South Africa has withdrawn its draft national artificial intelligence policy after it was found to contain fabricated academic references believed to be generated by AI.
The document, released for public comment earlier this month, was intended to guide the country’s approach to AI development, governance and regulation. It proposed the creation of new structures, including a national commission, an ethics board and a regulatory authority, alongside incentives to encourage private sector participation.
However, concerns were raised after several sources listed in the policy’s references were found to be non-existent.
Communications and Digital Technologies Minister Solly Malatsi confirmed that the draft had been withdrawn, acknowledging that the inclusion of unverified AI-generated citations had undermined the document’s credibility.
Investigations revealed that at least six of the cited academic sources did not exist, pointing to a failure in the verification process.
Malatsi described the issue as more than a minor oversight, noting that it had compromised the integrity of the policy. He also highlighted the irony that the document aimed to establish ethical guidelines for AI use, yet fell short in its own development.
The minister attributed the lapse to insufficient human oversight, warning against over-reliance on AI tools without proper checks and validation.
He said the incident serves as a lesson on the importance of maintaining strict review processes when using emerging technologies in professional and policy environments.
Malatsi confirmed that there would be consequences for those involved in drafting the policy, although it remains unclear when a revised version will be released.
Facebook Comments