South Africa’s Dept of Home Affairs suspended two senior officials after AI-hallucinated citations were found in a revised citizenship and immigration white paper.
Key Takeaways
The hallucinated references appeared only in the standalone reference list, not cited in the body; DHA maintains the policy content itself remains valid.
The Chief Director of citizenship and immigration was suspended immediately; a second director faces suspension the following week.
Two independent law firms were appointed: one to manage disciplinary proceedings, one to audit all policy documents produced since 30 November 2022 (ChatGPT’s public launch date).
A week earlier, the Dept of Communications and Digital Technologies withdrew its draft National AI Policy for the same reason: fabricated citations.
DHA will now require AI declarations and checks as mandatory steps in its internal document-approval workflow.
Hacker News Comment Review
The single comment frames this as a stakes-based enforcement story: high-consequence domains like law and government will discipline AI misuse; low-stakes work won’t, creating an uneven accountability landscape.
Notable Comments
@quantified: “In important work (law, government) there are consequences to using AI. In unimportant work, you can use it without worry.”
