South Africa Withdraws Draft AI Policy Due to Fake References
Introduction
Solly Malatsi, the South African Minister of Communications and Digital Technologies, has officially withdrawn the country's first draft of the national artificial intelligence (AI) policy. This decision follows the discovery that the document contained fake references.
Main Body
The draft policy was approved by the Cabinet on March 25 and was released for public feedback until June 10. The main goal of the framework was to make South Africa a regional leader in AI innovation. Furthermore, the government wanted to balance technological growth with the need to solve social and economic problems. The proposal suggested creating several oversight groups, including a National AI Commission, a Regulatory Authority, and an Ethics Board. To encourage private companies to participate, the policy proposed financial incentives such as grants and tax breaks. Acting Director-General Omega Shelembe emphasized that the document focused on ethical management, specifically aiming to reduce bias and address historical inequalities. However, an internal review revealed that the reference list included sources that did not exist. Minister Malatsi claimed that this was likely caused by the use of AI-generated citations that were not checked by humans. He criticized this as a failure of professional standards and asserted that the incident proves why human oversight is essential when using AI. As a result, the Minister confirmed that the staff responsible for drafting and checking the document will face disciplinary action.
Conclusion
South Africa has paused its AI policy development to fix these integrity issues. The government is now focusing on internal accountability and the importance of human verification in creating digital policies.