A major AI scandal has erupted in South Africa — Reuters
South Africa has scrapped the first draft of its national AI policy after it emerged that the document contained fabricated sources and had been generated by artificial intelligence itself. Authorities have promised to hold those responsible to account.
This was reported by Reuters.
The draft policy, released this month for public comment prior to final approval, aimed to position South Africa as a continental leader in AI innovation and to address ethical, social, and economic challenges.
The document outlined plans to create new institutions, including a National AI Commission, an AI Ethics Council, and an AI Regulatory Authority, as well as to introduce incentives such as tax breaks, grants, and subsidies to encourage businesses to participate in AI development.
However, it later emerged that the draft document contained quotes from sources that do not exist and were added “without proper verification.” As a result, the policy draft was shelved.
The country’s Minister of Communications and Digital Technologies, Solly Malachi, called the situation a “failure” that “jeopardizes the integrity and credibility of the policy draft” and promised consequences for those responsible for drafting the text.
“This unacceptable blunder demonstrates why vigilant human oversight of artificial intelligence use is critically important. It is a lesson we accept with humility,” he added.
As a reminder, VAKS will implement artificial intelligence for data analysis.
It is worth noting that the implementation of artificial intelligence in large companies follows a similar pattern: 20% of employees actively use AI, 20% categorically refuse to use it, and 60% use the tools only partially. Google falls roughly in the middle of this broad picture.
A study published in JAMA Network Open showed that artificial intelligence models from OpenAI, Google, and DeepSeek often produce errors due to a lack of data when making an initial diagnosis. An analysis of 21 models revealed that the error rate in differential diagnosis exceeds 80%.
