South Africa's AI Policy Crisis Deepens as Replacement Delayed to March 2027
South Africa's AI policy imploded in April 2026 when government withdrew its draft framework after discovering fabricated academic references. Communications Minister Solly Malatsi confirmed to TechCabal on October 15 that disciplinary proceedings continue against one suspended official, while another resigned over the scandal.
The replacement policy now targets March 2027—nearly a year later. Malatsi calls it a "false start" at the CollectiveXChange summit in Johannesburg. Instead of creating a dedicated AI regulator, South Africa plans to empower existing bodies across financial services, communications, and consumer protection. But coordination gaps loom large when AI harm crosses sector boundaries.
The stakes are real for consumers. Banks and insurers already use machine learning for credit scoring, fraud detection, and risk pricing. SABRIC recorded R2.4 billion ($143.8 million) in digital banking crime losses in 2025, up 26% from the previous year. TransUnion research shows deepfake impersonation driving fraud sophistication across Africa—yet no dedicated tracking exists for AI-enabled crimes.
Capacity is another hole. One regulator handles 200+ child sexual abuse material cases with just two staff. Malatsi insists ethical AI adoption must include redress avenues, but the National Consumer Commission admits existing laws don't guarantee regulators can investigate AI violations effectively.
March 2027 is a milestone, not the finish line. Public consultation will test whether the replacement combines ethical principles with enforceable rules, clear institutional responsibilities, and actual resources.
As Nigeria develops its own AI strategy, will we prioritize enforcement capacity over policy documents?
SOURCE: https://techcabal.com/2026/11/02/sa-ai-policy-enforcement-challenge/