“We need to focus on catching up.”

One of Africa’s most consequential conversations in the current moment is, what does a truly African Artificial Intelligence future look like? Depending on who you ask, AI in Africa promises to unlock new levels of productivity across healthcare, agriculture, education, and several other sectors. 

But as Abejide Ade-Ibijola, a professor of Artificial Intelligence (AI) and Applications at the Johannesburg Business School (JBS), tells TechCabal, that future comes with a warning. “The excitement around AI is real, but we risk building technologies that widen the very inequalities we are trying to solve,” he said in an interview.

While Africa only accounts for 2.5% of the global AI market, the continent’s innovations are expected to make an economic impact, potentially contributing $2.9 trillion by 2030.  Countries like Kenya, Nigeria, and South Africa are leading the charge, applying AI to local problems, including interpreting indigenous languages and boosting crop yields. But Ade-Ibijola says progress must be rooted in realism.

“The adoption of AI will likely create a two-tiered reality,” he said. Those in urban centers like Johannesburg and Kigali will enjoy the benefits of AI-powered mobile phones, virtual reality, and other cutting-edge technologies. Meanwhile, those in rural areas, often lacking basic infrastructure like electricity, risk being left behind.”

Without infrastructure, affordability, and intent, he believes that the benefits of AI will remain concentrated among the privileged few. Consideration must also be made about building ethical, unbiased AI systems as well as datasets that reflect Africa’s cultural and historical nuances.

This interview has been edited for length and clarity.

How do you envision AI shaping industries, societies, and daily life in the coming decades in Africa? 

Artificial intelligence is here to stay, and its impact will be profound across numerous sectors. In healthcare, for instance, AI has the potential to revolutionise diagnostics and treatment. However, we must address the issue of affordability. While AI-driven surgeries might offer greater precision, we risk widening the gap between those who can afford such advanced care and those who cannot. This existing socio-economic divide is a critical factor we need to consider as we integrate AI.  

In agriculture, drones equipped with sensors and irrigation systems can optimise crop management, reducing reliance on pesticides and manual labor. We can also foresee robotic systems handling tasks like tilling and planting, boosting efficiency and productivity.  

Education is another area ripe for AI integration. AI-powered tools can generate personalised learning plans, provide language translations for complex concepts, and even facilitate self-paced learning. While robotics in classrooms is still nascent, particularly at the university level, we could see physical AI-driven facilitators in the future.  

Manufacturing is already heavily reliant on AI, as seen in automated production lines in Japan and China. In Africa, we are beginning to see similar trends, with automated ordering systems in fast-food chains and food delivery robots in hotels. These advancements, while exciting, raise concerns about job displacement. If petrol stations and retail stores become fully automated, what happens to the workforce?

The adoption of AI will likely create a two-tiered reality. Those in urban centers like Johannesburg and Kigali will enjoy the benefits of AI-powered mobile phones, virtual reality, and other cutting-edge technologies. Meanwhile, those in rural areas, often lacking basic infrastructure like electricity, risk being left behind. This widening gap between the “haves” and “have-nots” is an inevitable consequence we must address proactively.  

The excitement surrounding AI is undeniable, but we must be mindful of its societal implications. We need to focus on equitable access and address the potential for increased inequality. As we embrace these transformative technologies, we must ensure that their benefits are shared by all, not just a privileged few.

What cultural or societal values should guide AI’s development in Africa?

Cultural and societal values that should guide AI development in Africa must focus on inclusivity and accessibility. Many of our people, especially in underprivileged rural areas, are still living in conditions akin to the Second Industrial Revolution. For example, some still use charcoal irons and lack access to electricity, television, or the internet.

When we talk about AI, we often speak from an urban perspective, but these technologies must reach those in remote areas to improve their daily lives. For instance, imagine a medical diagnosis app powered by AI that operates on a phone. If we could provide phones to villages and ensure the app works in local languages, it could act as a “doctor on their phone.” This app could offer preliminary diagnoses, recommend seeing a nearby doctor, or even suggest safe indigenous remedies when appropriate.

To make this feasible, we would need solutions like solar energy or lithium batteries to power these devices. This approach ensures that AI serves everyone, not just urban populations.

