Reframing "Linguistic Diversity"
Dominant narratives suggest AI will democratize language, enable translation at scale, and open new markets. In reality, AI consumes data and flattens culture, leaving us to ask whose interests are truly being served. Alix reframes “linguistic diversity” with Chenai Chair, exploring how communities can govern their language data and prioritize preservation over productization.
CHENAI CHAIR – LINGUISTIC DIVERSITY
Chenai Chair leads the Masakhane African Languages Hub. She founded My Data Rights (Africa) and has led initiatives at Mozilla Foundation (Africa Innovation Mradi, Common Voice: African Languages), the World Wide Web Foundation’s Gender and Digital Rights flagship, and Research ICT Africa’s youth and gender research. She brings feminist perspectives to data governance and AI ethics.
In this conversation, Chair addresses the current discourse around linguistic diversity in AI. She argues that the push to build new, varied linguistic datasets is being promoted by players with vested interests to access new markets in the Majority World. But this approach comes with political risks and sidelines long-standing actors in the ecosystem. In its place, Chair calls for a collaborative, bottom-up approach that centers communities. Drawing lessons from Masakhane’s experiences, Chair offers insights on how to include communities in such endeavors—and what to do when they refuse. The way forward, for Chair, lies in building on existing initiatives and sharing resources.
"We are moving towards another level of data which is language, and language is also personal identity…We often ask ourselves, who actually does own the language? Who gives us the permission?"