Diversity in AI - African Languages SLR

Overview

A systematic literature review led by researchers spanning Ethiopia, Uganda, and Nigeria is mapping the gaps in African language representation in AI, and surfacing the gender biases that prevent women from equally benefiting from AI-powered tools in healthcare, agriculture, and business.

Despite Africa's extraordinary linguistic diversity, the vast majority of AI systems are built on data that ignores local languages and underrepresents women effectively rendering large portions of the population invisible to the technologies shaping their lives. This digital suppression limits access to critical services.

The project employs a rigorous systematic literature review to identify where African languages are missing from AI datasets and where gender bias distorts outcomes. From this evidence base, the team is developing targeted interventions: advocating for women to lead AI development, designing inclusive data collection strategies, and producing policy recommendations that governments across the continent can act on.

The research is already informing inclusive AI policy frameworks across multiple African nations. By validating the presence of underrepresented groups in both written and spoken data and ensuring women are active designers, not just end users, of AI tools the project offers a replicable model for building AI ecosystems that work for everyone.