robotics often masks a hidden human cost, according to The Federal. Consumers and regulators must confront the true price of convenience. By late 2026, Indian regulatory bodies will likely face mounting pressure to establish clearer guidelines for data collection, potentially forcing startups like Human Archive and Pronto to re-evaluate their acquisition models.
