South Africa has withdrawn its first draft national artificial intelligence policy after officials discovered that several academic citations in the document were fake and likely generated by AI tools. The policy, which had been released for public comment, was intended to position South Africa as a leader in AI governance across Africa before the credibility issue forced its removal.
According to reports, at least six references in the policy bibliography pointed to academic papers that did not actually exist, even though the journals themselves were real. Communications Minister Solly Malatsi admitted the citations were likely generated through AI without proper verification and called the incident an “unacceptable lapse” that damaged the integrity of the draft policy.
The withdrawn policy had proposed major AI governance structures including a National AI Commission, AI Ethics Board, AI Regulatory Authority, and new incentives for private-sector AI development. Analysts say the controversy highlights a growing global problem where governments, researchers, and institutions increasingly rely on generative AI systems that can produce convincing but fabricated information.
Experts say the incident has become a major example of why human oversight remains critical when using generative AI for official work. The case has also intensified broader discussions about AI hallucinations, misinformation, and institutional readiness to regulate technologies that policymakers themselves may not fully understand.
