Navigating South Africa’s AI Future: Sovereignty, Justice, and the Power of Local Networks
🎧 [Listen to the full podcast here]
At Data Science for Social Impact (DSFSI), we are committed to ensuring that the evolution of technology in Africa is not just a matter of adoption, but a process of deliberate creation and sovereign governance.
We are excited to share a new podcast episode curated by Prof Vukosi Marivate, which uses NotebookLM to synthesise the complexities of the Draft South Africa National AI Policy (2026). This podcast serves as a critical bridge, linking the government’s new strategic blueprint to recent global benchmarks like the Stanford AI Index 2026 and the CIPIT State of AI in Africa Report 2025.
Thanks for reading DS@UP 🚀! Subscribe for free to receive new posts and support my work.
Photo by Luke Madziwa on Unsplash
Beyond Data: The Call for Algorithmic Sovereignty
A central theme of this discussion is the urgent need to move beyond mere data sovereignty. While acts like POPIA are essential for privacy, Professor Benjamin Rosman argues that South Africa must attain algorithmic sovereignty, the local capacity to design, develop, and govern AI systems.
Without building our own “algorithmic refineries,” we risk becoming perpetual exporters of “raw digital oil” (data) while importing expensive foreign insights that often misunderstand our local cultures, languages, and social realities.
Addressing Inequitable Openness: The NOODL Solution
The podcast highlights groundbreaking work from our own network. In response to “inequitable openness”, where African datasets are often extracted by Global North actors without returning benefits to local communities, Chijioke Okorie and Melissa Omino developed the Nwulite Obodo Open Data Licence (NOODL).
The name, meaning “raising up the community” in Igbo, reflects a tiered approach that differentiates conditions of use based on a user’s geography and development context, ensuring that African datasets serve the communities that curated them.
A “Bureaucrat’s Dream” or a Strategic Necessity?
The 2026 Draft Policy is ambitious, proposing a sweeping institutional architecture that includes a National AI Commission, an AI Ethics Board, an AI Regulatory Authority, and even an AI Insurance Superfund.
While some critics view this as a potential “bureaucrat’s dream,” the podcast explores whether this infrastructure is necessary to protect human rights and mitigate the “jagged frontier” of AI performance that often leaves African contexts behind.
Scaling Our Intellectual Infrastructure
A sovereign future is impossible without a robust local research ecosystem. We highlight the explosive growth of local academic networks from the African Institute of Data Science and AI (AfriDSAI) @ UP, the MIND Institute at Wits to the Data Science Law Lab at the University of Pretoria and grassroots communities like the Deep Learning Indaba. These institutions are the engines of “intellectual infrastructure” required to set global AI standards from a South African perspective.
The Lens of Ubuntu AI
We conclude the episode by reflecting on the policy’s use of Ubuntu as a guiding philosophy. By framing AI development through the lens of interdependence and community responsibility, South Africa has the opportunity to reject the individualist rationality of Western ethics in favour of a technology that serves the common good and human dignity.
This podcast was created to help researchers, policymakers, and citizens understand the nuances of this draft policy in context. We hope you find it a valuable shortcut into one of the most important policy debates of our time. 🇿🇦🤖
🎧 [Listen to the full podcast here]
Thanks for reading DS@UP 🚀! Subscribe for free to receive new posts and support my work.
Stay connected with our work:
- DSFSI: https://linktr.ee/dsfsi
- AfriDSAI: https://linktr.ee/afridsai