AI isn’t a future headline. It’s a present tense power shift; and South Africa is already living in the aftershock.
In this episode of Both Sides, Randall Abrahams sits down with Benjamin Rosman to unpack the speed of AI development, what makes today’s tools feel “general” and accessible, and why Africa’s biggest challenge might be language, data, and representation.
Rosman explains how African languages are structurally under-served by modern AI because many lack the large datasets these systems need to learn. That gap doesn’t just create inconvenience; it risks exclusion.
But the conversation isn’t only about what’s missing. It’s also about what’s possible: building AI for Africans, by Africans, developing local capacity, and ensuring the continent helps shape the technology rather than only adopting it.
Then comes the uncomfortable flip side: trust. If AI systems are trained to give answers that sound right – and people increasingly believe them – what happens when confidence replaces verification?
The future is moving fast. The real question is whether South Africa helps write the rules, or simply follows them. #eNCA
