The Exploitation of Kenyan Workers in AI Development
Kenya has become a hub for outsourcing labor in AI development due to its digitally connected and tech-savvy workforce. Yet, these workers are often subjected to substandard pay and harsh working conditions.
The Economic Inequality in AI Outsourcing
Tech giants such as OpenAI, Meta, and Microsoft outsource tasks like data labeling and moderation to countries like Kenya, India, and the Philippines. While these companies generate billions in revenue, their workers are paid as little as $2 per hour.
Kenyan vs. U.S. Pay Disparity
For the same work, U.S. counterparts receive up to $12 per hour or more. This stark pay difference perpetuates global inequality, as corporations exploit the economic desperation in developing countries to cut costs.
Workers in Kenya often label violent or inappropriate content, training AI systems like ChatGPT to filter and moderate content.
This task exposes workers to traumatic material, such as videos of violence, abuse, and torture.
Digital Desperation
Kenyan workers are highly educated, tech-savvy, and fluent in English, making them ideal candidates for these roles. However, the country’s high unemployment rate forces them to accept exploitative pay to survive.
The Psychological Toll of Content Moderation
One of the most disturbing aspects of this exploitation is the psychological harm inflicted on content moderators.
Trauma from Content Moderation
Workers are required to sift through the most heinous material imaginable—graphic violence, sexual abuse, and child exploitation. Many workers report long-term mental health issues, including depression, anxiety, and PTSD.
Some moderators have described feeling permanently altered, unable to return to normal life even after leaving the job.
Insufficient Support
Companies often provide minimal mental health resources to their workers, leaving them to cope with trauma on their own. For example, Kenyan workers employed by outsourcing firms contracted by Meta reported that the only “perk” for meeting their targets was free fast food, like KFC, rather than adequate mental health support.
Legal Repercussions and Worker Resistance
A growing number of former workers are filing lawsuits against companies like Sama, the outsourcing firm hired by Meta, for failing to provide adequate support or pay.
Kenyan Workers vs. Meta and Sama
In 2023, content moderators in Kenya sued Meta and Sama for exploiting workers and failing to address the psychological harm caused by their work. The workers also alleged that Sama violated Kenyan labor laws by underpaying staff.
How Corporations Justify Low Wages
The justification for these wages is often rooted in cost-cutting and the claim that these jobs provide employment in economically disadvantaged regions. However, this narrative ignores:
The True Cost of Labor
While $2 per hour may seem significant in Kenya’s local economy, it does not account for the nature of the work or its long-term psychological toll.
Global Inequality
These corporations profit immensely from AI systems trained on the labor of workers in poor countries, yet fail to distribute wealth equitably.
Broader Implications for AI Development
This systemic exploitation has several implications:
- AI Training on Labeled Data
The labeled data used to train AI systems like ChatGPT and DALL-E relies on human input. Workers are tasked with identifying and labeling images, videos, and text, which requires human intelligence and judgment.
This labor is essential but grossly undervalued, as evidenced by the minimal pay.
- Outsourcing to Poor Countries
By outsourcing to countries with low labor costs, the richest companies evade responsibility for fair wages and working conditions. - Policy Gaps
Governments in countries like Kenya struggle to regulate these exploitative practices, even as they try to create jobs in the tech industry.
Call for Change
The exploitation of workers in developing countries is not inevitable. Solutions include:
Stronger Labor Protections
Governments must enforce labor laws that protect workers from exploitation and psychological harm.
Fair Pay
Tech companies must offer wages that reflect the true value of the work performed.
Mental Health Support
Workers should receive comprehensive mental health services to mitigate the trauma of content moderation.
Global Accountability
International organizations must hold corporations accountable for exploitative practices in developing countries.
Conclusion
Kenyan workers and others in developing nations play a critical role in shaping AI systems, yet they are often left behind in the economic benefits these technologies bring. By paying fair wages and providing adequate support, tech companies can address the glaring inequality in AI development while respecting the dignity and humanity of their workers.