On Thursday in Diani, under the Mashariki Conference banner, I found myself in a room that felt both important and unfinished.
Important because of who was there—security and policy leaders from more than 74 countries. Unfinished because of what still needs to be said, plainly and without hesitation.
The gathering itself, driven by the good work of Noordin Haji, the National Intelligence Service Director and his team, is becoming something worth paying attention to.
It is not just another conference. It is slowly shaping into a space where countries can talk to each other honestly about security, intelligence and cooperation. That matters in a region where trust is often fragile and threats do not respect borders.
But beyond the speeches and the formalities, one theme kept returning in different forms: technology. More specifically, artificial intelligence. Not in the abstract, not as a buzzword, but as something already reshaping how societies function—and, just as importantly, how they can fail.
It is easy to speak about AI as an opportunity. And to be fair, it is. It can improve service delivery, support healthcare systems, strengthen agriculture and even help governments make better decisions. Those are real possibilities. But what stood out in Diani was not just what AI can do right—it was what happens when things go wrong.
Because in Kenya, and across Africa, AI does not land on a clean slate. It lands on systems that are already uneven.
Let’s start with inequality. Not everyone has the same access to the digital world. Some areas still struggle with basic connectivity. Others deal with unreliable electricity. Digital skills are not evenly distributed. So when AI tools are introduced, they don’t level the playing field—they often tilt it further. Those who are already connected benefit more, and those who are not fall further behind. That is not a future problem. It is already happening.
Then there is the issue of data. AI runs on data—large amounts of it. But the systems meant to protect that data are still catching up. Yes, progress has been made. The African Union has pushed forward frameworks on data governance, and countries like Kenya have taken steps to strengthen privacy laws. But laws alone are not enough. Enforcement is where things often slow down.
This creates a gap. A space where personal data can be collected, used and sometimes misused without clear accountability. And in a region where trust in institutions is already under pressure, that gap becomes dangerous.
Another issue we cannot ignore is bias. AI systems are only as good as the data they are trained on. If that data is incomplete or skewed, the outcomes will be too. This can show up in subtle ways—who gets access to opportunities, how people are represented online, or even how decisions are made in areas like finance or hiring. If we are not careful, AI can quietly reinforce the very inequalities we are trying to solve.
And then there is the question of truth.
AI has made it easier than ever to generate convincing content—text, images, even video. This creates new challenges for information integrity. In environments where media systems are already stretched, the spread of false or manipulated information can have real consequences. It can influence public opinion, deepen divisions, and erode trust in institutions.
Again, this is not theoretical. It is already part of the information landscape.
What makes all of this more complex is the governance gap. Many countries are working on policies and frameworks for AI. Some are moving faster than others. For instance, reports like the Government AI Readiness Index 2024 show that countries such as Rwanda are making notable progress in readiness. But readiness does not mean protection. It does not automatically translate into the ability to manage risks effectively.
In many cases, the laws exist on paper, but the institutions needed to enforce them are still developing. That gap between policy and practice is where most of the real challenges sit.
There is also a broader concern that does not get enough attention: dependence. Much of the AI technology being used in Africa is developed elsewhere. That is not necessarily a problem in itself, but it raises questions about control. Who owns the systems? Who sets the standards? Who decides how risks are managed?
If those answers sit entirely outside the continent, then even well-intentioned adoption can come with hidden costs.
So where does that leave us after Diani?
For me, the takeaway is simple. We cannot afford to treat AI as just another topic on the agenda. It needs to be at the center of the conversation—not just among technologists, but among policymakers, security experts, and the public.
And more importantly, we need to move beyond celebration.
Yes, innovation should be encouraged. Yes, progress should be recognised. But there has to be equal urgency around safeguards. Around building institutions that can keep up. Around making sure that the benefits of technology do not come at the expense of rights, fairness, and trust.
Because technology will not wait for us to be ready.
It will move forward anyway.
The real question—the one we must not dodge—is whether our systems, our laws, and our leadership will move with it.
