South Africa’s Draft National AI Policy has officially entered the Cabinet approval stage, signaling a pivotal move from high‑level principles to concrete regulatory action. The Department of Communications and Digital Technologies (DCDT) briefed Parliament on February 24, 2026, confirming that the policy has cleared the Socio‑Economic Impact Assessment System certification and secured inter‑departmental concurrence—key prerequisites for Cabinet consideration. The policy is expected to be gazetted for a 60‑day public consultation in March 2026, with finalization targeted for the 2026/2027 financial year. (bakermckenzie.com)
The Draft AI Policy adopts a sector‑specific, multi‑regulator model rather than establishing a standalone AI regulator. Oversight will be embedded within existing supervisory frameworks, such as communications, finance, and healthcare regulators. The policy is structured around five core pillars: capacity and talent development; responsible governance; ethical and inclusive AI; cultural preservation; and human‑centred deployment. (bakermckenzie.com)
This approach positions South Africa between the EU’s comprehensive AI Act and India’s lighter‑touch model. It aims to balance innovation with accountability, embedding AI oversight within existing regulatory systems while tailoring strategies to sector‑specific risks. Businesses operating in South Africa should begin assessing their AI deployments and governance structures in anticipation of increased oversight and forthcoming sector‑specific regulations. (techmoonshot.com)