Interactive Map Reveals Hidden Global Labor Powering AI

Brown University undergraduate Cunyan Ma, working with the university's Center for Technological Responsibility, Reimagination and Redesign, has created an interactive online map designed to expose the human labor behind artificial intelligence systems that power social media feeds, navigation apps, entertainment recommendations and other digital services. The project links technology firms and other large corporate customers to data-labeling companies and their delivery centers, highlighting how much AI training, moderation and evaluation work is outsourced to low-wage workers in the Global South, including Kenya, Indonesia and Colombia. The map combines existing databases, Ma's own records of data-labeling firms' delivery sites and news reports on working conditions. In its current form, it connects 13 data-labeling companies, 141 corporate clients and 31 countries where data worker sites are located. The tool is not presented as a complete account of the global AI labor chain, but as an illustrative effort to make opaque outsourcing networks more visible. Much AI training requires human oversight, from labeling objects and speech to rating chatbot responses, transcribing audio and identifying features in medical images. More sensitive tasks require workers to review violent, abusive or otherwise disturbing content so systems can recognize and filter it. Ma frames this arrangement as AI colonialism, describing it as the extraction of difficult, low-paid labor from the Global South to support profitable AI products developed and marketed in the West. Brown computer science professor Suresh Venkatasubramanian, who advises the project, said the map reinforces that AI is built from aggregated human knowledge and effort, not an autonomous sentient force. Ma said she plans to expand the project to other areas, including semiconductor manufacturing, and hopes future versions could help distributed data workers identify one another, strengthening prospects for coordination or unionization. The broader goal is to promote a more equitable distribution of both AI labor and the benefits it creates.