• Initiative builds data for voice, translation, speech technologies

• Effort aims to expand digital access, support inclusive AI development

Burkina Faso’s Ministry of Digital Transition, Posts and Electronic Communications on Tuesday launched a workshop to develop the linguistic foundations needed for artificial intelligence solutions. Discussions focused on building resources for AI models, particularly in four local languages: Mooré, Dioula, Fulfuldé and Gulmancema.

These resources are essential for applications such as voice recognition, machine translation and speech synthesis, which are key technologies for expanding access to digital services in a linguistically diverse country.

“By integrating national languages into emerging technologies, Burkina Faso is demonstrating its commitment to building inclusive, accessible artificial intelligence tailored to its sociocultural context. This workshop aligns with the 12 flagship projects of the digital transformation roadmap to 2030, particularly the initiative focused on AI for all Burkinabè,” the ministry said.

The initiative reflects a broader continental trend. Several African countries have recently launched programs to integrate local languages into AI technologies. In Benin, the “JaimeMaLangue” project aims to give local languages a digital presence. In Nigeria, authorities recently launched N-ATLAS v1, an open-source language model developed with Meta. The system supports several local languages, including Yoruba, Hausa and Igbo, as well as Nigerian English, with the aim of reducing linguistic barriers to technology adoption.

International technology firms are also increasing their focus on the issue. Google’s Africa-focused AI unit recently launched the Wazal project to improve support for African languages in AI models, notably through data collection and translation systems. The database covers 21 African languages, including Hausa, Yoruba, Luganda, Acholi, Swahili, Igbo and Fula.

In Burkina Faso, authorities are relying on locally produced data to build a more sovereign AI ecosystem. The project is being carried out with public institutions and research centres to assemble reliable corpora essential for training AI models.

Samira Njoya