The sight of garment factory workers in Delhi sporting head-mounted cameras has sparked alarm among local employees, who fear their work is being scrutinised to replace them with robots. Lalita, 32, was initially amused by the unusual request but soon grew concerned that her productivity was under the microscope and her job at risk.
These recordings, called 'egocentric data', are crucial in developing humanoid robots capable of navigating real-world environments. Unlike language models, which learn from vast amounts of online text, robots need to be trained on direct observations of human physical work – such as intricate hand movements and task precision. India is emerging as a key hub for collecting this valuable data due to its diverse labour force and extensive industrial base.
The country's numerous robotics companies, including EgoLab, Humyn AI, and Labellerr AI, specialise in building data pipelines for the industry. EgoLab reportedly counts Tesla among its clients, reflecting the high demand from leading tech firms. The process involves not only capturing footage but also meticulously cleaning and annotating it to ensure clarity and accuracy.
India's cost-effectiveness compared to Western countries is a major driver for companies seeking to scale their automation efforts. However, the practice raises ethical concerns: workers are often unaware that their movements are being used to train machines that could potentially replace them, leaving them anxious about their future employment. As the AI and robotics industry intensifies, finding a balance between technological advancement and worker welfare becomes increasingly complex.
The Indian government has yet to address these concerns, despite widespread criticism from labour rights groups. Meanwhile, experts argue that workers should be fully informed of the data collection process and its implications for their jobs. As AI-powered robots become increasingly prevalent in industries worldwide, it's essential to consider the human cost behind the technology.