The South African government’s withdrawal at the weekend of its recently gazetted national AI policy highlights the need for robust oversight of the use of AI.
The department of communications & digital technologies, led by minister Solly Malatsi, said it had discovered fake and fictitious sources cited in the document, likely introduced by AI.
“Following revelations that the draft national AI policy published for public comment contains various fictitious sources in its reference list, we initiated internal questions which have now confirmed that this was the case,” Malatsi said in a statement.
“This failure is not a mere technical issue but has compromised the integrity and credibility of the draft policy.”
As such, the minister said he would be withdrawing the draft, which his department had published for public comment earlier in the month.
AI investment has grown exponentially in recent years, sparked by the rapid adoption and popularity of OpenAI’s ChatGPT since it was launched in November 2022.
This failure is not a mere technical issue but has compromised the integrity and credibility of the draft policy.
— Solly Malatsi, communications & digital technologies minister
While businesses in South Africa figure out how to make the most of the new technology, the government is under pressure to create rules that protect rights such as privacy while also encouraging innovation and investment.
The government has now, ironically, fallen victim to some of the risk associated with AI use, which the national policy is meant to address and regulate.
“South Africans deserve better. The department of communications & digital technologies did not deliver on the standard that is acceptable for an institution entrusted with the role to lead South Africa’s digital policy environment,” Malatsi said.
“The most plausible explanation is that AI-generated citations were included without proper verification. This should not have happened. In fact, this unacceptable lapse proves why vigilant human oversight over the use of AI is critical. It’s a lesson we take with humility.”
If this is indeed the case, then the department experienced hallucinations — the generation of inaccurate or fabricated content — by whatever AI system or systems were used in conducting the research or producing the document.
Law firm Webber Wentzel has highlighted that hallucinations are “not a theoretical risk”, with such examples already playing out in South Africa’s courtrooms.
“The most obvious example is fabricated case citations, and South African courts have already seen practitioners file submissions referencing cases that do not exist,” said the Webber Wentzel team, including Priyesh Daya, Brittany Leroni and Caitlin Leahy.
“But the more insidious risk lies in the subtler errors: a real case cited with the wrong court, a finding slightly mischaracterised, or facts quietly adjusted to fit the argument being advanced. These errors are more difficult to catch because they look correct from the outset,” it added.
Even before Malatsi was forced to pull the draft, various stakeholders, including private sector players, were complaining that the draft policy has generated more questions for businesses than answers. The government acknowledged that it was far from being complete.
For now, the department will have to rework its proposal, deal with the fallout and try to regain the confidence of the country.
“I want to reassure the country that we are treating this matter with the gravity it deserves. There will be consequence management for those responsible for drafting and quality assurance,” Malatsi said.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.