ALEX TSADO & BETTY WAIREGI

More practical applications. Scaled talent training. Supercomputer access. These are the next 3 catalysts needed to spur Africa forward in the global race for artificial intelligence (AI). Africa, indeed, has come a long way from where it started in 2017 when most of the strong initiatives started. 

In fact, AI in the world actually started to move out of being a lab dream in just 2012 when a researcher named Alex Krizhevsky in Canada entered into a competition to use machine learning to predict different images and used the Nvidia GPU, different from the usual Intel chips that the remaining participants were using. What he found out was that to train his model with Nvidia GPU, he was able to leverage parallelized programming to run up to 1000 parts of the code at the same time. With this, instead of waiting many days for the model to train, it took only a few hours, finally making it practical to build AI solutions. That completely revolutionized the world of artificial intelligence and sent engineers and governments around the western and eastern world to invest heavily in applications of AI even for everyday life.

By 2017, you will find that the largest companies in the world and the largest governments in the world are powered by artificial intelligence. Think about Google Microsoft, Alibaba, Baidu, and the United States government, all with many applications of AI using it to win in competitive markets. At this time, the work of AI in Africa was not a coordinated one. It was one where there were innovators scattered around the continent working on small projects. The rest of the world had the bias that Africa was not participating in AI: global reports and indexes on AI would draw maps of AI in the world and leave Africa blank. Yes, they were incorrect in that there was a decent amount of AI happening in Africa. But they were correct in that a lot less was happening than needed to happen. It’s incredibly important for Africans, and more broadly speaking, every group of people around the world to participate in the development of artificial intelligence tools. It’s a lot more critical than any other tools of the past because there tends to be a direct correlation between the representation of the group that is designed for an AI tool and how effective the tool is going to be working for different groups of people. Said differently, as a Black person anywhere in the world that will use an AI tool, you really need the team that built that tool to include Black people that had the influence to ensure that the tool is tested on people like you 

By 2017 several Africans in the diaspora who work at the largest companies recognized this as they could see the challenges and dangers of the future head-on and started to create communities, and nonprofit organizations, to spark grassroots movements to get more Black and African people into artificial intelligence. Notable organizations include Black in AI which was formed in the United States of America, the Deep Learning Indaba founded in South Africa, and Alliance for Africa’s intelligence (Alliance4ai) which we founded in California with an office in South Africa. We founded Alliance4ai to be the centre of information and network for everything happening in AI in Africa. We created a database of over 100 notable Africans in AI, over 100 training centers, and communities in Africa and put together a policy group that advised several African government agencies on AI strategies. These included the presidential office of South Africa and the Africa Union’s Smart Africa Initiative which designed Africa’s first AI blueprint. Working with the grassroots, we created a program to help university students create AI clubs that would allow them to learn topics that the universities do not teach. The topics span three areas. Firstly, it was an introduction to the technical side of artificial intelligence. Secondly, was a strong embedding of AI in the history of African people, showing the strong contributions of Africans to ancient and modern AI. The third piece is around building workplace skills that are necessary to promote one’s work and get the funding or employment one needs after building technical skills. 

Today, Alliance4ai has trained hundreds of students who have graduated from the program as alumni to do things from getting scholarships to Masters and PhD programs to getting jobs at top companies in Europe and the United States, to founding startup companies that have won grants and awards. The most special testimony came from a 17-year-old girl in Ghana that said,

“Before my first Alliance4ai club meeting, I thought artificial intelligence was only meant for White and Chinese people. But after seeing the 100 leaders listed on the Alliance4ai website, I’m convinced I can play a role and will study AI in college.”

We are proud of the progress the continent has made, and that the rest of the world is finally beginning to notice what Africa is doing in AI as shown by the much higher investments in African AI companies, the opening of engineering offices by the largest companies in the world and the invitation of African innovators and inventors to international AI conferences like the UN’s AI for the good summit. 

The current challenge is that many of the initial attempts to build AI have been in spaces where they are targeting markets that cannot afford their services. For example, many diagnostic tools in agriculture have been marketed to subsistence farmers who have not been able to afford what’s being built. It’s now the work of African practitioners to find better ways of building solutions that work economically for the audiences they’re building for. 

On the government side, it’s most crucial to note that data is the new oil, and we must stop making the same mistake we made with oil. With oil, our governments only built processes to extract crude oil, export abroad, and import more expensive finished products. We are missing refineries. Today similarly to what happened with oil, our governments are working with NGOs to make it easy to extract African data. They are failing to build access to Supercomputers so the smartest African minds can get access to building quality finished products that work best for Africa. Failure to do this will leave the continent exploited again like in the last hundreds of years.