AI Development in Ghanaian Languages led by Paul Azunre - Unyielding Challenges Encountered Throughout the Process
In the heart of West Africa, Ghana is making strides in artificial intelligence (AI) development, particularly in the realm of local languages. Dr. Paul Azunre, co-founder of Ghana Natural Language Processing (Ghana NLP), is spearheading this effort, focusing on a community-driven approach to sourcing data.
However, the journey is not without its hurdles. Infrastructure deficits, including broadband connectivity and a reliable power supply, especially in rural areas, pose significant challenges. Moreover, the lack of local data, particularly in indigenous languages, makes it difficult to train AI models that accurately understand or generate these languages.
The shortage of local AI talent skilled in linguistic technologies is another issue. To address this, initiatives like the One Million Coders Programme and AI Community Centers, such as Ghana’s AiAfrica Labs and Google’s AI Community Center in Accra, are being proposed. These centres aim to foster collaboration and training, with a focus on rural areas and marginalized groups, including language accessibility.
Government policy also plays a crucial role. Ghana’s Draft National Artificial Intelligence Strategy and Data Harmonisation Act are designed to promote data sovereignty and boost research aligned with local realities, which is essential for local language AI.
Despite these challenges, the impact of Ghana NLP's work has been far-reaching. Doctors, teachers, and startups across Africa are using their tools. Dr. Azunre's team has managed to develop the world's first machine translation and speech systems for several Ghanaian languages, as well as Kikuyu from Kenya, with fewer resources than their international counterparts.
Dr. Azunre urges African AI developers to look beyond the Silicon Valley hype and build more efficient, smaller offline models. He argues that some tech giants promote open source ideals as a way to extract value from under-resourced regions.
The future of AI in Ghana is promising. The current government plans to review the AI strategy developed by the previous government and engage with industry to ensure Ghana's status as the AI hub in Africa. On January 31, 2025, Ghana's Minister of Communications, Digital Technology and Innovation, Samuel Nartey George, made a promise to transform Ghana into the center and hub of Artificial Intelligence (AI) within four years.
Dr. Azunre is happy to see that the work Ghana NLP is doing speaks for itself. However, he states that his team has not benefited financially for the past ten years. Foreign corporations, foundations, and other entities are often surprised when the subject of compensation for African language contributions is brought up.
Translating local languages in Africa with AI is complex, and Ghana's unique linguistic diversity complicates matters further. Dr. Azunre's team at GhanaNLP is already working toward building more efficient, smaller offline models. With continued support and resources, Ghana's AI ecosystem has the potential to set an example scalable across Africa.
[References] [1] [https://www.researchgate.net/publication/346039782_The_State_of_Artificial_Intelligence_in_Africa] [2] [https://www.sciencedirect.com/science/article/pii/S235234092100118X] [3] [https://www.nature.com/articles/d41586-021-00425-7] [4] [https://www.ghanaweb.com/GhanaHomePage/NewsArchive/Ghana-to-be-AI-hub-in-Africa-in-four-years-Communications-Minister-875801] [5] [https://www.theafricareport.com/113301/ghana-ai-development-priorities-infrastructure-talent-data-governance-policy-and-local-language-focus/]
Artificial intelligence (AI) developments in Ghana, particularly in local languages, are encountering infrastructure challenges, such as broadband connectivity and power supply, particularly in rural areas, as well as a lack of local data in indigenous languages. Dr. Azunre, focusing on a community-driven approach, is working towards building more efficient, smaller offline AI models to address the shortage of local AI talent and translate local languages with AI, setting an example scalable across Africa.
