Indian workers paid to film work for AI training data

In Bengaluru, India, a 22-year-old barber named Mohammed Ashim has begun wearing a head-mounted camera for five to six hours a day while cutting hair and trimming beards. The camera records only his hands and fingers; clients’ faces are blurred. He is one of hundreds of thousands of workers worldwide paid to record their activities, with the footage sold to technology companies and research organisations to train artificial intelligence—particularly physical AI, which can control robots in the real world.

Rashdan Technologies, a local company that collects and processes such human-generated data, pays workers $5 to $10 (£3.74 to £7.47) for eight to 10 hours of recorded work, according to co-founder and CEO Shahil Baig. He says consent and comfort are important: hands are recorded, not faces, and faces are blurred in multi-camera rigs. But clients are now requesting four to six cameras rigged around a person’s head to capture the entire scene, a harder ask that commands higher pay.

From barbers to teachers

Manish Agarwal, founder of Humyn Labs, which collects and processes human-generated data, argues that robots cannot be trained only in computer simulations. Physical tasks require high-quality signals involving limb movements, fingers, wrists, elbows, shoulders and dynamic force application—such as applying the right grip force so a plastic bottle does not crumble compared with a steel glass. He says there is no existing internet archive of embodied human activity, so data must be captured from scratch in real-world, diverse environments. Humyn Labs has bases in Bangalore and San Francisco and 96,000 data collectors across Latin America and Southeast Asia. It filters and labels the video before packaging it for clients.

Agarwal notes that India’s current role is largely as a supplier of raw material, with footage shipped to labs abroad. He argues that value will erode as more companies compete, and India needs to move up the value chain—cleaning, structuring, validating and setting standards for high-quality training data.

Some workers see the extra income as worthwhile. Mohammed Arman, who has worked in Bangalore’s textiles industry for 10 years, recently started wearing a head-mounted camera. “This is survival,” he says, explaining that he has a family in Bihar and every rupee matters. Jolly Arifur, a science teacher in Guntakal, Anantapur, has spent the past seven months recording her domestic work—scrubbing, cooking, sweeping tile floors and folding laundry—around her teaching job. She earns $2 to $3 per hour and says it is a good extra income.

Not everyone is optimistic. Bino Paul, a senior professor at the Tata Institute of Social Sciences in Mumbai, says it is too early to tell whether workers are collecting data for machines that will one day do their jobs. AI might be a useful tool for some, but robots could automate other jobs, and if that happens there may be little workers can do. “By the time workers realise what they have participated in, things may have gone too far,” he says. He wants governments to regulate AI, arguing it cannot be viewed merely as a technical system and that ethics must be upheld through human agency.

Despite such concerns, investment in AI continues, and some of it reaches workers in India, where many are prepared to wear a camera for a few dollars a day.

Source: BBC Technology.