The race to build robots that can mimic human physical intelligence is intensifying, and it's putting workers in a precarious position. In India, textile factory workers are being asked to record their every move, providing valuable data that could ultimately replace their jobs. This raises important questions about the ethics of data collection and the future of work.
Ashish Narayan, a machine technician in Nagpur, is one of the workers being asked to wear small recording devices on their foreheads. For hours each day, the devices track and record his every action, from adjusting loom tension to fixing jammed machines. Narayan feels like he's working in his own grave, knowing that the data he's providing will contribute to the development of robots that could one day replace him.
This isn't just a concern for Narayan. Ravi Shankar, President of Objectways, a US-based AI data solutions company, acknowledges the fears of workers. Objectways pays workers in India, the USA, Vietnam, Malaysia, and the Philippines to record tasks such as cutting fruits, cleaning utensils, and folding clothes. The data is then sold to robotics firms, contributing to the development of humanoid robots that can perform physical tasks.
The demand for such data is high. Manish Agarwal, co-founder of Humyn Labs, says there's an appetite for "millions and millions of hours" of egocentric data. This data is crucial for training robots to adjust to unpredictable conditions in dynamic environments, such as warehouses, factories, homes, and hospitals.
However, the power imbalance on factory floors is stark. Workers often don't know what's being recorded, where the footage is going, or how it may be used. They rarely have the option to refuse participation, especially in sectors with insecure jobs and weak worker protections. This means that workers are not only producing garments and maintaining machines but also generating valuable behavioral data, with little control over how it may automate or replace their work.
The ethical implications of this data collection are profound. As robots become more sophisticated, the potential for job displacement increases. While robots may perform tasks that humans don't want to do or can't access, the human cost of this technological advancement is a critical consideration. The question remains: how can we ensure that the benefits of automation are shared equitably, and that workers are not left behind in the race for physical intelligence?
