South Africa has withdrawn its draft national AI policy after it was discovered that some of its contents were AI-generated and that it cited fictional sources.
A draft policy on AI, intended to highlight South Africa as a leading country in AI innovation, was unveiled for public comments on April 10. The policy addressed the social, economic and ethical challenges of AI, and explained plans to establish an AI commission, ethics board and regulatory authority.
However, Communications Minister Solly Malatsi withdrew the policy after it was discovered by South Africa’s News24 that, despite referencing real academic journals, at least 6 of its 67 citations were AI-generated “hallucinations”.
The aftermath of the failure
The editors of the journals referenced by the draft policy each confirmed independently that the cited articles were fake.
Taking to X/Twitter, Malatsi explained the failure. “The most plausible explanation is that AI-generated citations were included without proper verification,” he said. “This failure is not a mere technical issue but has compromised the integrity and credibility of the draft policy.”
@SollyMalatsi / XThe X post by Solly Malatsi outlined the reasons for the failure.
“This unacceptable lapse proves why vigilant human oversight over the use of artificial intelligence is critical,” he continued. “It’s a lesson we take with humility.” Malatsi has said those responsible for drafting would face consequences for their oversight.
