In Kenya, a group of artificial intelligence workers has formed the Data Labelers Association (DLA). Their goal is clear: they want fair pay, better working conditions, and support for their mental health.
These workers play a crucial role in training AI systems for big tech companies. In just one week, 339 members joined the DLA, driven by the need to confront the injustices they face daily. There’s a common belief that AI operates independently, but the truth is, it relies heavily on human effort to perform even the simplest tasks.
This work, often called ghost, micro, or click work, involves breaking down various tasks essential for AI development. Kenya has emerged as a key player in this field, yet data workers remain severely underpaid, often earning mere cents for hours of effort. Many report not receiving payment at all for extensive work, with some members sharing screenshots that reveal instances of no compensation for up to 20 hours of labor.
Joan Kinyua, the DLA president, highlights the absurdity: “The workers power all these technological advancements, but they’re paid peanuts and not even recognised.” She points out that while their work supports products like self-driving cars, many aren’t even available in Kenya.
The reality is tough. Data work often demands degrees, fast internet, and high-quality equipment, yet the pay does not reflect these requirements, forcing workers to invest heavily in education and technology only to earn a pittance.
Ephantus Kanyugi, the DLA vice-president, notes that employers feel little pressure to increase wages or resolve payment issues. With a vast pool of laborers ready to step in when others leave, there’s little incentive to change. Workers also contend with exploiting weak labor laws, which companies use to keep costs down, sidestepping fair contract practices.
Michael Geoffrey Abuyabo Asia, the DLA secretary, points out that contracts often lack clarity. Some contracts are filled with complex jargon, others cover mere days, and some workers face situations with no contracts at all. This inconsistency leaves them vulnerable, as jobs can suddenly disappear, accounts can be deactivated without warning, and workers find themselves without recourse.
Moreover, the absence of healthcare, pensions, or labor unions exacerbates their precarious situation, making it hard to plan for the future. Many workers also face “content moderation trauma,” dealing repeatedly with disturbing content. When they voice concerns, they often face retaliation.
One member shared her experience: “Any time we raise our voice… we are dismissed automatically.” She found the contracts unhelpful, as they typically offer little more than start and end dates. After trying to advocate for others, she lost her job.
To tackle these challenges, the DLA aims to secure mental health resources, legal support for pay disputes, professional development opportunities, and advocacy initiatives to spotlight the common issues within the industry. They also want to establish collective bargaining agreements with data companies, bridging the gap between workers and the tech giants.
The DLA is already collaborating with the African Content Moderators Union and Turkopticon and engaging with the Distributed AI Research Institute. They’re reaching out to Kenyan politicians and the Ministry of ICT, seeking to raise awareness about the realities of their work and how conditions can be improved for platform workers.