Mohammed Ashim has spent four years cutting hair and trimming beards in Bengaluru. Recently, the 22-year-old has added another task to his routine: wearing a head-mounted camera to record his work.
He wears the camera for five to six hours a day, earning extra money while carrying out his usual job.
"It's good money, but honestly, the feeling of doing something this futuristic is what keeps me excited. I never imagined my razor techniques would be useful to a global tech company," he told the BBC.
Ashim is among workers worldwide being paid to record their everyday activities. Local companies sell the footage to technology firms and research organisations to train artificial intelligence (AI), particularly systems designed to control robots in the physical world.
How everyday work trains AI
Rashdan Technologies collects human-generated data for AI training. Its clients seek footage of industrial operations, agriculture, household chores and small-business tasks.
"Every robotics company, every frontier AI lab, has a specific kind of data they want: industrial tasks, agriculture, household chores, small business work like pottery, garment-making, embroidery, candle-making. We go wherever the task is," said Shahil Baig, the company's co-founder and chief executive.
Rashdan pays $5 to $10 for eight to 10 hours of recorded work. Baig said workers' consent and comfort were important. The company usually records hands rather than faces, blurring faces in footage that shows more of the body or surroundings. Names and contact details are excluded from the material supplied to clients.
Clients increasingly want more detailed recordings. "Now clients are asking for four to six cameras rigged around a person's head to cover the entire scene, not just the hands," Baig said. He added that workers receive higher pay for wearing multi-camera rigs, which are harder to agree to than a single camera.
Why robots need human-generated data
Manish Agarwal, founder of Humyn Labs, said robots cannot reliably operate in the real world if trained only through computer simulations. His company collects and processes human-generated data, mostly video, while some collectors wear devices on their hands and fingers to capture physical contact.
Humyn Labs has 96,000 data collectors across countries in Latin America and Southeast Asia, with bases in Bengaluru and San Francisco. It reviews thousands of hours of footage, selects useful material, labels activities and prepares the data for clients. Agarwal said recordings should reflect the environments where robots will be used: footage from Indian homes may not be suitable for machines designed for US households.
Could workers be training machines to replace them eventually?
Bino Paul, a senior professor at the Tata Institute of Social Sciences in Mumbai, said it was too early to know how AI would affect jobs. It could help some workers while enabling automation in other occupations.
"By the time workers realise what they have participated in, things may have gone too far," he said, calling for government regulation and ethical safeguards.
Survival
Mohammed Arman, who has worked in Bengaluru's textile industry for 10 years, recently began recording himself cutting and stitching garments.
Arman said the extra income mattered to his family in Bihar. "If someone offers me extra money to wear a camera while I do the exact same work I've been doing for 10 years, I'm not going to say no," he said.
Agarwal warned that India risks remaining a supplier of raw footage to overseas AI labs. As competition grows, he said, the country should develop expertise in cleaning, structuring and validating data rather than simply recording it.
Science teacher Jolly Arifur, from Guntakal in Anantapur, also films daily chores while caring for her young son and continuing to teach. For seven months, she has recorded herself cooking, sweeping and folding laundry.
"I get paid two to three dollars per hour. If I record three to five hours a day, that's a good amount in hand — it's become an extra source of income for me, on top of teaching," she told BBC.