South Africa Withdraws Draft AI Policy Over Fake Citations
South Africa’s Minister of Communications and Digital Technologies, Solly Malatsi, has withdrawn the country’s Draft National Artificial Intelligence Policy after it was found to contain fictitious sources in its reference list.
Malatsi disclosed this in a statement shared via his X (formerly Twitter) account on Sunday, describing the development as a serious breach that undermined the credibility and integrity of the proposed policy framework.
According to the minister, internal reviews were initiated following concerns about the authenticity of references cited in the document. The checks confirmed that several of the listed sources were not genuine.
He said the issue extended beyond a technical oversight, pointing instead to a broader failure by the Department of Communications and Digital Technologies to meet expected standards in policy formulation.
“This failure is not a mere technical issue but has compromised the integrity and credibility of the draft policy. As such, I am withdrawing the Draft National Artificial Intelligence Policy,” Malatsi stated.
He added that South Africans deserved higher standards from institutions responsible for shaping the country’s digital future.
Malatsi attributed the lapse to the likely inclusion of artificial intelligence-generated citations without proper human verification.
“The most plausible explanation is that AI-generated citations were included without proper verification. This should not have happened,” he said, noting that the incident underscores the importance of human oversight in the use of AI tools.
He further described the situation as a lesson in the risks associated with deploying artificial intelligence in sensitive areas such as public sector policymaking without rigorous checks.
The minister assured that the government was taking the matter seriously, adding that accountability measures would be implemented.
“There will be consequence management for officials responsible for drafting the policy and carrying out quality assurance checks,” he said.
The withdrawal follows an earlier report by News24, which alleged that the draft policy contained at least six fictitious references. Experts cited in the report noted that the false entries bore characteristics of AI hallucinations—instances where AI systems generate plausible but inaccurate information.
The revelations sparked criticism over the credibility of the policy drafting process and raised concerns about governance standards in emerging technology regulation.
The development is expected to delay South Africa’s efforts to formalise a national framework for artificial intelligence, as parts of the consultation process may need to be restarted.
South Africa is among a limited number of African countries actively developing national AI policies, amid growing adoption of the technology across the continent.
In Nigeria, the Minister of Communications, Innovation and Digital Economy, Bosun Tijani, recently announced that the country’s National Artificial Intelligence Strategy had been finalised and was awaiting submission to the National Assembly for legislative approval.
However, as of the time of filing this report, the approval process for Nigeria’s AI strategy remains pending.
