AI for Africa: How to build the future with limited data and local talent

Artificial Intelligence (AI) experts have said data, partnerships and investment in people are essential for boosting AI growth in African countries, including Nigeria.

At a recent Africa AI Impact Summit, the experts said by investing in local talent and partnerships, Africa can position itself to capitalise on the $1.5 trillion opportunity that lies ahead in AI.

“In building AI in Africa, the problem isn’t lack of data, but rather accessibility to the data and a lack of understanding of how to use it,” Tobi Olatunji, founder of Intron Health, a healthtech platform, said.

“In the African context, even though in healthcare a lot of the data is paper based, the data is there. There are billions of pages of health records in paper form. So you have to creatively think about how to get that data? How do I transform it? How do I use it? versus there’s no data. I think collaboration has definitely helped people discover where to find it. And how to use it,” he said

Bonaventure Dossou, a research scientist at Lelapa, an Africa-centric AI research lab, said less data can sometimes be more. “There are emerging data-efficient modelling approaches that can be used to exploit the most out of limited data availability.”

He also emphasised the importance of developing new models adapted to African languages and contexts, rather than relying on pre-trained models trained on massive datasets of high-resource languages.

Read also  Shell to use new AI technology in deep-sea oil exploration

Dossou added that Africans should focus on developing smart and creative ways to build AI models with limited data, rather than chasing after acquiring more and more data.

According to the UN Economic Commission for Africa, African data ecosystems are at “nascent stages of the African data revolution and the private sector is increasingly becoming a critical and dynamic player within African data ecosystem

The United Nations in 2021 projected that Africa could expand its economy by a staggering $1.5 trillion, by capturing just 10 percent of the speedily growing AI market, set to reach $15.7 trillion by 2030.

On people and collaboration Abdoulaye Diack, program manager of Google AI, said AI for Africa cannot be built without having people from underrepresented groups deeply involved in the research itself.

“We need more startups that work on fundamental research. Language is key to a lot of the challenges on the impact that we can have. If we can solve this challenge on the language side, we can have a tremendous impact,” he said.

He said to achieve great impact in the AI space; Africans should focus on people and partner with local universities, “especially the public universities, which I think are really struggling, by supporting faculties, local labs and AI communities, that can really push AI for impact.