We must be intentional about addressing inequalities. This means bringing technology directly to underserved communities – installing solar panels, providing access to devices, or even deploying mobile tech hubs with laptops for young people to learn and interact with advanced technologies.

These initiatives can inspire innovation and help children imagine their future roles in tech development. Continuous education for both children and their parents is vital. If building schools isn’t immediately possible, we can organize open days or trips to expose them to new ideas and opportunities.

Our data must feed into our own AI systems rather than perpetuate global disparities. By aligning AI with African values like Ubuntu and ensuring equitable access, we can create technologies that reflect our cultural heritage while promoting social responsibility.

What are some of the challenges of building unbiased, ethical AI systems in Africa? 

AI models are fundamentally built upon data. Without vast datasets, or big data, AI cannot function. The machine learns patterns and makes decisions based on this data. However, as humans, we are inherently flawed, and our data reflects that.

Consider the unfiltered content on platforms like X. It is filled with profanity, political criticism, and personal insults. If an AI is trained on this data, it will inevitably learn and replicate these behaviors. This is why discussions around ethical AI and unbiased AI are so vital. We must be intentional about how we approach data and training.

There are two primary strategies: Firstly, we can curate and sanitise the data before training the AI. Secondly, we can train the AI on unfiltered data but provide it with explicit guidelines on ethical behavior. For example, we can feed it examples of both polite and impolite language, allowing it to discern appropriate communication.

It is essential to acknowledge that AI will always reflect the biases of its creators and the data it is trained on. We must confront the issue of harmful bias, such as AI misidentifying people. In such cases, we need to implement rules that prevent these erroneous classifications. Instead of resorting to offensive labels, the AI should simply indicate “face not recognised. We can also refine the AI’s output at the inference layer. If the AI misclassifies a cat as a dog, we can manually correct it, teaching it to recognise the correct patterns.

If we want AI to be relevant to African contexts – recognising African artifacts, clothing, or medicinal plants – we need to provide it with extensive datasets containing these elements. Unfortunately, these datasets are currently scarce. We must actively build repositories that encompass African stories, perspectives, and knowledge systems. This includes our medicinal practices, educational systems, and even traditional numbering systems. By incorporating this rich cultural data, we can develop AI that is truly contextualised for Africa.

How should governments regulate AI more strictly while encouraging innovation without constraints?

Historically, excessive regulation has often stifled progress, but a complete lack of oversight can also lead to unforeseen and potentially dangerous consequences. The challenge is that policymakers often lack the technical expertise to fully understand the capabilities and potential dangers of these technologies.

Effective AI regulation requires a collaborative approach, bringing together government officials, AI experts, industry professionals, and manufacturers. This roundtable discussion would allow for the creation of informed and practical policies.

African policies must be communicated to the public, educating people about the potential benefits and risks of AI and related technologies, outlining the ‘dos and don’ts’ in a clear and accessible manner.

We must remain vigilant, adjusting regulations as needed to address emerging challenges and prevent loopholes. If regulations are too restrictive, they will hinder innovation; if they are too lenient, they will leave us vulnerable to abuse.

AI demands careful regulation, and Africa must proactively engage in these discussions, ensuring that policies are developed collaboratively and comprehensively while anticipating potential threats and creating robust frameworks to safeguard our societies.

What opportunities exist for African institutions to lead in global AI research? 

Our priority should be to measure up to existing standards. If you look at what countries like China, United States are doing, it is clear we need to focus on catching up rather than leading. For instance, China’s advancements with tools like DeepSeek show the level of competition we are facing globally.

Africa has historically struggled with leadership in many areas, and attempting to lead in AI now might be unrealistic. Instead, we should ask whether AI can help us leapfrog development challenges and improve our standing from a lower level to something more competitive.

First, we need investment in infrastructure – servers, and training programs. Second, our educational systems require a serious overhaul. They are outdated and overly theoretical; we need practical skills training starting as early as primary school. Third, societal discipline is crucial. Issues like drug addiction and the negative influence of social media must be addressed. For example, correctional facilities could focus on rehabilitation through technology training. Prisoners could learn skills like programming and contribute to technological innovations while serving their sentences. This approach would reduce costs while offering redemption opportunities. Our youth need guidance and discipline to prevent long-term societal issues.