Navigating the Pitfalls and Potential of AI in Africa

Every weekday, around midday, Alice anchors "The Brief Bulletin," a daily news show for online publication CITEZW. Reporting on pertinent political and social developments in Zimbabwe and neighboring Southern African countries, the show cuts between visuals of Alice reading the news and footage relevant to the topics being reported on. Alice’s voice and diction are clear and impeccable, with the news read in a stoic tone that doesn’t betray any bias.

Within seconds of watching "The Brief Bulletin" on any given day, viewers are informed that Alice is an artificial intelligence presenter, the first of its kind on the continent. Alice has anchored what is now the premier daily show CITEZW for more than a handful of months, reporting on Zimbabwe’s presidential election last August. Alice also hosts a weekly show alongside CITEZW founder Zenzele Ndebele, where popular personalities from across multiple fields are interviewed.

“That’s one of the best case uses of generative AI that I’ve been shown in Africa,” Michael Akpo, a freelance data scientist tells OkayAfrica. Akpo, a self-professed machine learning obsessive who’s based between California and Kigali adds that Alice is primarily a tool that has been well-trained and is immediately impactful despite some noticeable limitations. “I think this is one of the examples that should balance out some of the doom and gloom around AI, generative AI especially.”

Excitement and paranoia

In recent years, global conversations surrounding AI have shifted to reflect the paranoia many have towards technology that mimics human intelligence. Previously, they were mainly centered around the tantalizing potential of AI as a driving force in the Fourth Industrial Revolution. To an extent, the excitement of what AI can achieve hasn’t dwindled, but there are fears that such an advanced level of technology can upend things as we know them to be.

There’s still a wide gap between the rapid rate of AI technological advancement and uses in better-developed parts of the world and on the continent. According to experts and observers, that gap is a positive that can lead to tangible results.

According to PwC, a professional services company, AI is set to contribute a potential $15.7 trillion to the global economy by 2030, with $1.3 trillion in economic growth for Africa. The African Union’s New Partnership for Africa’s Development (NEPAD) estimates that AI could help double GDP growth rate by 2035. Governments and other organizations are also already working on how AI can help achieve the United Nations Sustainable Development Goals.

AI use across Africa

There are already dozens of AI technology applications across Africa, some of them solving real-life problems. Nigeria-based Intiron Health provides speech-to-text AI chat tools that understand over 200 African accents. In South Africa, computer scientists built a machine learning model that analyzes the segregation of housing opportunities based on race. AI and technology solutions company Omdena has employed machine learning and AI models to create solutions for pest control in Somalia and equipment management in Tanzania.

In 2023, Nigerian rap artist, producer and entertainment multihyphenate Eclipse Nkasi released the first African AI-powered album, Infinite Echoes, made using several AI software, including the ultra-popular ChatGPT. The project also featured Mya Blue, an AI singer. Speaking with CNN shortly after the release, Nkasi shared that he believes AI is a collaborative tool that can assist creativity, an optimistic stance that steers clear of the understandable cynicism that has followed some exploitive use, like the release of deep-fake songs by artists, deceased and alive.

“The best case uses of AI will enhance our standards of living,” Akpo says. However, Africa still has a lot of ground to cover in order to catch up with the other regions of the world, and even more work will be needed to ensure AI is integrated into all facets of African society. For these goals to be achieved, there’s a resource and knowledge gap that needs to be filled.

Ethical issues, challenges and opportunities

AI is a broad field that requires advanced equipment, extensive infrastructure, a lot of locally contextualized data and a high level of human expertise. This has led to the intellectual property issues, associated with generative AI especially, whereby AI platforms have been sued for using artists’ works without license, credits, or payment.

AI builds on pre-existing information and practices that should be properly digitized, but that in itself is an issue that still needs to be worked on. For example, many hospitals and clinics across many countries in Africa still store patient records in physical paper files and folders, making it difficult for the work being done in AI healthcare to be widely applied.

“When people’s work is in files and folders, as opposed to being on computers, it becomes much more challenging to automate,” systems analyst Chiagoziem Onyekwena tells OkayAfrica. Also speaking from the perspective of the threat AI poses to blue collar jobs, Onyekwena says that the cost of implementing AI solutions has to compete with the wages being paid to low skilled workers in Africa.

“How much disruption can you create if people aren’t being paid a decent wage?" Onyekwena said. "Also, a lot of roles haven’t been digitized, so the intermediary step before you get to AI isn’t there yet. Some of the roles that are ripe for digitization on the continent, the type of salaries that these jobs command are so low that they are already protected from the disruption.”

Regardless, that information and digitization gap still needs to be filled if AI is to make a tangible impact in Africa, which is where Onyekwena believes it would be most beneficial to Africans at home and in the diaspora. Whether it’s machine learning, deep learning or generative AI, data is a necessity, which means humans are needed to train models, from inputting data and labeling images, to writing code and reviewing processes.

Investing in AI in Africa

From Smart Africa to Google’s AI lab in Ghana, several incubator programs are doing their part in advancing AI in Africa. Last year, a coalition of the U.K. and Canadian governments, along with the Bill & Melinda Gates Foundation, pledged an $80 million donation to several AI projects in Africa, including dedicated funding of labs. Rwanda’s ICT minister Paul Ingabire said of the initiative, that it “empowers African countries to become producers, not just consumers, in the AI revolution.”

Regulating AI in Africa

Amidst the push to help develop Africa’s AI technology capacity, the African Union (AU) is in advanced stages of handing down a policy framework to regulate AI. In late February, the African Union Development Agency (AUDA) published a policy draft that outlines proposals on the confines of ethical AI use. It follows the EU’s new, pioneering AI act, the first comprehensive act of its kind that’s set to become law in Europe’s member countries.

The white paper and roadmap review document for the AU’s proposed regulation policies divide the framework into five pillars, including Human Capital Development and AI Economy. The breakdown for each pillar recognizes the forward-facing potentials of AI, and repeatedly mentions the importance of investing in AI startups. At the same time, it asserts that “AU Member States should collaboratively establish unified legal systems that clearly define AI ethics, offering protection and binding obligations across the continent.” While this is a checking measure, it can also be interpreted as a way for governments to steer AI-related projects by labs and companies.

The AU’s regulation proposal is divisive because, while many believe in the virtue of proactively regulating a technology that can be used to violate human rights, there are those who think regulating the technology could hamper the full scope of how AI technology can be used in Africa. Also, while the EU has the power to enforce its act, the AU’s act will mainly serve as a framework for member countries to base their own individual policies on.

“The regulations we’re seeing from the EU, they aren’t intended to stifle AI development. They are meant to protect people from the Machiavellian tendencies of tech companies,” Onyekwena says. “I don’t think that is where policymakers or governments should spend the most time. In the African context, we have bigger challenges to try to overcome as far as maturity of the work force and maturity of the data. I think [African] governments should focus on creating progressive policies that allow us to build from the low bases we have, to do more interesting things in the global AI conversations. There will still be spaces to protect the people of Africa.”