The development of artificial intelligence (AI) is often portrayed as the product of cutting-edge algorithms and engineering prowess. However, this narrative conceals the essential yet undervalued contributions of human data workers. These workers, who are instrumental in tasks such as dataset preparation, content moderation, and speech recognition, are rarely acknowledged as collaborators. Instead, their labor is minimized by AI companies that prioritize opacity and limit recognition of the human effort sustaining their technologies.
The creation of AI systems relies on a labor-intensive data pipeline that encompasses collecting, preparing, labeling, and verifying data. Human involvement is critical to provide contextual understanding and ensure ethical considerations. To manage the high costs of this work, many tech companies outsource to the Global South—regions like Africa, Southeast Asia, and Latin America—via digital labor platforms such as Appen, Amazon Mechanical Turk, and Scale AI (Remotasks).
A significant segment of this industry, the global data annotation market, is poised for rapid growth due to increasing demand for labeled datasets essential for training AI models. These datasets allow AI systems to discern patterns and correlations that form the basis of their predictions and classifications. However, this expansion also amplifies concerns over unethical labor practices, environmental impact, and bias, potentially exacerbating inequities in the developing world.
Data Annotation: A Modern Form of Exploitation
Researchers have drawn analogies between data annotation and historical colonialism, citing the extraction of cheap labor from developing countries to fuel the profits of global tech companies. In these largely unregulated environments, data workers face poor wages, precarious conditions, and a lack of job security.
For instance, a 2023 **TIME** investigation revealed that Kenyan workers labeling harmful content for OpenAI’s ChatGPT earned under $2 an hour while enduring grueling shifts evaluating hundreds of disturbing text passages. Similarly, reports from MIT Technology Review in 2022 highlighted exploitative practices faced by data annotators in Venezuela, Kenya, and North Africa, such as delayed payments, inconsistent workloads, and financial instability. These working conditions often neglect the psychological toll on workers tasked with confronting traumatic content.
In regions with limited employment opportunities, data work is often seen as an economic lifeline, offering a semblance of career advancement. Yet, systemic barriers such as high fees for cross-border payments, documentation hurdles, and limited access to banking services diminish workers’ earnings and opportunities. Persistent stereotypes about African labor, such as assumptions of low competence or lack of digital skills, further restrict access to the global tech industry for skilled professionals.
To navigate these challenges, some workers use tools like virtual private networks (VPNs) to obscure their geographic location, seeking fair opportunities. However, this system fosters competition among workers while reinforcing inequalities and leaving data workers undervalued in the global economy.
Cultural Impositions in Data Work
Economic exploitation is not the only issue; Western-centric frameworks dominate the AI industry, shaping how data is labeled and interpreted. Data workers are often forced to operate within rigid guidelines that fail to account for local cultural nuances. For instance, annotators may be asked to categorize race using U.S.-centric classifications that do not align with their cultural contexts.
This rigid approach disregards the depth of knowledge and cultural insight that data workers bring to their tasks. The result is an oversimplified labeling system that reflects Western biases rather than the complexities of diverse societies. By failing to recognize the cultural and intellectual labor involved in annotation, Western companies perpetuate a view of data work as mechanical and subordinate, undermining the expertise of workers from the Global South.