Hidden Workforce: How AI Depends on Exploited Global South Labor

Artificial intelligence, often touted as autonomous and self-sustaining, depends heavily on human labor, particularly from an estimated 160 million data workers worldwide, according to sociologist Antonio Casilli. These workers, primarily in the Global South, perform tasks such as training and maintaining AI systems, yet remain largely hidden by the companies that employ them.

Casilli, author of "Waiting for Robots: The Hired Hands of Automation" and co-writer of the documentary "In the Belly of AI," argues that these workers are "purposely hidden" because they are essential to AI's functioning. In an interview with Democracy Now!, he explained that data workers are not just labeling images or transcribing text but are crucial to keeping AI systems operational. "Sometimes they replace the machine itself," Casilli said, noting that these workers often earn low wages and may be compensated in non-monetary forms, such as food staples.

The World Bank has estimated that between 130 million and 400 million gig workers perform online tasks that support AI development, with many located in low-income countries where labor costs are lower. Casilli’s research, which includes surveys and interviews in over 30 countries, highlights the precarious nature of this work. In Madagascar, for example, some data workers receive payment in sacks of rice or sugar instead of cash.

The discussion around AI’s impact on labor has gained attention, with figures like Bill Gates warning of significant job displacement. However, the role of data workers in enabling AI is often overlooked. As Casilli noted, "Human workers make automation possible," adding that recognizing their contributions is essential for addressing labor inequalities in the AI industry.

The environmental and social costs of AI are also significant. Data centers, which power AI systems, have a substantial carbon footprint, relying heavily on fossil fuels for electricity. Meanwhile, workers in regions like Kenya and Venezuela describe exploitative conditions, with some calling AI’s promise of poverty alleviation "manipulation."

Critics argue that AI development must prioritize ethical labor practices and transparency. As Casilli emphasized, the hidden workforce behind AI underscores the need for fair compensation and recognition of their contributions to technology that is reshaping global economies.