South Africa’s ambitions to become a continental leader in artificial intelligence governance have suffered a major setback after the government postponed its National AI Policy to January 2027 following the discovery of fabricated references in the original draft.
The controversy has become one of Africa’s most significant examples of the risks associated with unchecked AI adoption in public institutions, particularly the growing concern around AI “hallucinations”, instances where generative AI systems produce inaccurate or entirely fictitious information presented as fact.
The Department of Communications and Digital Technologies confirmed the revised timeline during a parliamentary committee session on Tuesday. Acting Deputy Director-General Jeanette Morwane disclosed that a revised draft would be submitted to cabinet by November 2026 before entering a new public consultation phase in January 2027, pushing the process nearly nine months behind schedule.
PThe original policy draft, released in April 2026, was designed to position South Africa at the forefront of AI innovation and regulation in Africa. However, the document was withdrawn shortly after an investigation by local publication News24 revealed that several citations referenced sources and publications that did not exist.
The incident immediately triggered scrutiny over whether AI tools were used during the drafting process without adequate human verification.
Communications Minister Solly Malatsi admitted before Parliament that internal review systems failed to detect the fabricated references before publication. He described the oversight as significant and acknowledged concerns surrounding transparency in the use of AI-generated material within official government documentation.
While Malatsi maintained that the broader policy framework itself remained largely intact, the credibility of the document was undermined by the discovery of unverifiable citations.
In response, two government officials have reportedly been placed on precautionary suspension pending a formal investigation. Director-General Nonkqubela Jordan-Dyani described the incident as “highly regrettable,” noting that withdrawing the draft was necessary to preserve public trust in the policymaking process.
The government has since constituted a seven-member independent review panel tasked with auditing the original draft, removing questionable sections, and replacing fabricated references with verified sources.
The controversy highlights a deeper irony confronting governments globally: a policy intended to regulate artificial intelligence was compromised by one of AI’s most widely documented weaknesses.
Large language models and generative AI systems can produce convincing but inaccurate information when not properly supervised. Experts have repeatedly warned that without rigorous human oversight, these systems risk introducing misinformation into sensitive sectors including healthcare, finance, education, and public governance.
The country has spent the past two years positioning itself as a leader in AI governance across Africa, particularly after the African Union adopted a continental AI strategy in 2024. Several African governments, including those in Nigeria and Kenya, are currently developing their own national AI frameworks while monitoring South Africa’s progress.
The delay now raises broader questions about institutional preparedness, regulatory discipline, and whether governments are moving faster than their verification systems can support.
For South African businesses, startups, and developers awaiting regulatory clarity around AI deployment, the postponement introduces further uncertainty. Companies seeking guidance on compliance standards, liability frameworks, ethical obligations, and data governance will now wait at least eight additional months before formal public consultations resume.
Beyond the practical implications, the scandal has become symbolic of a wider challenge facing emerging economies eager to demonstrate technological sophistication.
A policy framework built on unverified information risks creating the illusion of readiness while weakening confidence in the institutions responsible for oversight.South Africa’s experience may ultimately serve as a cautionary lesson for governments across Africa racing to establish AI governance structures.
As global competition around artificial intelligence intensifies, policymakers face mounting pressure to introduce frameworks that attract investment and signal digital readiness. However, the South African incident demonstrates that speed without verification can erode credibility faster than innovation can build it.
The scandal reinforces a principle increasingly echoed by AI researchers and governance experts worldwide: AI-generated content should assist institutional work, not replace rigorous human review.
For African governments navigating the early stages of AI regulation, the lesson is becoming increasingly clear, trust in AI policy will depend not only on technological ambition but also on the strength of the verification systems behind it.
