Algorithmic colonialism: how cross-lingual deepfakes reproduce linguistic inequities in African media spaces

Abstract

This paper interrogates the intersection of cross-lingual deepfake technologies and linguistic inequities in African media. Through digital forensics, critical algorithm studies, and postcolonial theory, it demonstrates how synthetic speech models-predominantly trained on Global North linguistic corpora exacerbate the marginalization of African vernaculars. Drawing on case studies from Nigeria, Ghana, Mali, and South Africa, the paper shows how cross-lingual deepfakes reinforce colonial language hierarchies, threaten digital sovereignty, and undermine local media ecologies. The conclusion offers solutions for policy, governance, and technology to combat these growing inequalities.