Artificial Intelligence Development in Ghanaian Languages, led by Paul Azunre, hasn't been a smooth ride
In the heart of West Africa, Ghana is making strides in the field of Artificial Intelligence (AI) and Natural Language Processing (NLP), with Dr. Paul Azunre, co-founder of Ghana NLP, leading the charge. Dr. Azunre and his team are working tirelessly to build NLP systems for Ghanaian languages, a task fraught with challenges due to systemic inequities.
However, the financial landscape for open source AI research and development in Africa is gradually improving. Major investments, such as Google's $37 million funding package, are explicitly supporting African languages and AI research capacity across the continent. This funding not only emphasizes the development of digital skills but also strengthens the ecosystem for sustainable open source AI research and applications in various sectors, including agriculture and food security.
Initiatives like the AI Collaborative for Food Security, funded through a $25 million Google.org grant, exemplify how financial resources can enable AI solutions addressing locally relevant challenges. These investments not only enhance food systems but also build digital infrastructure and research capacity for broader AI innovation, including language processing.
Collaborative research programs, such as the South African-German AI research initiative, provide multiyear funding and partnerships that encourage sustainable AI development involving local universities and public research institutions with an interdisciplinary approach. These collaborations help tackle societal problems, including those dependent on language technologies and cultural contexts.
Efforts to preserve indigenous knowledge and local languages via AI are gaining momentum. Research highlights that AI integration offers unique opportunities for language preservation, cultural revitalization, and sustainable development, which hinge on long-term financial and infrastructural support.
Dr. Azunre's team at GhanaNLP has made significant strides, developing the world's first machine translation and speech systems for languages like Twi, Ewe, Ga, Dagbani, Gurene, and even Kikuyu from Kenya, with fewer resources than their international counterparts. However, Ghana NLP's work still faces challenges, with local institutional support being lacking, according to Dr. Azunre.
One of the key challenges in adopting AI in Africa is using it to process local languages. Dr. Azunre's approach to sourcing data for Ghana NLP is community-driven, focusing on working with native speakers. High-quality data, according to Dr. Azunre, results in smaller volumes compared to what corporations do when they automatically scrape large volumes of data.
Despite the progress, much of the advancement in AI still fails to speak to the majority of African populations. Dr. Azunre sees opportunities in building practical, commercial tools-apps that solve urgent local problems. He encourages African AI developers to think beyond the Silicon Valley hype and focus on building more efficient, smaller offline models.
In a promising development, Ghana's Minister of Communications, Digital Technology and Innovation, Samuel Nartey George, made a promise to make Ghana the center and hub of Artificial Intelligence (AI) within four years. The current government in Ghana plans to engage with the industry to ensure that Ghana becomes the hub of AI on the continent.
Looking ahead, Dr. Azunre is not disclosing more details about what's next for GhanaNLP, but he encourages staying tuned. Despite not benefiting financially from Ghana NLP's work for the past ten years, Dr. Azunre remains optimistic about the future of AI in Africa. He urges foreign corporations, foundations, and other entities to acknowledge the value of African languages and compensate contributors fairly.
References:
- TechCabal
- Quartz Africa
- COMPASS AI/ML
- ResearchGate
- South African-German AI Research Initiative
Technology advancements in the AI and artificial intelligence sector are not only empowering the development of digital skills in Ghana but also strengthening the ecosystem for sustainable research and applications in various sectors, such as agriculture and food security. For instance, Dr. Azunre's team at GhanaNLP, with fewer resources than international counterparts, has developed machine translation and speech systems for local languages like Twi, Ewe, Ga, Dagbani, Gurene, and even Kikuyu from Kenya.
