South Africa Suspends Two Officials After AI Policy Draft Contains Fake References

The Department announced Thursday that it removed the draft National Artificial Intelligence Policy from public consultation. Officials found that the document cited sources that do not exist, prompting an immediate halt to the consultation process.

Director‑General Nonkqubela Jordan‑Dyani said the suspensions are a precautionary measure while an internal investigation examines how the policy was produced. The identities and seniority of the two staff members were not disclosed.

The department attributed the breach to “irresponsible use of AI tools,” stating that reliance on generative AI software introduced false citations and compromised the document’s integrity. The statement framed the suspensions as part of a broader accountability drive.

An internal review will determine whether the misuse stemmed from inadequate oversight, lack of training, or deliberate manipulation. The outcome will be communicated after the investigation concludes.

The incident highlights growing challenges for governments integrating AI into policy work. While AI can accelerate drafting, unchecked outputs risk misinformation and legal exposure. South Africa’s experience may prompt other nations to tighten verification protocols for AI‑generated content.

Watch for the investigation’s findings and any subsequent policy on AI tool usage within public sector drafting processes.