Every Algorithm Carries Bias: Africa’s Urgent Need for Inclusive AI
The author argues that global AI development—and its concentrated funding and infrastructure—risks deepening economic and representational gaps for Africa. They call for internally driven innovation, better data representation to avoid “dat
As AI continues to rise, so does the growing gap between developed and developing nations. Today, we’re witnessing billions of dollars being pumped into AI research and development in the Western world. While many of those experiments may fail, the few that succeed will inevitably reshape the global economy and deepen the influence of those governments and corporations across the world. These AI products will not just power economies—they will own consumer attention, behaviour, and even belief systems.
A lot of this funding is coming from bold, overzealous investors and venture capitalists, not because they have money to burn, but because they understand this is a winner-takes-all game. And unfortunately, the side that’s losing doesn’t just lose economically—they lose the power to shape their destiny.
But beyond the economic value of AI lies something more subtle and dangerous: the concentration of power. Just like the internet gave rise to global tech empires like Meta and Alphabet, we’re starting to see the same centralized control happening with AI infrastructure. Today, tools like ChatGPT and similar platforms are winning—not necessarily because they’re perfect, but because the infrastructure, funding, and user base are stacked in their favour.
This imbalance is already having massive consequences for regions like Africa, where investors are understandably risk-averse and highly profit-driven. It’s not their fault—the capital available here is limited, and people are fighting tooth and nail to protect what little they have. But this mindset, though valid, is suffocating innovation. It’s nearly impossible to build anything meaningful here if it doesn’t promise immediate financial returns.
One major implication of this imbalance is what I call "data bias blindness." Today, AI systems still struggle to accurately generate images of Black Nigerians or Africans. They often confuse us with Black Indians or other vaguely similar groups. This isn’t about racism—it’s about absence. We’ve failed to feed the machines enough of ourselves for them to recognize us, even in basic visuals. So, how can we expect to be recognized by global institutions, governments, or companies when even machines can’t place us properly?
This tells us one thing: we need to show up better. We must begin to innovate beyond just monetary gain. We need to stop waiting for fellowships, grants, or external saviours to move forward. Innovation in Africa must be internally driven. Yes, we’ve had bad actors who took money and ran, building nothing. But these are a few among a sea of genuinely passionate builders. And let’s be honest—other countries have their frauds too; the difference is, they just don’t scream about it like we do.
In the age of AI, we must rethink what’s possible. We must return to first principles—how can AI solve real problems in the economy and humanity? This cannot be about academic papers or adding projects to portfolios. Ironically, Africans may need AI the most, but we’re also the slowest to embrace it.
That’s why, at Genesis Intelligence, we are committed to building inclusive AI systems. We’re not just building tech—we’re building people, and people build societies. Our long-term vision is to place Africa at the center of AI and innovation, just like fintech has redefined economic systems here. We believe that if we leverage AI properly in Africa, it could significantly raise our GDP. But only if we stop fearing failure, build useful products, and scale what works—instead of chasing profit from day one.
I’m not calling out African investors, not the government, not our institutions, and not our youth. This isn’t about blame. It’s about responsibility. We all have a role to play.
So, stand up and get to work.
LETTERS
Like this letter? Like, Share or Subscribe for future version
No algorithm. No manufactured frequency. Unsubscribe at any time.