South Africa has withdrawn its first draft national artificial intelligence policy after authorities confirmed that false references were included in the docum...
South Africa has withdrawn its first draft national artificial intelligence policy after authorities confirmed that false references were included in the document’s bibliography, creating a setback for the country’s digital policy agenda.
Communications and Digital Technologies Minister Solly Malatsi announced the decision after an internal review found that several sources listed in the policy could not be verified. He said the most likely explanation was that AI-generated citations had been inserted without being properly checked.
Advertisement
“The most plausible explanation is that AI-generated citations were included without proper verification. This should not have happened,” Malatsi said in a public statement.
The draft AI policy had been published earlier this month for public comment and was meant to guide how South Africa develops and regulates artificial intelligence. It aimed to place the country among Africa’s leaders in AI innovation while addressing social, ethical and economic risks linked to the technology.
Among its proposals were plans to establish a National AI Commission, an AI Ethics Board and an AI Regulatory Authority. It also suggested incentives such as grants, subsidies and tax support to encourage private sector investment in AI development.
Malatsi said the discovery of fabricated references had damaged trust in the document and made its withdrawal necessary. He added that the issue was more than a technical error because it affected the integrity of a major national policy.
The minister said officials responsible for drafting and reviewing the policy would face consequences. However, he did not provide details on disciplinary action or say when a revised version of the AI policy would be released.
The incident has highlighted growing concerns around the use of artificial intelligence tools in official work without strong human supervision. AI systems can generate convincing text and references, but they can also produce false material if outputs are not carefully checked.
South Africa has been positioning itself to play a larger role in the digital economy, with government support for innovation, data infrastructure and technology investment. The withdrawal of the policy is likely to delay part of that strategy while authorities prepare a corrected version.
Despite the setback, the government said it remains committed to building a credible AI framework for the country. Officials are expected to restart consultations once a revised draft has been completed and verified.
For South Africa, the episode serves as a warning that even AI policy itself requires strict human oversight. The country now faces the task of rebuilding confidence in a document meant to shape its digital future.
