October 24

Timothy Kotin

Making AI Work for Africa: Why Google Must Go Local with its New Research Center in Ghana

In the 21st century, perhaps no technology has garnered more attention than artificial intelligence (AI). As AI develops systems that perform tasks which normally require human thought, it has led to a spectrum of emotions, from optimism to fear. While scientists, ethicists and policy makers grapple with concerns over the rise of super-intelligent computers that will supplant humans, AI developers explore the potential that the technology holds to solve countless problems – including in developing countries. Given that optimistic Africa watchers have proclaimed that the future is African, citing its growing connectivity, economic growth and demographic scale, there is no better place to harness the sweeping impact of AI than Africa.

The continent is on the cusp of leveraging its untapped potential for AI. In June, Google announced that it would open an AI research center in Accra, Ghana later this year. AI has the potential to bring major positive change to Africa, stimulating new economic activity in some of the sectors that need it most, such as agriculture, healthcare and finance – industries that have historically suffered from infrastructure inadequacies and a lack of skilled labor. But in order to make meaningful contributions and avoid merely creating an African outpost for its pre-existing research efforts, Google must invest in regional talent to develop locally relevant AI solutions.


The Missing Element in Industrialization Plans

Many African governments, encouraged by international financial institutions, have focused intensely on industrialization as a means to economic progress, attempting to replicate East Asian successes by channeling energy and resources into the manufacturing sector. Governments from Ethiopia to South Africa have put out major industrialization plans that hardly mention automation and AI. One possible reason for this is the fact that, while the opportunities that AI could bring to African countries are enormous, the risks of getting it wrong are even greater.

Research has shown that over 80 percent of the manufacturing work that will soon be automatable is in the developing world. The world’s least developed countries will be hit the hardest by an abrupt technological transition—losing a whopping 52 percent of jobs in Kenya, 46 percent in Nigeria, and 41 percent in South Africa. As African countries largely missed out on the wide-scale manufacturing that lifted East Asia to economic advancement, these countries must avoid relying solely on that model, particularly when AI could offer similar – if not even greater – gains.


Making Google’s AI Efforts More Impactful

It’s true that Google is taking baby steps to support efforts to teach AI courses in Africa, such as its machine learning course at the African Institute of Mathematical Sciences (in partnership with Facebook). But it should focus its efforts on training African AI and data science talent in a manner that can scale for years to come. Despite the handful of universities in Africa with AI programs, the continent has far too few universities in general to accommodate its young population. Existing universities consistently underemphasize technology, with computer science departments across Sub-Saharan Africa chronically undersized and underfunded, further hampering the continent’s ability to adapt to a more automated world. Google picked Ghana for the AI research center in part because of its relatively strong private university network, but private universities alone are too limited in size to produce enough technical talent to keep up with the industry’s potential. Therefore, it will be crucial for Google to diligently look outward—tapping talent elsewhere on the continent and funding its own developer programs.

If a large talent pool can be created, the AI center in Accra can make its parents company’s mission of democratizing AI a reality. In addition to sourcing local and regional talent, Google will have to work closely with local enterprises and developers in order truly understand what types of AI solutions are needed for an African context. AI has already been harnessed by developers in Africa to circumvent existing inefficiencies, for example to analyze and interpret speech in rare dialects or to assess and disburse loans to customers that lack a credit history. Google should throw its weight behind efforts to address challenges like these, capitalizing on its open source tools like TensorFlow to collaborate with the local developer community to create relevant technologies. If it does, the AI center in Accra will have a better chance of delivering on Google’s promise to benefit 10 million Africans through its regional work, and in ways that truly deliver transformational impact.


Moving Beyond Symbolism

Given growing recognition of the AI skills shortage and issues of bias and discrimination in trained AI systems, there is much work to be done for African countries to truly harness the potential of this powerful technology. Some regional developers have already seen success in applying AI to local problems, but the talent pool is still shallow, and funding to scale can be sparse.

Google is taking an important step toward addressing these problems by opening this research center, but it must not rest on the symbolism of its investment. It should focus efforts on engaging with the community, through training and technology transfer, investment and mentoring, and aggressive sourcing of regional talent, if it seeks to meaningfully contribute to Africa’s technological transition.

It has been suggested that AI, along with other technological and demographic trends, could transform society “ten times faster and at 300 times the scale, or roughly 3000 times the impact” of the Industrial Revolution. Africa, perhaps more than any other continent, could benefit from this powerful revolution, and Google’s new research center and initiatives could play a major role in the transformation.


Timothy Kotin is a Ghanaian entrepreneur and the co-founder of SuperFluid Labs, a data analytics platform based in Ghana and Kenya.


Photo: A street scene in Ghana. Credit: eva_lutter.




artificial intelligence, global development, manufacturing