Ghebru Bethel
| 2026, 23
| pp.1~58
| number of Cited : 0
This article examines how artificial intelligence (AI) can support the preservation and revitalization of endangered African languages and scripts through the lens of multicultural education and digital language justice. Bringing Banks's theory of multicultural education into conversation with May's analysis of linguistic minoritization, together with culturally sustaining pedagogy and community data sovereignty, it develops four analytical dimensions spanning recognition, rights, community agency, and intergenerational learning. It proposes a function-based typology of AI applications, examines Ongota, Yaakunte/Yaaku, Nǀuu, and African Ajami manuscript traditions as illustrative preservation scenarios, and treats African script technology as a cross-cutting infrastructural layer rather than an equivalent case.
The article further develops a seven-phase implementation roadmap and identifies five interrelated risk domains: extractive data governance, exposure of restricted knowledge, AI error and epistemic distortion, infrastructural unsustainability, and unequal institutional power. It argues that AI-supported preservation should be evaluated not only through technical performance or data volume, but through its contributions to linguistic recognition, educational access, community authority, intergenerational learning, and cultural continuity.