KARACHI/ISLAMABAD: While much of the global race for artificial intelligence is dominated by billion-dollar chipmakers and Silicon Valley giants, one Pakistani startup believes the country's biggest opportunity lies on factory floors, where skilled workers are helping train the next generation of AI-powered robots.
Founded in January 2026, World Context records factory workers using proprietary sensors, wearable gloves and head-mounted cameras, then converts those recordings into training datasets sold to robotics companies in the United States and China. As companies race to develop humanoid and industrial robots capable of working in warehouses and factories, demand is growing for real-world data showing how experienced workers perform complex manual tasks.
Pakistan's large manufacturing workforce could give it an advantage in that market. According to the latest Pakistan Labour Force Survey (2024–25), 14.8 percent of employed Pakistanis work in manufacturing, much of it involving manual labor. In urban areas, manufacturing accounts for 20.7 percent of total employment.
"We build industrial datasets for robotics companies," Sherdil Ali Khan, CEO of World Context, told Arab News. "We build and prototype sensors here in Pakistan and deploy them across different sectors, across real factory workers and across real industrial environments."
The company says it is currently collecting data from approximately 8,000 workers across textile, automotive, logistics and other manufacturing sectors.
Workers wear specially designed gloves and camera systems that record how they perform everyday industrial tasks, from the movement of their hands and fingers to the pressure they apply during repetitive work. Those recordings are then transformed into datasets that teach AI-powered robots to replicate complex human actions.
Khan believes Pakistan's competitive advantage lies in the expertise factory workers have built over decades.
"There's a lot of handwork," he said. "There's a lot of expertise ingrained into the wisdom of the average factory worker. That expertise is exactly what is most valuable in this sector, as robots must learn from experts."
He argues that, unlike countries where manufacturing has become highly automated, Pakistan still relies heavily on skilled manual production, making it well placed to supply the industrial data needed to train robots.
Khan believes Pakistan's future role in AI will depend less on building frontier AI models than on supplying the infrastructure those models require.
"We don't actually play in the superintelligence field," he said. "The coordination and commoditization of training data will be Pakistan's first empirical contribution to the global sector."
He said Pakistan could become a regional hub for AI data services because of both its industrial workforce and its ability to work with American and Chinese technology companies.
"Pakistan is uniquely positioned to interact with both the Chinese and the American superintelligence economies," he said, adding that the country must now "build infrastructure to generate wealth and generate proximity to the superintelligence race."
"THIS IS GOING TO REMOVE HUMANS"
Among the workers contributing to that effort is 32-year-old factory worker Sunny Manzoor, who wears World Context's sensor equipment during his shift.
Manzoor said the head-mounted device initially felt uncomfortable but quickly became easy to wear without interfering with his work.
"In the beginning it was a little tight, so there was some pain in my head."
Like many workers taking part in the program, Manzoor said he did not initially know exactly how the equipment would be used. After learning the recordings would help train AI-powered robots, he admitted he worried that automation could eventually replace workers like him.
"Obviously, we're human beings," he said. "Obviously, we work for the sake of our children. If this is going to remove the humans, then where will they go? It will become difficult for them to survive."
Still, he believes skilled factory workers retain an advantage over machines.
"The finishing is very detailed," he said. "A robot has to be taught. If something goes wrong, it stops. A human can adapt."
World Context says collecting AI training data could also create an additional source of income for factory workers by paying them not only for manufacturing products but also for helping generate the data used to train future robots.
Khan acknowledged that increasingly capable AI systems would inevitably reshape labor markets but argued that countries like Pakistan would benefit more by participating in the industry than by staying on the sidelines.
"This is not going to change whether or not we participate in this industry," he said. "But whether we get a foot in now or get kind of steamrolled later is definitely up to us."
