Mozilla recently submitted detailed comments to South Africa’s Department of Communications and Digital Technologies (DCDT) regarding the proposed National AI Policy Framework. This framework seeks to advance AI development in South Africa responsibly, emphasizing a human-centered approach with a focus on ethical AI deployment, fairness, and bias mitigation. These elements resonate with Mozilla’s vision of trustworthy AI, where individuals have control over their data and algorithmic outputs. We also commend the framework’s commitment to data privacy and its initiatives to integrate AI into educational curricula and public services.
However, transparency in the framework’s development process remains an area of concern, particularly as it was not made readily accessible to the public. Mozilla believes that inclusive, participatory governance is essential, especially involving the youth to ensure diverse perspectives shape AI policy. Furthermore, while the framework mentions transparency in AI systems, it lacks a concrete commitment to open-source AI—a critical element for transparency across the AI lifecycle.
Mozilla believes that the framework’s implementation will benefit from greater transparency around AI system operations and public awareness campaigns. While the framework suggests initiatives to educate the public on AI, it could include more actionable steps to make AI systems understandable and accountable. By providing clear, accessible information on how AI operates and affects users, South Africa can build stronger public trust in AI technologies, fostering a culture of responsibility across sectors.
The framework could also strengthen its support for local AI infrastructures and competitive markets, preventing reliance on global players and ensuring a thriving, diverse AI ecosystem. Additionally, connecting AI governance to existing data policies would ensure a more cohesive approach to ethical AI development. While ethical guidelines are proposed, the framework would benefit from outlining clear enforcement mechanisms, making regulatory compliance a practical reality. Addressing the environmental impact of AI and worker ethics is another area for potential enhancement.
We are glad that DCDT has made a commitment to take into consideration our inputs. By fostering an open and inclusive process, supporting local AI talent, and ensuring ethical governance, the DCDT’s policy framework has the potential to advance a trusted, inclusive AI ecosystem that reflects the needs and values of all South Africans. We look forward to seeing South Africa lead the way in AI governance, creating a model that other nations in the region can follow to promote equitable and responsible AI development.