The rapid rise of artificial intelligence (AI) has been hailed as a technological revolution, but the exploitative labour practices fueling its growth are increasingly hard to ignore. The parallels to the outsourcing boom of the 1990s, when U.S. companies took advantage of cheap labour in China, are evident today as AI giants like OpenAI contract global data annotation companies, such as Sama (formerly Samasource), which rely on a vulnerable workforce in Kenya.
Per Federation of Kenya Employers (FKE) data, youth unemployment in Kenya is alarmingly high at 67%, making AI data labelling jobs for global tech firms seem like a ticket out of poverty. However, for many, including Naftali Wambalo, a mathematics graduate, the reality has been far from ideal.
Despite signing up for platforms like Sama, which advertise AI training jobs at $12 per hour, Wambalo and others earn a mere $2 per hour, a staggering 80% reduction in the advertised pay rate. Outsourcing firms, meanwhile, pocket up to $12 per hour from US tech companies, as revealed by CBS News.
This pay disparity exposes a broader issue in AI’s global supply chain, where workers in countries like Kenya are paid pittance wages while their labour drives Silicon Valley’s growth. Kenya, eager to become a tech hub, has promoted itself as a “Silicon Savannah” with attractive incentives and relaxed labour laws to entice companies like Google, Microsoft, and OpenAI. But the tech jobs on offer are far from the promised salvation. Instead, they involve endless hours of labelling disturbing, graphic content—images of violence, abuse, and exploitation—to train AI systems.
“The workforce is so large and desperate that they could pay whatever and have whatever working conditions, and someone will pick up that job,” says Kenyan civil rights activist Nerima Wako-Ojiwa. She likens the situation to “modern-day slavery,” with big American tech companies advertising these jobs as pathways to success, but ultimately perpetuating exploitation.
Naftali Wambalo, a college-educated Kenyan, was among many drawn to the promise of AI jobs, only to face a harsh reality. “I was living paycheck to paycheck. I have saved nothing because it’s not enough,” Wambalo laments. As if the meagre pay isn’t enough, workers like Wambalo are still required to pay taxes, further cutting into their already underwhelming earnings.
The human cost behind AI development
The mental toll of these “humans in the loop” roles is also devastating.”I looked at people being slaughtered, people engaging in sexual activity with animals, people abusing children physically, sexually,” Wambalo recounts, describing the routine exposure to graphic content and lack of support workers face. “You continue isolating yourself from people,” says another worker, Faustine. “You don’t want to socialise. It’s you and it’s you alone.”
Despite claims of ethical AI practices by companies like Meta, Kenyan workers experience a starkly different reality. Sama, the California-based outsourcing firm that partnered with OpenAI, allegedly exploits Kenya’s lax labor laws, paying workers far less than the minimum wage they would receive in their own country. Sama, which prides itself on being an “ethical AI” provider, has been accused of pressuring workers to finish six-month projects in three months, leaving many unpaid for the remainder of their contracts. Some workers even received token rewards like a soda or KFC chicken for their efforts.
In 2019, Meta announced a $22 minimum wage and benefits for US-based contract workers. However, the company has yet to clarify if or when it will implement similar standards for international workers. The gap in treatment between workers in the U.S. and those in developing countries like Kenya underscores the double standards within the tech industry. Other firms like Scale AI, which operates the Remotasks platform, have been accused of similar exploitation, with workers locked out of their accounts just before payday under vague claims of policy violations.
The “AI sweatshops” as Wako-Ojiwa describes them, created by outsourcing companies like Sama — based in the California Bay Area and employing over 3,000 workers in Kenya — thrive on temporary contracts, with workers often lacking job security or benefits, despite the emotionally and mentally taxing nature of their work. Workers have filed lawsuits against Sama alleging poor conditions and unpaid wages. Some employees also claimed the company blacklisted them after their contracts with Meta ended, preventing them from securing similar work with Majorel, Sama’s replacement contractor.
With Kenya’s nearly 20-year-old labour laws failing to account for the realities of digital labour, workers remain vulnerable to exploitation. The problem is not confined to Kenya; other countries like India, the Philippines, and Venezuela also face similar issues, with tech companies benefiting from cheap, highly skilled labour while bypassing ethical concerns.
“The work we do matters,” says Cori Crider, co-founder of Foxglove, an advocacy group for tech workers. As dissatisfaction among workers grows, more than 180 Sama moderators have sued the company for unfair termination and poor conditions. In a landmark ruling, Kenya’s Employment and Labour Relations Court also determined that Meta could be held accountable for Sama’s practices.
Demand for change
Kenyan activists are calling for stronger regulations to protect workers from exploitation. “These companies would never dare impose such conditions in their home countries. Why should Kenya accept this treatment?” Wako-Ojiwa demands.
Scale AI, another American AI training company, has faced similar criticism for failing to pay workers and abruptly shutting down operations in the country.
“What happens is when we start to push back in terms of protection of workers, a lot of these companies, they shut down and they move to a neighbouring country,” says Wako Ojiwa.
The AI industry’s explosive growth brings with it a responsibility to protect the human workers who fuel its progress. It’s important that companies are held accountable for ensuring fair wages, reasonable working hours, and mental health support for their global workforce. The future of AI must not come at the expense of those who power it. If the ethical standards tech giants claim to uphold are to mean anything, the workers driving AI progress must be treated with dignity and respect.
