South Africa's draft artificial intelligence policy was pulled in April 2026 after a damaging discovery: fabricated academic references had made it into the official document. The withdrawal triggered immediate questions about oversight, process, and who exactly was minding the work on one of the country's most consequential technology frameworks.
Five months on, Communications and Digital Technologies Minister Solly Malatsi confirmed that disciplinary proceedings are ongoing. One official has been suspended, another resigned in the wake of the scandal. Speaking to TechCabal on October 15, Malatsi was direct: "Having such a major stain on an important policy is something that cannot just be allowed to slide." He also addressed the delay publicly at the CollectiveXChange 2026 summit in Johannesburg, calling the whole episode a "false start."
Accountability matters, but it is only one piece of the problem. The deeper question is whether South Africa's existing institutions are equipped to actually enforce AI rules and protect people from harm. The government is considering strengthening current regulators rather than creating a brand-new AI authority, which would sidestep unnecessary bureaucracy. The catch is that it would demand far more of institutions that would need to investigate AI-related harms, coordinate enforcement, and deliver real redress.
Without clear responsibilities and adequate resources, any replacement policy runs the risk of reading like a wish list rather than a working framework. Cabinet is expected to finalise that replacement by the end of March 2027, with public consultation to follow. Whether the new document will have the teeth the first one lacked is the question worth watching.
Originally published by TechCabal.