Training AI takes heavy toll on Kenyans workers


Training AI takes heavy toll on Kenyans workers

 Recent investigations have highlighted the exploitation of Kenyan data labelers by major tech companies, revealing severe mental health challenges and poor working conditions. A 2023 Time investigation uncovered that OpenAI outsourced data labeling tasks to Kenyan workers earning less than $2 per hour. These workers were tasked with reviewing and labeling vast amounts of content, including graphic and disturbing material, to train AI models like ChatGPT. The exposure to such content led to significant mental health issues among the workers. ​​

In a related case, former Facebook content moderators in Kenya filed a lawsuit against Meta and its outsourcing partner, Samasource, alleging severe mental health impacts from their work. Between 2019 and 2023, 185 moderators were exposed to distressing content, leading to diagnoses of PTSD, depression, and anxiety. The lawsuit highlights the lack of adequate mental health support and the harsh working conditions these moderators faced. ​​

These revelations have spurred efforts among Kenyan workers to advocate for better conditions. For instance, Mophat Okinyi, a former content moderator, co-founded the Content Moderators Union to protect AI data workers' rights in Africa. The union has grown to over 400 members and is working towards official recognition. Additionally, Okinyi established Techworker Community Africa, an NGO focused on educating African AI workers about their rights and providing mental health support. ​​

In November 2024, Ai Kenya, in partnership with the Pulitzer Center, hosted a town hall event titled "Redefining AI Labor: Ensuring Fairness and Equity for Data Workers." This event brought together data workers, tech leaders, labor advocates, and policymakers to discuss labor practices in AI-driven industries and propose actionable solutions to improve the AI labor ecosystem. ​​

These cases underscore the urgent need for ethical AI development practices that prioritize the well-being of data labelers. As AI continues to evolve, it is imperative to address the human costs associated with its development and ensure fair treatment and support for all workers involved.

Training AI takes heavy toll on Kenyans working for $2 an hour​​



Comments

Popular posts from this blog

African Innovation-driven economies

The WikiExplorers and the Shell Island of Joal-Fadiouth

The Influence of Corporately Owned Celebrities on Political Elections