South Africa withdraws national AI policy after at least 6 of 67 academic citations found to be AI-generated hallucinations
https://thenextweb.com/news/south-africa-ai-policy-hallucinated-citations
Publish Date: 2026-04-28 17:24:36
Source Domain: thenextweb.com
Summary
South Africa’s Communications Minister Solly Malatsi recently withdrew the nation’s draft national AI policy due to the discovery that six out of its 67 academic citations were AI-generated fabrications published in real journals but never existed. This exposed a fundamental flaw in the policy’s authorship and verification process. Initially approved by the Cabinet and made available for public feedback, it became clear that generative AI used for drafting was not checked thoroughly. Malatsi termed this an “unacceptable lapse” that had compromised the policy’s credibility and implied that accountability measures would follow. This incident raises serious concerns about South Africa’s institutional capacity to regulate AI, highlighting broader global issues about the reliability and verification needs of AI-assisted documents within critical governance spheres.
The reverberations extend beyond policy, questioning the technical competency of the department responsible for overseeing AI governance and contributing to growing public skepticism over AI usage. Though the immediate consequence is that the draft policy will need to be redrafted and resubmitted, the broader issue warns that the failure mode of AI in document generation often flies under the radar, requiring detailed human verification to uncover its potential missteps.
Key Points:
- Withdrawal of AI Policy: South Africa withdrew its draft national AI policy after finding AI-generated fabrications in its citations.
- Governance and Credibility Issues: The failure calls into question the institutional capacity of the department tasked with regulating AI.
- Global Verification Challenges: Similar instances globally highlight the growing challenges and the need for rigorous verification in AI-generated documents.
- Need for Human Oversight: The incident underscores that human verification remains essential in contexts where AI is used to generate policy documents.
- Public Trust Erosion: This event contributes to skepticism towards the use of AI in governmental processes due to its inherent unreliability in producing factual content.
