Data Enrichment Work and AI Labor in the Global South

Background:

The rapid development of AI, particularly generative AI, relies on millions of workers engaged in data enrichment — curating, labeling, and annotating data to train AI models. This emerging category of labor also includes content moderators, who are often tasked with defining the boundaries of permissible content on social media platforms, as well as other new forms of AI-supporting labor. These workers often work in precarious contract jobs or in pay-per-task environments in the Global South and are virtually invisible in the global economy. Little is known about the size, demographics, or working conditions of this workforce, especially in the Global South. Yet, the emergence of the AI labor economy has significant implications for developing regions, and it is not yet well understood how these workers contribute to the global AI industry and to the economies in which they live, and what measures might help to protect them.

Researchers and policymakers often conceptualize today’s AI industry as an AI value chain — an interconnected system including software and hardware vendors, cloud service providers, model developers and deployers, and firms offering AI-specific services in testing, auditing, and security. Data enrichment workers, and the companies that employ or contract with them, are key nodes in this network of actors, as they both rely on data and communications infrastructure for the global distribution of AI-supporting tasks and increase the value of data and models deployed and used by others.

USAID aims to engage one qualified research team to explore the data enrichment labor market in the Global South and its role in global AI value chains. Teams are strongly encouraged to develop unique proposals and research questions drawing on their expertise or experience.