“One needs to work five or six hours to complete what effectively amounts to an hour of real-time work, all to earn $2. In my point of view, it is digital slavery.” – Mutmain, a worker in Pakistan.

An article in Wired exposes how large tech companies are using workers in poorer countries, from Philippines to Colombia, to label data for training AI at wages amounting to a pittance.

Wired cites the case of Oskarina Fuentes who faced the economic turmoil of Venezuela in 2016 that unleashed soaring inflation and scarce job opportunities. Then, a friend introduced her to Appen, an Australian data services company that sought crowdsourced workers for tagging training data for AI algorithms. Fuentes and many educated Venezuelans turned to such platforms as a lifeline.

Though Appen’s clients include tech giants like Amazon (NASDAQ: AMZN), Facebook (NASDAQ: META), Google (NASDAQ: GOOGL), and Microsoft (NASDAQ: MSFT), much of this business is part of a hidden industry. Notably, the global data collection and labeling market size was estimated at $2.22 billion in 2022, and is projected to grow to $17.1 billion by 2030.

However, Fuentes’ journey on Appen is far from easy. She deals with power blackouts, earning as little as $1 or $2 on slow days. Fuentes works tirelessly from her bed, often starting tasks at 2 am due to Appen’s international client base.

This “microtasking” pattern is not unique to Venezuela. Workers from developing countries like east Africa, India, and the Philippines, as well as refugee camps, offer cheap labor on platforms like Appen. However, this new AI-driven industry is raising questions about “data colonialism,” because workers in Latin America label images that feed AI in wealthier nations. And this is leading to concerns about pathetically low wages, job fulfillment and evolving skill requirements.

The model benefits clients by managing a crowd instead of individuals, but workers like Fuentes want compensation for time spent waiting for tasks. They’re tied to their screens, with Mutmain, a worker from Pakistan, expressing frustration about low earnings and underestimation of their labor.

Fuentes and her fellow Venezuelan Appen workers are seeking recognition as more than mere tools. They yearn for employment status or unionization in an industry that currently lacks stability and worker rights.

The story of Appen and its global workforce sheds light on the complexities of AI data labeling and its impact on workers in developing regions.

Related Story: Kenyan Workers Subjected To Trauma When Labeling Explicit Content For OpenAI

Photo by orva studio on Unsplash