On April 27, 2026, a US courtroom will host a dispute that feels less like litigation and more like a reckoning. At issue is not merely a contract or a corporate disagreement, but the philosophical DNA of artificial intelligence itself. Across the aisle will sit Elon Musk and Sam Altman—two figures who once shared a vision, now divided by what that vision has become.
In its earliest incarnation, OpenAI was conceived as a counterweight to concentrated technological power. It was meant to pursue research in the open, to ensure that artificial intelligence would benefit humanity broadly rather than enrich a narrow elite. Musk, one of its early backers, has argued that this founding ethos has been compromised—that the organization has drifted from a public-spirited mission toward a more conventional, profit-driven model. Altman, by contrast, represents a different school of thought: that the scale, cost, and competitive intensity of modern AI demand capital, structure, and, inevitably, a degree of closure.
The case, then, is not simply about what was promised, but about what is possible. Can a technology as powerful as AI remain open in a world defined by competition and geopolitical rivalry? Or does its very importance necessitate control? The answer will reverberate far beyond the courtroom. In the years since OpenAI’s founding, a constellation of new firms—many seeded by its alumni—has embraced proprietary models and closed systems. The industry has, in effect, voted with its architecture. Should the courts compel a return to openness, the implications would be profound, challenging not only business models but the underlying assumptions of the AI economy. For South Africa, the timing of this trial is particularly striking. The country has recently invited public comment on a draft national AI policy ahead of June 10, 2026—a document that reveals both ambition and unease. It acknowledges, with unusual candor, the risks of dependence on foreign infrastructure, particularly in a world shaped by the strategic rivalry of the United States and China. Its recommendation—to reduce reliance on external hardware ecosystems—signals a shift in mindset. It is, in essence, a call for technological sovereignty. If pursued with discipline, that single idea could alter South Africa’s trajectory. It suggests a move away from passive consumption toward active production—a transition that has defined the rise of every major technological economy. Equally significant is the policy’s emphasis on data. It proposes prioritizing the availability of large public datasets for training AI systems, a step that may seem technical but is, in fact, foundational. Data is the raw material of intelligence. Without it, there can be no meaningful participation in the AI economy. And yet, here the policy encounters a more human obstacle. Public institutions are not always designed for openness. Information is often fragmented, poorly organized, or guarded—sometimes out of habit, sometimes out of fear. The act of sharing data requires not only infrastructure, but a cultural shift toward transparency and accountability. It is, in many ways, the harder problem.
What emerges is a paradox. The policy reads as both blueprint and aspiration—a vision of what could be, tempered by the realities of implementation. It is a reminder that strategy, however elegant, must ultimately contend with institutions, incentives, and human behavior. The broader lesson, however, is clear. As the world debates whether AI should be open or closed, countries like South Africa are being forced to answer a more immediate question: whether they will own any part of this future at all. Openness, in this context, is not an abstract ideal. It is a lever. To democratize data is to lower the barriers to entry, to invite participation, and to seed innovation in places that have historically been excluded. It is, quite simply, a way to build. The trial between Musk and Altman may clarify the direction of the global AI industry. But the choices made in Tshwane (Pretoria), and in capitals across the continent, will determine something equally important: whether Africa remains a consumer of intelligence, or becomes a producer of it. In the end, the future of AI will not be decided in a single courtroom. It will be shaped by a series of decisions—legal, political, and institutional—about who gets to build, who gets to access, and who ultimately benefits. And in that unfolding story, openness may prove not just a principle, but a strategy.
These are matters that require more robust engagement and I invite you to join me with other information thinkers as we reflect about AI in Africa over the next few days.
