Togo has launched a national initiative to develop artificial intelligence tools capable of understanding local languages, aiming to improve access to public services for citizens who do not speak French.
The three-year programme, known as Voix du Togo (Voice of Togo), was launched in early October by Cina Lawson, Minister of Public Service Efficiency and Digital Transformation. It will collect audio recordings, written transcriptions and translated texts to build datasets for training AI models in Togolese languages.
Through an online platform, citizens can contribute by recording spoken sentences, transcribing audio and translating written material. The collected data will be reviewed and validated before being used to develop language-processing technologies.
The initial phase will focus on Ewe, Kabiyè and Moba, with plans to include additional national languages as the programme progresses. For each language, the initiative aims to collect 50 hours of validated audio recordings, 6,000 pairs of translated sentences and contributions from at least 500 speakers.
The resulting datasets are expected to support AI applications in healthcare, justice and government administration. The technology could enable citizens to interact with public institutions in their preferred languages and help patients describe symptoms without having to communicate in French.
Voix du Togo forms part of the government’s broader strategy to strengthen local AI development and deployment. The Ministry of Public Service Efficiency and Digital Transformation is working with Togo AI Lab, which focuses on applying AI across sectors including agriculture, healthcare, education and the environment.
In September 2026, Togo AI Lab conducted public tests of the platform in Lomé to assess usability and gather feedback ahead of its wider rollout.
By developing language resources based on local speech and written material, the initiative aims to create AI tools that reflect Togo’s linguistic diversity and make digital public services more accessible.
