NAIROBI, Kenya- When AI chatbots like ChatGPT deliver human-like responses, it’s easy to marvel at the tech. But what about the people behind the scenes—the ones teaching these AI systems to think, speak, and identify patterns?
In Kenya, thousands of workers contracted by Sama, a subcontractor for Meta and OpenAI, say their stories reveal a darker side of the AI boom: low wages, grueling work conditions, and lasting psychological harm.
An exposé by 60 Minutes revealed that Kenyan workers tasked with training AI models were promised $12.50 (Sh1,618) per hour but received just $2 (Sh259).
For Naftali Wambalo, a father of two and mathematics graduate, this wage was barely enough to cover basic needs—and that’s before taxes.
Wambalo’s tasks weren’t mundane. He trained AI to identify faces, objects, and places, all while grappling with the knowledge that his contributions powered cutting-edge global technology.
“It’s unfortunate to be the hands behind such powerful tools and earn so little. If big tech is going to do business here, they must do it the right way,” he said.
Sama, which employs over 3,000 workers, defended its pay structure, calling it fair for the region. But for employees like Wambalo, fairness feels far off.
Despite efforts to complete assignments ahead of deadlines, their only reward was fast food and soda—hardly a substitute for proper wages or appreciation.
For many workers, the most disturbing part of the job wasn’t the pay—it was the content they had to sift through.
Beyond the graphic nature of their work, employees reported a lack of mental health support. Many developed severe conditions, including PTSD, depression, and insomnia.
Some, like Wambalo, say the job has even affected their personal lives. “After repeatedly seeing such content, I hate sex,” he admitted, underscoring the long-term impact of their roles.
While companies like Meta and OpenAI deny the allegations and assert a commitment to fair wages and safe work environments, the workers’ accounts paint a grim picture.
The revelations have ignited a broader conversation about tech ethics and labor practices. Kenyan civil rights activist Nerima Wako-Ojiwa labeled the operations an “AI sweatshop,” highlighting the exploitative nature of short-term contracts and low pay.
“We’re so focused on creating opportunities that we forget to ask if they’re good opportunities,” Ojiwa argued.
The backlash isn’t new. In 2023, Kenyan moderators sued OpenAI and Sama, calling for laws to regulate tech outsourcing and ensure worker protections. They demanded government intervention, citing the psychological harm inflicted by their work.
Meanwhile, another AI training platform, Scale AI, faced similar accusations through its Remotasks platform.
Kenyan workers claimed their accounts were closed before payday, leaving them unpaid for their efforts. The company exited the Kenyan market in March 2024, but the lingering impact remains.
