Masakhane – “We Build Together”

Abstract

This case study explores Masakhane, a grassroots, pan-African initiative building Natural Language Processing (NLP) systems in underserved African languages. Addressing the digital language divide, Masakhane demonstrates how AI and machine translation can be developed ethically, collaboratively, and in culturally grounded ways. The team focuses on creating open-source parallel corpora, translation tools, and named entity recognition for dozens of African languages. Their work combats the exclusionary practices of dominant tech platforms, which overwhelmingly support English and other European languages, contributing to language extinction and information inaccessibility. By enabling scientific research, public health communication, and everyday technologies to be available in native African languages, Masakhane promotes data sovereignty, decolonial AI, and technological self-determination. This case raises important questions around transparency, equity, and power in AI design, especially in contexts historically shaped by colonial knowledge systems. It also encourages students to reflect on the ethics of AI development, the consequences of language exclusion, and the value of participatory, multilingual innovation. Masakhane’s model offers a powerful example of how community-led, inclusive technologies can challenge global inequalities in digital infrastructure and epistemic representation.