Lessons From the Margins: Contextualizing, Reimagining, and Hacking Generative AI in Global South

Harvard Data Science Review

Team: Diana Mosquera, Francisco Gallegos, Andrés Domínguez

GENERATIVE AIGLOBAL SOUTHPOLITICAL ECONOMYVALUE CHAIN OF AI

Our new article is now available in the Harvard Data Science Review, where we analyze the impact of generative AI in the Global South and the crucial role that these regions play in the global value chain of this technology.

The dominant narrative around generative AI tends to present it as a neutral and universal tool, capable of solving global problems from a "technologically objective" logic. However, this vision conceals the material, social, and ecological costs that make its existence possible. From the extraction of minerals like lithium, cobalt, or nickel, many of them in Global South countries, to the outsourcing of precarious digital labor for tasks such as content moderation or data labeling, the infrastructure that sustains Generative AI is deeply marked by colonial and extractive power relations. All of this occurs while economic benefits, patents, and prestige are concentrated in the technological centers of the Global North. But faced with this panorama, other forms of action also exist.

Our article makes visible how researchers, technologists, activists, and communities from the Global South not only denounce these injustices, but also actively build alternatives. From practices of reappropriation and redesign of technologies, to the development of linguistic models that prioritize underrepresented languages and AI tools developed with a feminist perspective, there is simultaneously a diverse movement that reimagines what artificial intelligence can and should be. These initiatives show that it is possible to think about an AI that is not based on scale at all costs, nor on centralized control of computational resources. An AI that does not ignore the plurality of ways of knowing, living, and building technologies—smaller, open, contextual, built from care, epistemic justice, and ecological sustainability. Many of these experiences invite us to explore other routes: those that place communities, their autonomy, and their contexts at the center.

This research was written in collaboration with Andrés Domínguez Hernández, researcher at the Alan Turing Institute, with whom we share the commitment to think about artificial intelligence from a critical, situated, and global perspective. We believe that the Global South is not a passive recipient of technology, but an active space of reflection, innovation, and resistance.