Kenyan AI Workers Face Harsh Realities Amid Promises of Tech

November 24, 2024 – Nairobi, Kenya

A recent investigation by CBS News has uncovered the grim conditions faced by Kenyan workers employed in the artificial intelligence (AI) sector, revealing a stark contrast between the promises of tech job opportunities and the harsh realities on the ground.

Kenyan civil rights activist Nerima Wako-Ojiwa highlighted the exploitation faced by workers lured by U.S. tech companies with the promise of AI jobs. Despite Kenya’s high unemployment rate, which stands at 67% among young people, the influx of tech giants has not brought the prosperity many hoped for. Instead, workers are overworked, underpaid, and subjected to poor working conditions.

The Attraction of AI Jobs in Kenya

Nairobi has become a significant hub for AI-related jobs due to its large, educated, yet underemployed population. The Kenyan government, keen to position the country as a “Silicon Savannah,” has offered financial incentives and promoted lax labor laws to attract companies like Microsoft, Google, Apple, and Intel.

Naftali Wambalo, a mathematics graduate and father of two, was thrilled to secure a job in this emerging field. Known as “humans in the loop,” Wambalo and his colleagues spend hours labeling data to train AI systems for companies such as Meta, OpenAI, Microsoft, and Google. Despite the importance of their work, which includes training AI to recognize everything from cars for autonomous vehicles to abnormalities in medical images, the pay and job security are abysmal.

Unfair Labor Practices

The pay for these crucial roles ranges from $1.50 to $2 an hour before tax, a stark contrast to the $12.50 an hour companies like OpenAI pay to outsourcing firms such as SAMA. Wambalo and others like Nathan Nkunzimana and Fasica Berhane Gebrekidan report a lack of stability, with some contracts lasting only a few days to a month.

Workers describe the conditions as akin to “AI sweatshops,” with unrealistic productivity demands and inadequate compensation. SAMA, an American outsourcing company employing over 3,000 workers in Kenya, has been particularly criticized for these practices. Documents reviewed by CBS News confirm that OpenAI paid SAMA significantly more per worker than the wages received by employees.

The Mental Toll of AI Training

Beyond financial exploitation, the nature of the work has taken a severe toll on the mental health of workers. Many, like Wambalo, were tasked with reviewing highly distressing content, including violence, pornography, and abuse, to train AI systems to filter such material. The psychological impact of this work has led to severe mental health issues, with inadequate support provided by the companies.

Wambalo’s marriage has suffered, and Berhane Gebrekidan finds it difficult to engage in daily conversations, preferring to cry over speaking. Both are among approximately 200 digital workers suing SAMA and Meta over the working conditions and the psychological harm they have endured.

Calls for Reform and Responsibility

Wako-Ojiwa and the affected workers are calling for stricter labor laws and better protection for digital laborers. They argue that the responsibility lies not only with the outsourcing firms but also with the tech giants providing the work.

Despite the challenges, there is hope that increased awareness and legal action will lead to better conditions for Kenyan AI workers. The tech companies involved, including Meta and OpenAI, have stated their commitment to fair wages and safe working conditions, but workers and advocates demand concrete actions to match these promises.

As Kenya continues to position itself as a leader in the digital economy, the plight of its AI workers serves as a critical reminder of the need for ethical labor practices and the protection of workers’ rights in the age of technology.