Epistemic Justice in AI

Abstract Despite appearing neutral, digital technologies are shaped by histories of power, exclusion, and epistemic dominance. Algorithmic systems that mediate access to welfare, finance, health, and civic life often embed dominant worldviews while marginalizing others, particularly across the Majority World. These exclusions go beyond data gaps or design flaws; they reflect deep assumptions about what constitutes knowledge, progress, and legitimacy. This article examines whose knowledge underpins digital futures. It draws on feminist, decolonial, and anthropological perspectives to interrogate how colonial legacies and extractive development models continue to influence contemporary data infrastructures. Rather than centering bias mitigation, the article adopts epistemic justice as a foundational framework that foregrounds plural ways of knowing, being, and governing. Across various contexts, communities are not only resisting exclusion but are also actively shaping alternative digital futures grounded in care, consent, and collective accountability. From Indigenous data sovereignty to feminist practices of decision-making, Majority World approaches disrupt dominant design logics and expand the possibilities of digital governance. The article proposes the Four Fs for epistemic plurality: foreground, fracture, futures, and fellowship. These principles offer a shift from designing for the world to designing with it, recognizing its full diversity and complexity. In doing so, the article contributes to wider debates on digital rights, knowledge justice, and the political imagination required to create inclusive, plural, and relational technological futures.