Generative Artificial Intelligence and Indigenous Knowledge Systems in the Global South: A Systematic Review on Data Colonialism and Epistemic Justice

Abstract

Introduction: The accelerated expansion of Generative Artificial Intelligence (GenAI) is reshaping contemporary ecologies of knowledge production, validation, and circulation, particularly in contexts marked by historical inequalities such as Latin America and the Global South. When these technologies are deployed over Indigenous Knowledge Systems (IKS), they not only mediate processes of cultural preservation and documentation but also intervene in broader disputes over data coloniality, epistemic sovereignty, and the risks of digital epistemicide. In this scenario, the relationship between GenAI and Indigenous knowledge becomes a strategic space for questioning who controls data infrastructures, under which frameworks Indigenous worlds are represented, and what conditions allow these knowledge systems to continue naming the world on their own terms.

Objectives: To examine how applications of Generative Artificial Intelligence (GenAI) within Indigenous Knowledge Systems (IKS) can contribute to epistemic and cultural justice in Latin America and the Global South.

Methods: A systematic review was conducted following the PRISMA protocol, with searches in Scopus and Web of Science between 2023 and 2025, complemented by Scopus AI–assisted retrieval. After screening, full-text reading, and applying inclusion/exclusion criteria, a final corpus of 15 studies was obtained. Thematic analysis was guided by four central categories: GenAI, Indigenous knowledge, Epistemologies of the South, and epistemic/cultural justice.

Results: The analysis identified four interrelated axes: A) Preservation and revitalization of knowledge; B) Culturally situated educational and evaluative tools; C) Data governance frameworks and knowledge cocreation; and D) Decolonial imagination of creativity. Together, these axes reveal a field that is both promising and shaped by dynamics of extractivism and epistemic dispossession.

Conclusions: GenAI can only be considered an ally of epistemic justice when anchored in data sovereignty, community cogovernance, and slow temporalities of deliberation, which stand in tension with the technoscientific acceleration associated with the shock of the future. It is concluded that horizons such as Slow AI and sociotechnical infrastructures designed from and with Indigenous peoples are necessary conditions for contesting data coloniality and reauthorizing their knowledge systems in the Global South.