by STEPHANIE WANGARI & GAYATHRI VAIDYANATHAN
New data reveals the hidden network of African workers powering AI, as they push for transparency from the global companies that employ them indirectly.
- The people behind AI data training and other digital work are often hidden.
- A new map shows workers in 39 African nations employed by outsourcing firms in the U.S., Europe, and Asia.
- Some intermediaries did not sufficiently comply with legal requests for worker data.
Firms that provide outsourced digital labor for big tech companies tend to be secretive. They are often bound by legal contracts that limit what they can say, allowing tech companies to distance themselves legally and ethically from their workers, experts told Rest of World.
“This creates a circle of invisibility around this work,” Antonio Casilli, a sociologist at Polytechnic Institute of Paris who studies the human contributors to artificial intelligence, told Rest of World. Casilli was not involved in the research mentioned below. “Sometimes, I interview people that work for a big company, [and] they sometimes don’t even know when and how many workers they have.”
A new dataset, visualized as maps, reveals the extent to which African workers are indirectly employed in the tech sector, doing content moderation, customer service, and data annotation for AI models, among other jobs.
Outsourcing firms and their African offices
Companies in the U.S., Europe and Asia hire African workers for training AI models, content moderation and other digital jobs
One of the maps shows the flow of data and knowledge out of 39 African nations to subcontractors, mostly located in the United Arab Emirates, North America, and Europe, with four outsourcing firms in Africa. From there, it goes on to clients such as Meta, OpenAI, and Samsung. The research was conducted by the African Content Moderators Union (ACMU) and Switzerland-based nonprofit, Personaldata.io.
Rest of World for more