South Africa delays AI policy to 2027 after citation scandal forces rethink

South Africa’s long-awaited national artificial intelligence (AI) policy has been delayed to January 2027 after the government withdrew an earlier draft over fabricated academic references.

The setback has triggered renewed scrutiny over how generative AI is being used in policymaking and exposed weaknesses in government oversight.

A delegation from the Department of Communications and Digital Technologies, led by Communications Minister Solly Malatsi, briefed parliament on Tuesday morning about new efforts to rebuild confidence in the country’s AI governance agenda after what officials described as a major credibility crisis.

The delay reinforces the tension facing African governments racing to regulate AI while still building institutional capacity to understand and govern it. South Africa had hoped to position itself as a continental leader in AI regulation and innovation, but the collapse of its first draft policy has exposed risks around overreliance on generative AI, weak internal oversight, and the challenge of crafting credible rules for a fast-moving technology that is already reshaping business and public services.

The original draft policy, approved by Cabinet in March and gazetted in April for public comment, was withdrawn weeks after reports revealed that several references cited in the document appeared to be fictitious or attributed to journals that had never published the work in question.

Addressing MPs, Malatsi acknowledged that the department had failed to detect the problems before the media exposed the scandal.

“The department had not picked up that there were issues with the references in the draft policy document before the events were exposed in news reports,” he said.

The minister revealed that two officials had been suspended over the embarrassing dent on South Africa’s efforts to be a continental leader in policing AI.

“It was then that we got the responses to protect the integrity of the policy development process and, obviously, the stain that it has caused not just on the department but also on the government’s overall process of formulating and finalising policy,” Malatsi added.

The minister said the department would tighten internal controls and implement responsible AI-use measures to avoid a repeat of the incident.

On May 14, the government appointed an independent AI review panel that will rebuild the withdrawn document and recommend revisions before it is resubmitted to Cabinet later this year.

The panel will be chaired by Prof. Benjamin Rosman of the Machine Intelligence and Neural Discovery Institute at the University of the Witwatersrand. It includes experts in AI research, law, governance, and digital policy, among them Prof Vukosi Marivate, Prof Alison Gillwald, Bowman’s partner Heather Irvine, Dr Tshepo Feela, cybersecurity expert Jabu Mtsweni, and cyber lawyer Lufuno Tshikalange.

With the revised framework only expected to be opened for public comment in January 2027, South Africa remains without a formal national AI policy even as businesses and government institutions rapidly adopt AI-powered systems. The delay now places mounting pressure on Malatsi to steer policymakers toward regulating a technology evolving far faster than government processes can keep pace with.