Colonial structures in AI: a Latin American decolonial literature review of structural implications for marginalised communities in the Global South

Abstract

This article presents a critical literature analysis exploring the interconnections between artificial intelligence (AI), structural power relations, and the colonial matrix of domination, employing a Latin American decolonial theoretical framework. Through content analysis and a structured methodology involving pattern–category–concept identification, the study systematically maps key themes in existing literature, revealing how AI development perpetuates historical colonial structures and intensifies systemic inequalities across sociotechnical and epistemic domains. Two major structural dimensions of AI coloniality are identified: first, foundational colonial logics—including coloniality of power, knowledge, and being—and second, their contemporary reconfigurations as data colonialism, labour coloniality, and digital feudalism. These dynamics disproportionately affect Indigenous and marginalised communities in the Global South. In addition, the article critically addresses a significant epistemic gap identified in the reviewed literature—its limited empirical engagement with Indigenous epistemologies and alternative worldviews. Findings demonstrate that AI is not a neutral artefact but a sociotechnical system embedded within Eurocentric and capitalist logics. Addressing these issues requires confronting technological and economic power asymmetries. While acknowledging emerging resistance strategies, this article centres on structural reproduction rather than bottom-up responses, offering a conceptual framework that supports future critical, community-driven research on AI and decoloniality.