African AI researchers build language tools to counter Western tech dominance

African AI researchers are challenging Western dominance in artificial intelligence by developing tools that address the specific needs of African communities and languages. This work represents a significant push against the current AI landscape where major models primarily serve American and European interests while neglecting the linguistic and cultural diversity of Africa—perpetuating historical power imbalances in technology development and distribution.

The big picture: African researchers at the Distributed AI Research Institute (DAIR) are creating AI solutions focused on historically underserved communities rather than multinational corporations or Western users.

• Key researchers like Nyalleng Moorosi and Asmelash Teka Hadgu are addressing critical gaps in AI development for African languages and cultural contexts.
• Hadgu’s Lesan AI specifically targets low-resource languages like Amharic and Tigrinya, building high-quality datasets for millions of speakers typically ignored by major tech companies.

Key challenges: Leading AI systems like ChatGPT perform poorly on African languages, often producing meaningless output for languages such as Amharic and Tigrinya.

• Approximately 90% of AI model training data comes from Europe and North America, with Africa contributing only 4%, according to the Data Provenance Initiative.
• The unique tonal systems and oral traditions of many African languages remain poorly represented in Western AI models.

Why this matters: The current pattern of AI development mirrors colonial extraction dynamics, with big tech companies gathering local knowledge without proper compensation while prioritizing Western languages and contexts.

Where we go from here: Seven African nations have already developed initial AI governance frameworks, with more governments drafting national strategies to protect local interests and ensure community data sovereignty.