What happens when the technologies shaping our future cannot understand the languages we speak? In Togo, a new national initiative is addressing this question by bringing local languages into the development of artificial intelligence. Launched under the leadership of Minister Cina Lawson, the programme aims to make AI more inclusive, strengthen the country’s technological capabilities and open new possibilities for accessing essential public services in the languages people use every day.
Artificial intelligence can translate documents, answer questions, generate text and increasingly interact with people through spoken language. Yet many of the world’s languages remain poorly represented in the data used to develop these technologies.
For millions of Africans, this raises an important question: will the growing use of AI make information, education and public services more accessible, or create new barriers for people whose languages are not adequately represented in digital systems?
In Togo, where approximately 50 local languages are spoken, the government is taking steps to address this challenge.
On 2 October 2026, Cina Lawson, Minister of Public Service Efficiency and Digital Transformation, officially launched the National Local Languages Challenge, the first stage of a three-year programme known as Voix du Togo (Voice of Togo).
Led by the ministry and implemented through Togo AI Lab, the initiative aims to develop the linguistic data and artificial intelligence models needed for digital systems to understand, process and eventually communicate in Togo’s national languages.
The ambition extends beyond translation. It is about enabling people to interact with digital technologies in their own languages, while building the foundations for AI applications that reflect the country’s linguistic and cultural diversity.
Teaching AI to understand local languages
Artificial intelligence systems learn from data. To recognise speech, translate text or respond meaningfully in a particular language, they need access to recordings, written material and examples of how that language is used.
For languages with limited representation in existing digital datasets, developing these capabilities requires a concerted effort to collect and organise linguistic resources.
This is where the National Local Languages Challenge comes in.
Through an online participation platform, Togolese citizens are invited to record spoken sentences, transcribe audio recordings and translate written texts. The contributions will help create the datasets needed to train AI models adapted to the country’s languages.
The programme begins with three languages: Ewe, Kabiyè and Moba. According to the Togolese government, the intention is subsequently to extend the initiative to the country’s other national languages.
For each language, the initial collection targets include 50 hours of validated audio recordings, 6,000 pairs of translated sentences and contributions from at least 500 different speakers.
The collected material will be checked by linguists and the communities concerned before being used to train AI models. The programme also envisages competitions involving students, researchers, developers and start-ups, opening opportunities for Togolese talent to contribute directly to the development of the technology.
Recognising that not everyone has reliable access to smartphones or the internet, the initiative also provides for field teams to gather contributions from people who might otherwise be excluded.
When language becomes a barrier to public services
The implications of this work extend well beyond the technology itself.
Consider someone seeking medical advice who finds it difficult to explain their symptoms in French, or a citizen who needs to complete an administrative procedure but is more comfortable communicating in a local language.
As public services increasingly move online, the languages supported by digital systems can determine who is able to use them independently and effectively.
Voix du Togo aims to create the foundations for future applications that could allow people to communicate with public administrations, healthcare providers and other services in the languages they know best.
The government has identified potential uses including spoken interaction with administrative services, support for communication in healthcare and translation in judicial proceedings.
These applications remain longer-term objectives rather than services already operating at scale. Nevertheless, they illustrate how advances in language technology could address practical challenges and make digital public services more accessible.
In September, Togo AI Lab held a public testing session in Lomé to gather feedback on the Voix du Togo platform, assessing its usability and how well prospective users understood its features.
Such testing is an important step towards ensuring that new technologies respond to people’s needs rather than simply demonstrating what is technically possible.
More than translation: language, knowledge and identity
The significance of the initiative also lies in what language represents.
Languages carry histories, cultural identities, knowledge, traditions and ways of understanding the world. They are not simply interchangeable means of communicating information.
When a language is poorly represented in digital technologies, the issue extends beyond the inconvenience of translation. It raises questions about whose knowledge is available, whose experiences shape technological development and whose cultural perspectives are reflected in the systems people increasingly rely upon.
For African languages, these questions are particularly relevant as artificial intelligence becomes more influential in education, research, business and public administration.
Creating linguistic resources can help make languages more accessible in digital environments while supporting their documentation and preservation.
It can also open new possibilities for locally developed educational content, language-learning tools, research and other applications that respond to the linguistic realities of African societies.
The broader question is not simply whether AI can be made to speak African languages, but whether African knowledge and cultural identities will help shape the development of AI itself.
Building Africa’s own AI capabilities
Togo’s initiative also reflects a wider ambition to strengthen national expertise in artificial intelligence.
Rather than relying exclusively on language technologies developed elsewhere, the programme seeks to create resources that Togolese researchers, developers and entrepreneurs can use to build applications responding to local needs.
The United Nations Development Programme (UNDP), which is supporting the initiative, is also assisting Togo AI Lab in developing computing infrastructure for AI research and model training.
At the programme’s launch, Minister Cina Lawson described language models as an essential part of the public infrastructure needed for the digital age.
This approach recognises that the foundations of technological innovation extend beyond devices and connectivity. Data, research expertise, computing capabilities and institutions that can support their effective use are equally important.
The initiative could therefore contribute not only to linguistic inclusion but also to the development of local technical expertise, research opportunities and new applications.
Its longer-term impact will depend on the quality of the linguistic data collected, the performance of the resulting models, their responsible development and the extent to which they can be translated into useful, accessible services.
Questions of data protection, informed consent and community participation will also be important as the programme develops, particularly when collecting and using people’s voices and linguistic contributions.
Africa’s knowledge, Africa’s languages, Africa’s future
The initiative connects directly with the theme of eLearning Africa 2027, Africa’s Knowledge Power: From Human Capital to Shared Prosperity, which will be explored in Lomé, Togo, from 26–28 May 2027.
The conference will examine how African knowledge, research, skills and innovation can generate economic and social value, and how the benefits of technological progress can be shared more widely.
The development of AI systems that understand African languages provides a particularly compelling example of these questions in practice.
It brings together linguistic and cultural knowledge, scientific research, technological capability, public service innovation and opportunities for locally developed solutions.
It also challenges a fundamental assumption about digital transformation: that societies must adapt themselves to the languages and structures of existing technologies.
What if technology could instead be developed to reflect the languages, knowledge and priorities of the people it is intended to serve?
Through Voix du Togo, the country is beginning to explore what that could mean in practice.
The question is not only whether artificial intelligence will learn to speak Africa’s languages. It is whether African people, knowledge and ingenuity will have a defining voice in shaping its future.
Sources:
UNDP Togo, 5 October 2026: Et si l’intelligence artificielle comprenait enfin nos langues locales ?
