Training Data Justice, Cultural Sovereignty, and Decolonial AI Governance: A Global South Perspective

Abstract

The rapid expansion of generative artificial intelligence is reshaping global systems of cultural production, knowledge circulation, and epistemic authority. Yet contemporary AI governance frameworks remain overwhelmingly dominated by Western technological paradigms, corporate platform logics, and data extraction models that systematically marginalise Global South cultural traditions and knowledge systems. This paper argues that AI governance must be understood not merely as a technical or regulatory challenge, but as a question of cultural sovereignty and epistemic justice. Drawing from emerging debates on training data governance, algorithmic bias, and digital cultural infrastructure, the paper introduces the concept of 'training data justice' as a normative and analytical framework for examining how cultural communities participate in, contest, and govern the extraction of cultural knowledge in the age of AI. Particular attention is given to how Islamic ethical traditions — including amanah (stewardship), maslahah (public interest), 'adl (justice), and shura (collective consultation) — may offer alternative foundations for AI governance beyond dominant market-driven or techno-solutionist models. Situating the analysis within Global South and postcolonial scholarship, the paper proposes governance frameworks grounded in cultural participation, equitable benefit-sharing, and the protection of epistemic diversity. It concludes that Global South intellectual traditions, including Islamic ethical thought, should serve as critical foundations for imagining more pluralistic, culturally grounded, and socially legitimate digital futures.