The Department of Communications and Digital Technologies has announced that the final draft of South Africa’s national AI policy will be published in January 2027. This decision follows a comprehensive review process that will include recommendations from a panel of experts.
Communications Minister Solly Malatsi explained in Parliament that South Africa’s previous draft national AI policy was officially withdrawn due to the inclusion of unreliable AI-generated references. The controversy prompted the precautionary suspension of two officials in the department for misuse of generative AI in preparing the document.
Malatsi stated that it wasn’t until News24 published an exposé that the department became aware of the problematic references in the policy. “We contacted our Director-General and the officials involved to confirm the claims,” he said, highlighting the internal investigation that followed.
The draft policy was retracted to maintain the integrity of the policy-making process and to mitigate the repercussions of the scandal. The investigation revealed significant oversight and a lack of transparency regarding the use of AI in drafting the document.
To address these issues, the department has enlisted experts in AI, governance, and academia. In May, Malatsi mentioned the formation of an independent panel to review the national AI policy before its eventual publication. Prof. Benjamin Rosman from the Wits MIND Institute will chair this panel, which consists of distinguished academics, legal specialists, cybersecurity experts, and a commissioner from the National Planning Commission.
This panel will not only review the revised draft of the AI policy but also evaluate the withdrawn draft, determining which elements should be retained and which should be discarded. The purpose is to ensure that the new policy is based on accurate, verified information and to replace any fictitious references.
Additionally, the experts will provide written responses to the minister’s inquiries regarding AI and the policy document. The panel’s review of the new draft will take place by the end of the drafting process, culminating in a final advisory opinion on its adherence to the expert recommendations.
According to Malatsi, the review process is expected to start in May 2026 and conclude by October 2026, with the draft being presented to the Cabinet for final approval. The aim is to submit the revised policy to Cabinet by November 2026, targeting a public comment publication in January 2027.
An internal department team will support the drafting process, ensuring that the department remains actively involved in finalising the national AI policy. Meetings between departmental officials and the expert panel are set to commence in the coming weeks.
