Artificial intelligence embarrassed the government by writing its own policy

South Africa has withdrawn its national artificial intelligence policy document after it was discovered that part of it was written using AI and contained non-existent scientific sources. The government's document, which aimed to make it a leader in AI innovation, unexpectedly put country officials in an awkward position.

Last month, South African Minister of Communications Solly Malatsi announced his decision to withdraw the national AI policy for review. This was due to the fact that at least six out of the 67 scientific sources cited in the document were fabricated by artificial intelligence, referencing academic papers that do not actually exist.

"The most plausible explanation is that references generated by artificial intelligence were included in the document without proper verification. This should not have happened," Solly Malatsi said. He emphasized that this is not a simple technical error, but a serious flaw that has undermined the integrity and credibility of the policy draft.

The AI policy put the government in an awkward position

South Africa’s National Artificial Intelligence Policy aimed to position the country as one of the leading nations in AI innovation. At the same time, the document also intended to address ethical, social, and economic challenges associated with the technology.

However, once it was revealed that the document used non-existent scientific sources, the initiative turned into an embarrassment for the government. Media outlets reported that a significant portion of the document may have been prepared using artificial intelligence.

"This unacceptable flaw proves how important constant and meticulous human oversight is when using artificial intelligence. This is a lesson we accept with humility," the Minister of Communications added.

References to non-existent academic papers provided

South Africa's News24 was the first to report that the country's draft AI policy contained fake scientific sources. The document had recently been submitted for public consultation.

According to the publication, the references included papers that do not actually exist. Among them were sources attributed to academic journals such as the South African Journal of Philosophy, AI & Society, and Journal of Ethics and Social Philosophy.

This situation has once again brought to the forefront the characteristic of artificial intelligence known as "hallucination," where AI can present information that looks convincing but is actually incorrect or fabricated.

The document will be revised and those responsible will be held accountable

Solly Malatsi also announced that certain measures would be taken against individuals responsible for drafting the national AI policy. Emphasizing that there will be "consequences" for those responsible, the minister stated that the document will be resubmitted for public consultation only after being revised.

This incident demonstrated that relying solely on artificial intelligence is not enough when preparing crucial documents like state policies. The information it provides, especially scientific sources and citations, must be meticulously verified by humans.

This incident in South Africa serves as a reminder that while utilizing the capabilities of artificial intelligence, one must not forget the responsibility of human oversight and data verification.