JOHANNESBURG — South Africa has released its draft National Artificial Intelligence Policy for public comment, marking a decisive shift from high-level principles to concrete regulatory action as the country positions itself as a continental leader in AI governance.
The policy was approved by Cabinet on 25 March 2026 and officially gazetted on 10 April, opening a 60-day public consultation period that will close on 10 June. The 86-page document represents years of deliberate policy development, from early work following the 2020 Presidential Commission on the Fourth Industrial Revolution to South Africa’s leadership in developing an AI blueprint for the African Union in 2021.
Six Pillars for Responsible AI
The policy is structured around six core pillars: capacity and talent development; AI for inclusive growth and job creation; responsible governance; ethical and inclusive AI; cultural preservation and international integration; and human-centred deployment.
Minister in the Presidency Khumbudzo Ntshavheni emphasized that the policy aims to ensure both the benefits and risks brought by AI are evenly distributed across society and generations, while strengthening government’s ability to regulate and adopt AI responsibly.
Sweeping New Institutional Architecture
The draft’s most striking feature is its institutional ambition. The policy outlines plans to establish new institutions, including a National AI Commission, an AI Ethics Board and an AI Regulatory Authority to coordinate policy, enforce ethical standards, monitor compliance, and provide mechanisms for redress and compensation in cases of AI-related harm.
Additional proposed bodies include an AI Ombudsperson Office to allow citizens to challenge AI-driven decisions, an AI Innovation Hub, an AI Insurance Superfund, and a National AI Regulatory Forum that would coordinate oversight among existing regulators including the Information Regulator, the Competition Commission, and the Financial Sector Conduct Authority.
The creation of seven new oversight bodies has raised questions about implementation capacity and funding in a country where public resources are already stretched thin. The policy itself does not attach specific budget figures to any of the proposed bodies, though it calls for funding to be secured during the second year of a three-year implementation roadmap.
Sector-Specific, Multi-Regulator Approach
The government has opted for a sector-specific, multi-regulator model rather than a single AI regulator, meaning AI governance will be embedded within existing supervisory frameworks. This approach allows for tailored implementation as different sectors face varying AI risks and applications.
Key regulatory areas include data protection and privacy, with the policy recommending robust data governance standards aligned with global trends such as the European Union’s General Data Protection Regulation, emphasizing that data should be collected, processed, and stored transparently and securely.
Balancing Innovation and Sovereignty
The policy envisions AI as a driver of economic transformation, with education, healthcare, agriculture, and public administration flagged as priority sectors. By fostering an AI ecosystem that promotes innovation, South Africa aims to harness AI to improve competitiveness and stimulate job creation, particularly through support for AI startups and public-private partnerships.
However, the document also expresses concerns about technological dependence. It notes that reliance on foreign infrastructure could compromise the security of sensitive South African data and calls for plans to reduce South Africa’s current hardware dependence on the U.S. and China amid their ongoing geopolitical rivalry.
Phased Implementation Timeline
Finalisation of the policy is anticipated during the 2026/2027 financial year, with sector-specific strategies and supporting regulatory measures expected to follow from 2027/2028. The implementation roadmap includes publishing national AI policy guidelines, implementing key regulatory requirements for high-risk use cases, and developing sectoral AI strategies.
What Comes Next
The Department of Communications and Digital Technologies confirmed that the Draft AI Policy has successfully cleared the Socio-Economic Impact Assessment System certification and achieved concurrence across all Director-General clusters, suggesting broad inter-departmental support.
For organizations operating in South Africa, legal experts recommend using the 60-day comment period strategically. Companies should conduct comprehensive internal reviews of existing AI deployments, identifying high-impact systems, mapping data flows, assessing model explainability mechanisms, and evaluating alignment with existing governance frameworks.
The draft policy positions South Africa alongside other nations navigating the complex terrain of AI regulation. With the EU AI Act now in force and countries from Kenya to Nigeria developing their own frameworks, South Africa’s measured, sector-specific approach reflects a country attempting to balance urgency with the complexity of governing a rapidly evolving technology that moves faster than any policy cycle can comfortably track.
As the public comment period opens, the question is not whether South Africa needs an AI policy—that much is clear—but whether the ambitious institutional architecture proposed can be adequately resourced and coordinated to deliver on the promise of responsible, inclusive AI governance for the continent’s most industrialized economy.
