The future of AI: Why Google is betting big on Africa

19 Feb 2019

Image: © atosan/

The continent of Africa is one of the most diverse and youngest populations in the world. And it is a cradle of promising talent when it comes to AI and machine learning.

At a Google Making AI event in Amsterdam late last year, the potential of Africa as a hotbed for AI talent was impressed on me.

Having only ever been to the north of the continent, what I don’t know about the many countries, cultures and peoples of Africa sometimes frightens me. We ought to know more.

‘If AI can make our lives better in Africa, then all the debate about jobs being lost is not the correct one’

Almost reading my mind, Moustapha Cissé, research scientist and head of Google AI in Accra, Ghana, said levelly to a crowded room: “The potential for having a positive impact is as high here, if not higher.”

Last year, Google announced its first artificial intelligence research centre in Accra, focusing initially on fairness in machine learning (ML), interpretability of ML models and the use of AI for medical diagnosis and treatment. The unexpected move by the internet giant signalled a vibrant future through technology for Africa, a place that has for too long exploited by other nations but now holds the keys to solving today’s and tomorrow’s problems, especially through AI.

Cissé, who grew up in Senegal, studied maths and physics at university. He fell in love with AI almost like a hobby, tweaking chess algorithms between classes. “I got a taste for it,” he said.

“We are building a new AI centre in Ghana to solve challenges on the ground but also to continue working on foundational and cutting-edge research. The reason this is important is because when a new technology is experiencing rapid progress, it is important to bring as diverse a set of perspectives to ensure that the benefits are available to everybody.”

Crucially, what Google has realised is that Africa has a large, untapped pool of tech talent that is already showing a passion for AI. And when you hear Cissé talk about AI in Africa, it is not just about Ghana but across the entire continent where a love of maths and AI can be fostered.

A new narrative for Africa

Young man in blue shirt sits on a stage surrounded by foliage.

Moustapha Cissé, research scientist and head of Google AI in Accra, Ghana. Image: John Kennedy/Silicon Republic

“Africa has the youngest, fastest-growing populations on the planet and there is an enthusiasm for AI,” Cissé explained. “The local challenge is how can we use AI for better impact. That’s the goal of our centre in Accra. In the past 10 years the technology has experienced rapid growth. The situation is changing. You have more and more graduate programmes in Africa bringing AI to students. There is still a lot to do, such as deep learning and there are communities where people come together.”

Cissé cited the work of the African Institute for Mathematical Science, a pan-African network of centres of excellence focused on enabling talented students to become innovators and drive the continent’s economic self-sufficiency.

“There is masses of work being done in machine intelligence, helped by investments by Google, and many of the leading scientists come to Kigali [Rwanda] to teach in the programmes. It is an exciting time to be doing machine learning in Africa and we want to help strengthen the ecosystem.”

The key is to establish research facilities all across the continent to bring opportunities to people.

A fresh perspective

Cissé said that many of the technologies coming out of Africa are influenced by the local environment and problems on the ground. “They reflect important challenges locally. It is about bringing a fresh perspective. When we build a new technology it is AI by Africa rather than just for Africa. Everything is falling into place to make this happen.”

Early successes include language applications as well as machine learning for an early warning system for crop disease that is being built in Kimpala, in the Democratic Republic of the Congo.

Another example is Harambee, a social enterprise in South Africa that uses TensorFlow to interact with more than 1m young people. Harambee is credited with helping 50,000 young people find employment using precise geographical attributes and preferential behavioural metrics.

An interesting characteristic of Africa’s AI revolution is that rather than fear ‘brain drain’ through emigration, Cissé believes that the opposite will happen.

“Brain drain? I don’t believe in it, especially in science being used where it is most useful. People tend to go back to where they started – many Africans actually do this. We should create conditions for people to go back and contribute where their skills are most impactful. Science is universal and as soon as people are contributing, it is irrespective of where they are based.”

How AI could drive the destiny of Africa

While much has been made recently of China vying against Europe for influence in Africa, Cissé is adamant that the future of Africa will be determined by the African people themselves.

“Africa doesn’t belong to the Europeans or Chinese. We want to collaborate with countries with mutually beneficial results. It is also up to Africans to explore other partnerships and it is up to us as leaders to ensure technology is beneficial for the people.”

Cissé said that there are pockets of AI sprouting up in countries such as Nigeria. From a start-ups perspective, Google has established a Launchpad Africa accelerator in Lagos, Nigeria where there are more than 60 start-ups taking part, as well as similar hubs in Ghana, Kenya and South Africa. He said that the key is to establish more collaborations to tackle specific problems.

“One way is to use AI to improve healthcare in Africa, where we have the youngest and fastest-growing population in the world. It is a challenge but a huge opportunity for all of us.”

Cissé said that computer hardware and internet connectivity can be problems but this is also where Google comes in with powerful open-source computing resources such as TensorFlow.

“Making GPUs available for free and available to everyone is a very important step towards democratising the use of AI.

“We have by far the youngest population in the world. I’m 33 and I’m older than 70pc of the population. There is a tradition of excellence in applied math, and while the talent pool is not similar to north America or Europe, it is being improved. Encouraging more bright young Africans to do research in ML will improve things.

“There are many countries in Africa that are forward-looking. Across the whole continent, something is happening. There is a vibrant science ecosystem that is growing and growing.

“From a socio-economic perspective, tech can be a huge enabler if we manage to create the conditions for its wider adoption and put technology into the hands of people with problems to solve as well as investing in entrepreneurs.”

Alluding to Africa’s colonial past, Cissé said: “The whole narrative will change. The overwhelming majority of the population of Africa were born after independence. Economists call AI a general-purpose technology and everybody agrees it is a revolution. We are in a period of time where it is accelerating so fast and transforming so many industries. Within 20 years of the first automobiles arriving in New York there were no more horse cabs. If jobs are lost then new ones will be created by the fact that the technology exists.”

And if AI has the potential to save lives and drive prosperity in Africa, Cissé believes the narrative of fear about the technology in western economies needs to change too.

“We believe there are huge avenues of collaboration between the tech and humans,” he said. “If AI can make our lives better in Africa, then all the debate about jobs being lost is not the correct one.”

John Kennedy is a journalist who served as editor of Silicon Republic for 17 years