Enabling Languages through Equitable Data and Scholarship

Since 2015, my work has centered on advancing Natural Language Processing (NLP) for African and other low-resource languages, where a critical barrier is not only the scarcity of datasets but also their accessibility and ownership. Too often, linguistic resources are locked away under restrictive terms, limiting their use for research, innovation, and community benefit. In this talk, I reflect on our efforts to open up the space of African language technologies by creating, curating, and sharing datasets under equitable licenses that prioritize community rights, scholarly reuse, and long-term sustainability. I will highlight how open data practices and responsible licensing frameworks enable new forms of collaboration across academia, grassroots communities, and industry—allowing us to grow resources in ways that respect linguistic heritage while fueling technological innovation. By embedding equity and openness into our data strategies, we move beyond static language repositories toward living, shared knowledge infrastructures. The future of NLP for African languages will be defined not only by model architectures but also by the choices we make around data governance, licensing, and scholarship. Through collective stewardship, we can ensure that language technologies remain inclusive, context-aware, and truly in service of the diverse communities they aim to represent.

Panel