Africa’s 1.4 billion people across 54 countries hold collective bargaining power that no single nation on the continent can match. Whether that power is exercised, or simply left on the table, may determine who writes the rules governing artificial intelligence for a generation. That was the central argument University of Johannesburg Vice-Chancellor and Principal Professor Letlhokwa Mpedi put to business leaders, technologists, policymakers and academics at the CNBC Africa AI Summit 2026 on Thursday, 27 August.
Speaking during a fireside chat titled “Beyond Regulation: Governing the Future of AI in Africa,” moderated by CNBC Africa Editor-in-Chief Godfrey Mutizwa, Prof Mpedi framed the choice plainly. Africa can import governance structures designed elsewhere, or it can become an active architect of the technologies, policies and rules that will increasingly shape daily life across the continent. The difference is not academic. It is a question of who AI ultimately serves.
“Africa has a genuine window to write rules that are protective and generative,” Prof Mpedi said.
The governance challenge runs the full length of the AI value chain, from developers and deployers to the institutions that deploy these systems. That challenge becomes acute when AI influences decisions in finance, healthcare, policing and justice. Prof Mpedi cautioned against treating algorithmic accuracy as a substitute for human judgment. AI systems operate probabilistically and can reproduce biases embedded in their training data. In high-stakes environments, accountability cannot be delegated to an algorithm alone.
Digital sovereignty, he argued, extends well beyond physical infrastructure. It encompasses data, intellectual property, knowledge and culture, along with the capacity to negotiate collectively. The African Continental Free Trade Area could serve as a platform for coordinated African positions on AI governance, including governance of the data used to train AI systems. A unified continental approach to training data could become a meaningful assertion of that sovereignty.
Language is a particular vulnerability. Africa is home to more than 2,000 languages, yet AI systems predominantly operate in English, French and Portuguese. This gap is not merely technical. Language carries culture, knowledge and distinct ways of understanding the world. If AI systems cannot adequately engage with African languages and contexts, they risk failing African societies and economies outright. The implication is direct: Africans must code, write and publish in their own languages so that African knowledge and perspectives are embedded in the data shaping future AI systems. Without that, the continent risks remaining a source of raw material for technologies developed and owned elsewhere.
Intellectual property law offers a useful precedent here. South Africa’s protection of rooibos through geographical indication shows how African knowledge and cultural outputs can be shielded from appropriation and monetisation abroad without attribution or benefit to originating communities. Similar frameworks could protect African cultural heritage and indigenous knowledge in the age of generative AI.
Meanwhile, building genuine AI sovereignty requires sustained investment in people, research capacity and institutions. South Africa currently invests approximately 0.6% of GDP in research and development, compared with more than 2% in many developed economies. That gap points to a critical role for universities as research and innovation partners to industry and government. The private sector alone cannot address every research challenge. Universities can connect graduate students to real industry problems while cultivating the next generation of African research leaders. Stronger intra-African university partnerships, with fewer bureaucratic barriers, could build a more robust continental research ecosystem.
The educational imperative is equally pressing. Graduates entering an AI-driven economy will need more than the ability to use AI tools. They require technical skills to develop and deploy AI, alongside critical thinking, ethical reasoning, legal literacy and civic responsibility. In South Africa, where youth unemployment exceeds 45%, that combination is not a luxury. It is a survival requirement.
The University of Johannesburg’s own work reflects this focus on societal impact. AI and machine learning are being applied to healthcare, agriculture, conservation, climate resilience and access to justice, in what the university describes as a “lab to life” approach. The Institute for AI and the Law and the iKamva Initiative connect technological innovation with justice, accountability and human dignity. Through a metaverse-based legal clinic, students engage with AI-driven virtual clients and simulated cases covering gender-based violence and Road Accident Fund claims. Further detail on this work is available at https://news.uj.ac.za/news/africa-must-become-a-rule-maker-not-a-rule-taker-in-the-age-of-ai-says-ujs-vice-chancellor-prof-mpedi/
The ultimate measure of AI’s value, Prof Mpedi argued, lies not in what the technology can do in isolation, but in whether it expands opportunity, improves lives and helps address society’s most pressing challenges. Africa is not simply responding to the AI era. The open question is whether the continent will build the talent, institutions, infrastructure and partnerships needed to shape the systems before those systems shape it.