Pressure mounts after SA AI policy cites fake sources

New questions are being raised about South Africa’s draft artificial intelligence policy after it emerged that some of the research it cites may not exist.

This follows earlier reporting that flagged “fictitious references” in the document — a problem experts say is consistent with AI-generated hallucinations.

Since those reports, scrutiny has intensified around how the policy was compiled and whether sufficient verification processes were followed before publication.

The concern is not just technical. It goes to the credibility of a national policy meant to guide how AI will be governed in South Africa.

Officials have not publicly detailed how the incorrect references made it into the draft. This has not been confirmed.

What has changed is the level of attention. The issue has moved from a niche tech concern into a broader governance question about accountability, especially as AI tools become more widely used in government work.

For readers in the Eastern Cape, where digital access and public trust in institutions remain uneven, the implications are practical. Policies built on weak or unverified information can directly affect how resources, education programmes, and technology rollouts are implemented in rural areas.

The controversy also highlights a wider risk: using AI tools without proper human oversight. Pondoland Times’ own AI policy stresses that all AI-assisted content must be verified by humans to avoid exactly this kind of error.

What happens next will depend on how government responds. Experts are likely to push for a review of the draft policy, stricter fact-checking processes, and clearer disclosure when AI tools are used in official work.

For now, the key issue remains unresolved: how much of the policy can be trusted in its current form.

Readers should treat the draft as under scrutiny until further clarification is provided.

We will update this story as more information becomes available or if government issues a formal response.