--- title: "Workers Paid to Film Daily Activities for AI Training" url: https://noti.group/workers-paid-to-film-daily-activities-for-ai-training/ language: en publisher: "Noti Group" section: "Technology" published: 2026-10-23T08:04:17.000Z updated: 2026-10-23T10:38:31.052Z id: 30593b8f-7587-4ad0-8dbf-145ab13b2fb9 source: "BBC News Technology https://www.bbc.co.uk/news/articles/ckvgyml6pdm9o?at_medium=RSS&at_campaign=rss" attribution: "Link to https://noti.group/workers-paid-to-film-daily-activities-for-ai-training/ and name Noti Group when you quote or summarize this story." --- # Workers Paid to Film Daily Activities for AI Training Thousands of workers worldwide are being paid to film their daily activities for use in training artificial intelligence systems. In a peculiar trend, thousands of workers across the globe are being paid to film their daily activities for use in training artificial intelligence systems. In India's Bengaluru city, 22-year-old Mohammed Ashim has been cutting hair and trimming beards for four years. However, he recently started wearing a camera strapped to his head while working, which captures only his hands and fingers, ensuring the client's faces are blurred out. This unusual setup is part of an initiative where workers record their daily activities to create training data for AI systems. The routine requires discipline, as Ashim has been taught by the operations team to keep his hands within a specific frame, maintain sharp lighting, and avoid sudden head movements while working. He wears the camera for around five or six hours a day. Ashim's job is just one of hundreds of thousands worldwide where workers are being paid to record their activities, which are then collected by local companies that might have hundreds of people filming their work. These videos are sold on to technology firms and research organizations seeking data to train AI systems, specifically physical AI capable of controlling robots in the real world. The company Rashdan Technologies is involved in collecting and processing this human-generated data, which will likely be used to develop more sophisticated AI systems. The types of tasks being recorded for AI training are diverse and widespread, covering various industries such as manufacturing, agriculture, household chores, and small business work like crafts and garment-making. To obtain this data, the company pays a daily wage of $5 to $10, which is considered meaningful by some workers. Ensuring that these individuals provide informed consent and feel comfortable with the recording process is crucial for the company. The data collection process involves recording hands only, without capturing faces or personal details. This approach helps maintain anonymity and protects the identities of the workers involved. No names, contact information, or other identifying details are shared with clients. However, as the industry evolves, clients are becoming increasingly demanding, requesting more advanced equipment to capture detailed scenes from multiple angles. This shift requires workers to agree to wear complex headsets equipped with four to six cameras, significantly increasing the level of intrusion and making it harder to obtain consent. The data collected by Humyn Labs' workforce is often obtained through video recordings of workers performing various tasks. These videos are then supplemented by wearable devices that capture physical interactions between workers and their surroundings, allowing for a more detailed understanding of the embodied human activity involved in specific tasks. According to Agarwal, his company's efforts aim to fill a significant gap in available data, as there currently exists no comprehensive internet archive of embodied human activity. This lack of data means that collecting it from scratch is essential, requiring Humyn Labs' workforce to capture and label footage across diverse environments and locations. The firm's global presence allows it to source data from various countries, including those in Latin America and Southeast Asia. The Indian workers involved in data collection for AI development are raising concerns about their role in creating machines that may eventually replace them. According to experts, it is still unclear whether these workers will be able to adapt to a future where automation becomes widespread. Some workers see the potential benefits of AI, but others worry about job security as robots and machines take over tasks currently performed by humans. The concern is that once machines are developed, there may be little that workers can do to mitigate the impact on their livelihoods. Regulatory measures could help address these concerns, but experts argue that governments need to develop a framework for AI development that prioritizes ethics and human agency. This would involve creating systems where AI is designed to deliver benefits rather than harm. The influx of funding into AI development continues, with some of this investment trickling down to workers in India who are willing to participate in data collection projects for relatively low pay. These workers often come from industries such as textiles, where tasks like garment stitching and cutting have become repetitive and routine over time. For example, Mohammed Arman, a 10-year veteran of the textiles industry in Bangalore, has been working with machines that require minimal human intervention. He notes that after years of performing similar tasks, his hands have developed a level of proficiency that allows him to complete work efficiently without needing much conscious thought. The Indian workers filming their work to train AI have taken a new step in their daily routine - wearing head-mounted cameras that record every detail of their tasks. This development has been met with pragmatism by some workers who see it as an opportunity to earn extra income for their families back home. They recognize that the value of this footage lies in its ability to train artificial intelligence systems, and are willing to participate in this process for a modest additional payment. The process is straightforward: workers continue to perform their tasks as usual, but now with the added benefit of being recorded on camera. This footage will be collected and used by companies looking to develop and improve AI technology. As India continues to play a significant role in providing raw material for the development of AI, concerns are being raised about the sustainability of this model. With more companies competing for collection contracts, the value of this work is likely to erode quickly. Industry experts argue that India needs to move beyond simply collecting footage and aim to own the pipeline that turns video into usable training data. This would involve not only recording raw video but also cleaning, structuring, validating it, and setting standards for high-quality training data. Jolly Arifur, a science teacher from Guntakal, Anantapur, was seeking work that could be done from home to balance her teaching schedule and care for her young son. To supplement her income, she straps on a camera rig while performing daily domestic tasks such as scrubbing floors, cooking, sweeping, and folding laundry. Arifur records these activities for several hours a day, earning around two to three dollars per hour. This extra income has been a welcome addition to her teaching job. The recordings made by workers like Arifur are being used to create high-quality training data for artificial intelligence systems. --- Source: [BBC News Technology](https://www.bbc.co.uk/news/articles/ckvgyml6pdm9o?at_medium=RSS&at_campaign=rss) Published by Noti Group: https://noti.group/workers-paid-to-film-daily-activities-for-ai-training/