Shaping Africa's AI Future: Beyond the Kiddie Table

As we consider the broader implications of AI's deployment across Africa, a session from my recent 30th reunion at Harvard Business School (HBS) serves as a cautionary tale and a potent reminder of the challenges along the winding road to either AI Utopia or – God forbid – AI Dystopia, depending on how it all plays out for the vast majority of humanity.

Last fall, I was back in Boston, eager and excited to reconnect with lifelong friends from those two transformative, life-changing years at business school. Besides bonding, these reunions offer alumni a chance to dive back into the world of cutting-edge management thinking, brought to life by brilliant professors shaping tomorrow's leaders. Of the several sessions on AI that I attended, one in particular was an eye-opening experience.

The title sounded promising enough – "Building AI for All." However, it quickly became apparent that a more appropriate title would have been: "Building AI for 'All Left Out,'" the overwhelming majority of humanity variably referred to as the "Third World," "Emerging Markets," and, of more recent vintage, the "Global South." These are the voices that typically do not have a seat at the adult AI table, where the real AI action is. Instead, they have to watch the action from the sidelines – from the kiddie table, like Oliver Twist, begging for more.

I took a seat and surveyed the room, looking for the AI elite squad – the investment bankers, the management consultants, and, of course, the hard-charging Tech Bros of our class, all of whom had been present in full force at the standing-room-only AI session in the large state-of-the-art auditorium. Instead, those present and accounted for in the standard-sized lecture hall were the usual suspects – the "social entrepreneurs" and the "do-gooders." The AI elite, AKA the "do-wellers," were nowhere in sight. In the business world, "AI for All," "AI for Good," and similar euphemisms are typically code for charity or corporate social responsibility. In other words, not core to business, not mission critical.

The presentation aimed to showcase AI's potential across various regions of the Global South. As an African, I was immediately struck by the opening slide and the choice of imagery for each region – a high-tech hub for Asia, a similar scene for another emerging market, and, alas, for Africa, a lone farmer standing in a field in a remote village surrounded by a flock of chickens. The rusty ramshackle shed in the background completed the look – a scene straight out of central casting.

This image alone exposed a fundamental problem that Africa has to contend with within the context of AI and the broader Africa narrative:

• Who is telling the Africa story, and how does it relate to the ongoing discussion on AI?
• What is the risk that those who are currently leading AI research and development could inadvertently impose their pre-existing notions of Africa onto their work, reinforcing stereotypes?

Africa has 1.4 billion people. Not every African is poor, not every African is a farmer, and not every African raises chickens. Nor is it the image that comes to mind when I think of AI in Africa. Even more concerning are the potential consequences when these biases are perpetuated and amplified, which inevitably happens once they are encoded into AI algorithms.

Although the choice of imagery in the HBS presentation might, on the face of it, seem of minor concern, it exemplifies how deeply ingrained biases can lead to far more serious outcomes when applied to crucial areas like health, free speech, education, finance, and more. Left unchecked, these biases will hinder the full realization of AI’s benefits, ultimately undermining its potential to drive positive change and accelerate growth not just for Africa, but also for the majority of the global population – including many consumers in developed markets that are not part of the AI elite.

Here are a couple of examples of less benign AI bias.

Imagine an AI-powered heart disease diagnostic tool developed using data primarily from middle-aged white men. While highly accurate for this demographic, the AI system fails to accurately assess risk in women, leading to delayed diagnoses and increased complications. Its predictive accuracy is also lower in African American and Hispanic populations, resulting in underdiagnosis and inadequate preventive care. A well-intentioned AI invention has helped part of the population but has inadvertently caused harm to another.

Let's examine AI bias from another angle, our First Amendment Right – the right to speak, write, and share ideas and opinions without facing punishment from the "Nation State." But where is the same protection for "citizens" of the "Platform State," or, even worse, for citizens of Federated "Platform States" acting in concert.

Imagine a widely-used professional or social networking platform where users are encouraged to share insights and express opinions through diverse formats such as thought pieces, articles, and social media posts. However, the platform employs algorithms that systematically suppress positive stories from or about Africa, while amplifying narratives entrenched in stereotypes of poverty, corruption, and conflict. This approach isn't a mere oversight; rather, it is a deliberate choice to maintain a status quo that depicts Africa through a lens of negativity and limitation. By favoring stereotypical depictions over positive or even neutral news, the platform engages in a form of digital authoritarianism, deliberately distorting public perception and discourse.

Worse, this bias has real world consequences on the continent's economic development and growth. By persistently casting Africa in a negative light, these algorithms may scare away crucial investments, tourism, and partnerships that are pivotal for genuine economic and social advancement – in contrast to the conventional charity-driven models that are ineffective and perpetuate the "white-savior" narrative. The narrative shaped by these algorithms risks turning into a self-fulfilling prophecy. In this scenario, we find ourselves in a world where the African narrative is disproportionately influenced by algorithms curated by a select few – and not necessarily the few who truly understand or care about the continent.

Regulation is necessary but not sufficient.?Africa needs to be all-in embracing AI:

• Research
• Coding
• Product and platform development
• Testing
• Marketing AI solutions that are relevant to?the continent's?1.4 billion people

It is likely that Africa’s needs are not being met by existing offerings, as they are not a business priority for the few who currently hold the bulk of the world’s AI expertise, computing power, and financial resources.

Failure of Africa to seize the moment risks the continent being permanently relegated to a seat at the kiddie table, with no choice but to make do with a generic one-size-fits-all approach to AI. In AI, ?one size doesn’t even come close to fitting “all,” and is more likely to harm “many.”

Forward-looking AI leaders in the Global North seeking to tap into the African AI opportunity, should:

• Invest in local talent
• Collaborate with African researchers and developers
• Prioritize projects that address specific challenges faced by the continent

This will make AI development locally relevant, and consequently, in higher demand in a world that is becoming more African. This proactive approach cannot be done at the kiddie table, nor as charity or corporate social responsibility. It must make sound business sense – the good news for Africa, and for the rest of the world, is that it does.

Analysts estimate that AI could expand Africa's economy by $1.5 trillion by 2030 – about 50% of its current GDP. But to meet the moment and fully capitalize on this potential, Africa needs its own African AI elite squad present and accounted for in the AI sandboxes – and a seat at the grown-ups’ AI’s table. And with these kind of numbers, the global AI elite squad will show up to do business, not charity. They'll need Africa, the world's youngest and fastest growing demographic, to sustain the slowing growth in their mature markets.

And that, along with the benefits of first-mover advantage, is something they definitely teach everyone at Harvard Business School – including the investment bankers, the management consultants, and the hard-charging Tech Bros, all of whom I was delighted to reconnect with as always. We need them too, and I love them all – a special shout out to those from section "D."