AI for Africa by Africans
The need for Africa-centric AI tools in local languages
Last month, I wrote an article about African Artificial Intelligence (AAI), I wanted to start a conversation about how one size does not fit all.‘One-size-fits-all’ solutions are rarely the answer; in fact, they usually leave everyone disappointed. The feedback on the first article about AAI prompted me to think about the challenges of AI and its implementation to benefit Africans.
A significant issue with AI, and something that mathematician and data scientist Cathy O'Neill wrote in her book Weapons of Math Destruction - How Big Data Increases Inequality and Threatens Democracy, is the built-in bias of AI. We live in the age of the algorithm. Increasingly, the decisions that affect our lives, where we go to school, whether we get a car loan, and how much we pay for health insurance are being made not by humans but by AI models.
In theory, this should lead to greater fairness: Everyone is judged according to the same rules, and bias is eliminated. But, as we know, every individual and their circumstances are different. One of the biggest problems is that these 'models' can cause significant harm and reinforce discrimination.
For example, if a poor student can't get a loan because a lending model deems him too risky (by where they live, they are then cut off from the kind of education that could pull them out of poverty, and a vicious spiral ensues.
African challenges
One of the major challenges we face here in Africa when using and integrating AI is that it is not made with Africans in mind or our diverse languages. As Large Language Models (LLMs) dominate AI development, we must consider whether we should rely solely on these tools or create African AI tools and models tailored to the continent's unique needs.
Africa is home to over 2 000 distinct languages and thousands of dialects spoken across 54 countries. However, Africa's languages are woefully underrepresented in mainstream AI tools. LLMs are powerful, but their capabilities are shaped by their training data. Who is training them in Swahili, Yoruba, Afrikaans, or Oshiwambo?
Our dream as a Namibian ICT company is to see Africa-specific AI tools prioritising local languages, such as the African Artificial Intelligence (AAI) tool. Creating and training tools can bridge the gap between technology and those currently excluded due to language limitations. A farmer in rural Kenya or Namibia should be able to ask an AI tool about agricultural best practices in their native tongue and receive comprehensible, culturally relevant advice. Just imagine how transformative that could be for the communities.
Practical benefits
Beyond the practical benefits of an AAI tool, there are economic benefits as well. Creating an AAI represents an opportunity for economic growth and innovation. Developing AAI capabilities would stimulate local tech industries, create jobs, and build expertise across Africa. It also reduces reliance on external providers and protects the continent's digital sovereignty. African nations, depending solely on imported AI tools can exacerbate dependency and economic imbalance. Investing in developing Africa-specific AI tools empowers local innovators and ensures that technological advancements align with Africa's priorities.
If we are to develop and bridge the digital divide, creating AAI tools in local languages is not without its challenges. Building and training AI models require vast amounts of data and investment. Efforts would also need to be made to digitize and curate linguistic resources while addressing ethical concerns such as consent and privacy.
Critical factor
Funding is another critical factor; developing AI tools is capital-intensive, and African nations must prioritize investment in technology alongside other pressing needs. Collaboration between governments, private sector players, and international organizations should be the way forward.
Africa could power an AI revolution and empower itself. Its linguistic and cultural diversity, combined with its unique socio-economic landscape, necessitates the development of AI tools tailored to local needs. The potential benefits of AAI far outweigh the obstacles. We can make it happen as a continent.
* Kehad Snydewel is the Managing Director of Green Enterprise Solutions.
** Opinion pieces and letters by the public do not necessarily reflect the opinion of the editorial team. The editors reserve the right to abridge original texts. All newspapers of Network Media Hub adhere to the Code of Ethics for Namibian Media, a code established jointly with the Media Ombudsman.