As companies race to build smarter humanoid robots, concerns are mounting among workers that the very data they are helping create could eventually automate their own jobs. (Objectways)The next big raw material in the global artificial intelligence (AI) race may not be chips or compute power, but human actions themselves. Across factories, warehouses, kitchens and retail floors in India, workers wearing head-mounted cameras are recording hundreds of thousands of hours of “egocentric” or first-person data that is increasingly being used to train robotics and embodied AI systems.But as companies race to build smarter humanoid robots, concerns are mounting among workers that the very data they are helping create could eventually automate their own jobs.Egocentric data refers to video and sensor recordings captured from the perspective of the person performing a task. Typically collected using cameras mounted on the head, chest or wrist, the footage shows exactly what a human worker sees while carrying out an activity — picking up an object, assembling components, navigating cluttered spaces or stocking shelves.Unlike traditional robotics datasets captured from static third-person cameras, egocentric datasets are designed to mimic the visual viewpoint a robot itself would have during deployment. For robotics firms, this first-person perspective is critical because it captures fine-grained hand movements, object interactions and real-world environmental complexity that conventional video often misses.Owing to cheap labour availability and weak worker protections, India is emerging as a hub for such data collection.The demand stems from the lack of point-of-view human data, a core challenge in robotics, which prevents them from being able to manipulate their systems in a similar way as humans. Egocentric data reduces this gap by allowing AI systems to learn directly from the perspective they will later operate in.The datasets are particularly valuable for training Vision-Language-Action (VLA) models, which are AI systems that combine visual understanding, language instructions and physical movement. Such models are seen as foundational for future humanoid robots capable of performing household chores, warehouse operations and industrial assembly tasks.Story continues below this adExpert Explains | As chatbots serve ads, what is the conflict between users and AI companies?As is the case with any training dataset, real-world diversity matters enormously. Controlled laboratory settings often fail to capture the unpredictability of actual workplaces, changing lighting conditions, cluttered spaces, interruptions and partial object visibility.Raw first-person footage alone, though, is not sufficient to train robotics systems. Once footage is captured, it is annotated — a labour-intensive process where human reviewers label specific actions and interactions frame by frame. These labels can include hand poses, object locations, grasp points, action boundaries, contact timing and natural-language descriptions of tasks being performed.For example, a clip showing a worker assembling a device may be broken down into micro-actions such as “pick screw”, “align panel”, “tighten fastener” and “place finished unit”. AI systems learn patterns from these annotations to imitate or predict human behaviour.Worker anxietiesThe rapid growth of egocentric data collection has also triggered ethical questions around surveillance, consent and labour displacement. A worker that The Indian Express spoke to was worried that he was effectively helping train machines that may later replace him.Story continues below this adThere is also a major surveillance fear. The same wearable camera and sensor systems used for data gathering could eventually be repurposed as workplace surveillance tools. Continuous first-person monitoring can allow companies to track worker productivity, movement patterns, task completion speed and even idle time in granular detail, and could be a reason why many factories in India are cutting deals with data collection companies.NewsletterFollow our daily newsletter so you never miss anything important. On Wednesday, we answer readers' questions.SubscribePrivacy is another major issue. Egocentric recordings can inadvertently capture faces, private conversations, sensitive workplace processes or bystanders who may not have consented to being recorded.There are also broader questions around ownership and compensation. Workers generating valuable behavioural datasets are often paid for data collection tasks, but may not share in the long-term economic value created when that data helps build commercially successful robotics systems.Soumyarendra Barik is a Special Correspondent with The Indian Express, specializing in the complex and evolving intersection of technology, policy, and society. With over five years of newsroom experience, he is a key voice in documenting how digital transformations impact the daily lives of Indian citizens. Expertise & Focus Areas Barik’s reporting delves into the regulatory and human aspects of the tech world. His core areas of focus include: The Gig Economy: He extensively covers the rights and working conditions of gig workers in India. Tech Policy & Regulation: Analysis of policy interventions that impact Big Tech companies and the broader digital ecosystem. Digital Rights: Reporting on data privacy, internet freedom, and India's prevalent digital divide. Authoritativeness & On-Ground Reporting: Barik is known for his immersive and data-driven approach to journalism. A notable example of his commitment to authentic storytelling involves him tailing a food delivery worker for over 12 hours. This investigative piece quantified the meager earnings and physical toll involved in the profession, providing a verified, ground-level perspective often missing in tech reporting. Personal Interests Outside of the newsroom, Soumyarendra is a self-confessed nerd about horology (watches), follows Formula 1 racing closely, and is an avid football fan. Find all stories by Soumyarendra Barik here. ... Read More © The Indian Express Pvt LtdTags:Express Explained