In Cape Town, South Africa, the government had asked for public comment on a draft AI bill that was created using AI. Oh, the irony. But it gets even better, it had hallucinated sources, citing at least six academic papers that didn’t exist. Needless to say, Cape Town Authorities removed the draft shortly thereafter.
The bill had cited 67 academic papers, with at least six having been hallucinations. Of course, it doesn’t take a genius to figure out that this draft was most likely AI-generated. And that the fake papers were included without human verification, which “should not have happened,” but it is now becoming a very popular occurrence.
These AI hallucinations have invalidated the entire draft, which is why the Minister of Communications and Digital Technologies decided to redo the bill. This once again demonstrates that AI should always be used with human oversight, and not just taken as fact.
“I want to reassure the country that we are treating this matter with the gravity it deserves. There will be consequence management for those responsible for drafting and quality assurance,” the minister said.
An important lesson in why you shouldn’t trust everything AI says
While AI can be very useful, making your everyday tasks easier, it’s important to double-check what it is outputting. We’ve seen the memes of asking Gemini if 2027 is next year, only for it to say it’s not and give out a very complicated answer about why it’s not. And that’s just a simple question, never mind the complicated approach of putting together AI regulations.
Even with using AI for research, it’s still important to double-check, especially if something sounds fishy. While many AI chatbots do provide sources, that doesn’t mean those sources are accurate, or even exist. As we’re seeing here with the AI Regulations, South Africa wanted to put out.
The post South Africa’s AI bill was written by AI, and it hallucinated its own sources appeared first on Android Headlines.
