Same Stereotypes, Different Term? Understanding the “Global South” in AI Ethics

In recent years, the impact of artificial intelligence (AI) systems on global politics has received growing attention. However, much work remains to understand the connotations, usage, and contradictions of the term "Global South" within AI ethics. In this paper, we complicate existing conceptualizations of the "Global South" by interviewing 20 AI ethics scholars and practitioners who have engaged with global politics. We find a range of definitions for the "Global South," from geographic regions to places with colonial legacies. We identify four key limitations of the term "Global South": (1) its tendency to homogenize countries with diverse political power and technological development, (2) its use as a proxy for underdeveloped in AI innovation and regulation, (3) its implication of technological illiteracy, and (4) its suggestion of unidirectional influence with AI technology marketed from centers to peripheries. The "Global South" can thus perpetuate an imperial gaze, similar to the terms "Third World" and "developing countries." Despite these limitations, many participants we interviewed feel pressured to use the term "Global South" due to research and funding structures that reinforce the United States as a center of resources and knowledge production. Ultimately, rather than simply adopting another term that may inherit the same harmful stereotypes, we encourage scholars and practitioners to: (a) ground their work in specific regions and communities, (b) leverage particular analytical frameworks, such as anti-caste theories, when drawing connections across communities, and (c) make deeper changes to build alternative funding structures and foster solidarity among countries at the peripheries.