South Africa’s Department of Communications and Digital Technologies published a draft national AI policy on Friday, April 10, opening it for public comment through June 10. The policy builds on the National Artificial Intelligence Policy Framework published in August 2024 and was informed by 32 stakeholder submissions, according to iAfrica.

“The aim of the AI policy is to ensure that both the benefits and risks brought by AI are evenly distributed across society and generations,” Cabinet said in a statement approving the publication, per iAfrica.

Six Strategic Pillars

The draft is organized around six pillars, as detailed by iAfrica:

  1. Capacity and talent development, including integrating AI into school curricula from primary through tertiary levels
  2. AI for inclusive growth and job creation, with plans to establish specialized AI research institutions
  3. Responsible governance, covering safety controls and legal remedies for deepfakes and defamatory AI content
  4. Ethical and inclusive AI, focused on fairness, transparency, accountability, and preserving democratic values
  5. Cultural preservation and international integration
  6. Human-centered deployment, maintaining human control over AI decision-making, particularly in generative AI systems

Distributed Oversight, Not a Single Regulator

A significant structural decision: the policy chooses distributed oversight across existing regulatory authorities rather than creating a centralized AI regulator. The Information Regulator handles privacy, while other sector-specific agencies will coordinate on AI governance. Law firm Baker McKenzie described this as “a coordinated oversight approach rather than the creation of a centralised AI regulator,” according to iAfrica.

In the interim, AI deployments remain governed by existing legislation including the Protection of Personal Information Act, the Consumer Protection Act, and the Cybercrimes Act.

Three-Phase Implementation

The policy follows a phased three-year rollout:

  • Phase 1 (2025-2026): Finalizing the draft policy and identifying regulatory requirements. Now complete.
  • Phase 2 (through March 2027): Publishing national AI guidelines, implementing rules for high-risk use cases, and beginning institutional framework design.
  • Phase 3 (2027-2028): Full implementation of all policy interventions, with provisions for updates as technologies evolve.

Public comments are invited by June 10. The full policy document is available in the South African Government Gazette.

The Global AI Governance Pattern

South Africa joins a growing list of non-G7 economies building proactive AI regulatory infrastructure. Brazil, India, and the UAE have all established or proposed AI governance frameworks in 2025-2026, each with explicit provisions for autonomous systems. The timing also matters for the local ecosystem: VALR, the South Africa-based crypto exchange, launched a dedicated AI Service for autonomous agent trading earlier this week, a product that falls directly under the scope of this draft policy.