This month Togo launched a crowdsourcing drive that asks residents to submit texts, audio recordings and translations in order to teach artificial-intelligence systems the country’s fifty indigenous languages. The effort seeks to fill a gap that leaves most West African tongues absent from today’s large language models, and it positions the government as a data-source rather than a passive consumer of foreign AI services.
Policymakers and technology leaders across the continent have flagged the scarcity of African languages in mainstream models as a strategic vulnerability. When a model cannot process the linguistic nuances of a nation’s citizens, its utility for public-service delivery, education and commerce diminishes sharply. The Togo initiative illustrates a broader consensus that language inclusion is a prerequisite for any AI-driven public-sector transformation.
Research firm Gartner projects global AI investment to reach roughly two trillion dollars this year, as major firms pour resources into computing power and infrastructure. Yet the World Bank notes that Africa, home to eighteen percent of the world’s people, hosts only six-tenths of one percent of worldwide data-centre capacity, and merely five percent of that infrastructure is configured for advanced AI workloads. The disparity highlights a structural bottleneck that limits the continent’s ability to host large-scale models.
Presently Africa’s contribution to the AI surge comes largely through the export of minerals such as copper and cobalt, which power the chips that run machine-learning algorithms. The African Development Bank estimates that widespread, inclusive AI adoption could lift the continent’s gross domestic product by up to one trillion dollars by 2035. Realising that potential, however, depends on building the energy and connectivity foundations required for AI to move beyond a niche tool and become a driver of productivity in agriculture, markets and public administration.
In its latest Economic Update, the World Bank advises African economies to prioritize adapting existing AI technologies rather than attempting to develop frontier models from scratch, a path it deems costly given limited fiscal space. Togo’s crowdsourced language project exemplifies this approach: by enriching a model with locally sourced data, the government can make the system immediately relevant for delivering services such as health alerts, agricultural advice and civic information.
The overarching message for the continent is that dominance in AI model size is less critical than ensuring the technology serves the daily needs of its people. African nations must move beyond roles as mere raw-material suppliers or passive consumers of off-the-shelf AI products, and instead cultivate ecosystems where AI can generate the largest economic gains. Such a shift requires coordinated investment in data infrastructure, energy reliability and local talent development.