South Africa has watched the value of its resources leave before. IOL warns it is about to happen again with AI

South Africa has a habit of gaining access to a strategic resource, benefiting from it for a while, and then watching the real value leave for whoever actually owns the system. That is the pattern IOL identifies in an analysis published on 25 September, and it warns that artificial intelligence is the next place it could play out.

The analysis traces the pattern through three earlier chapters of the country’s economic history. Mining gave South Africa access to extraordinary mineral wealth. Electricity gave it the capacity to power an industrial economy. Telecommunications connected it to the world. In each case, according to IOL’s account, the country got partial benefit and then saw the greater share of the value accumulate elsewhere, with the entity that held ownership of the underlying system.

The argument is not that South Africa missed out entirely. It is that access and control are different things, and that the gap between them is where the value went.

Artificial intelligence now presents the same fork, according to the analysis, and the warning is that the window for choosing differently is smaller than many people in policy positions seem to recognise.

That is a serious claim, and it deserves to be read on its own terms rather than dismissed as techno-pessimism. South Africa is not a passive observer of the global AI build-out. It has a growing community of machine learning researchers and engineers, a financial sector that has been an early adopter of automation, and universities producing graduates who work in AI systems for companies around the world. The question IOL raises is not whether South Africans can use AI. It is who will own the infrastructure, the data and the models that determine where the value sits.

The distinction matters for ordinary people because it shapes what kind of jobs are created, what kind of skills are rewarded, and whether the productivity gains of the next decade are retained in the domestic economy or exported. It also shapes what public services can do with the technology, from health diagnostics to traffic management to the administration of grants and identity documents.

The analysis does not claim to know which way this will go. It argues that the choice is still open, and that it is closing.

For a reader trying to work out what to watch, the useful question is not whether South Africa adopts AI. It is whether the country builds and controls enough of the stack to keep the returns. That means the data centres, the compute, the models, the platforms and the skills pipeline, not just the applications that sit on top of someone else’s system.

IOL’s piece is a warning, and it is framed as one. But the underlying point is constructive. South Africa has the human capital, the market size and the institutional depth to be a meaningful participant rather than a customer. The pattern it describes is not a law of nature. It is a set of choices that were made, and could be made differently.

The Enquirer will continue to report on the companies, policies and people shaping South Africa’s position in artificial intelligence, and on what those decisions mean for the country’s economy and its citizens.

Source: IOL, South Africa’s AI Crossroads: Access Is Not the Same as Control