Towards a decolonial AI sovereignty model for indigenous educational settings

Abstract

This article proposes a critical reorientation of Generative Artificial Intelligence (GenAI) in Indigenous educational settings. We argue that GenAI, built on Western epistemologies, reproduces colonial hierarchies and risks what de Sousa Santos calls epistemicide. To confront these harms, we introduce the Decolonial AI Sovereignty (DAIS) model–a conceptual framework grounded in Indigenous data sovereignty, relational ethics, and decolonial refusal. To operationalise DAIS, we draw on seven illustrative case studies of Indigenous-led or Indigenous-engaged AI initiatives. Together, these cases demonstrate how Indigenous communities can actively shape technological futures that challenge extractive models and assert sovereign control over knowledge. We argue that Indigenous governance of GenAI must extend beyond data inclusion to encompass full epistemic jurisdiction, including the right to withhold participation in extractive digital systems. Unlike posthuman or critical pedagogical frameworks that emphasise AI as a co-creator, DAIS is unique in grounding AI’s role within Indigenous sovereignty, refusal, and ceremonial protocols, placing political and colonial conditions at the foundation rather than the margins. By bridging critical Indigenous theory, AI ethics, and educational futures, this article contributes to an emerging discourse on Indigenous AI sovereignty and offers a roadmap for reclaiming generative technologies as tools of resurgence, not assimilation.