TOM EATON | AI policy seems out to make the end of humanity controlled and sustainable

When it comes to the recent fiasco around South Africa’s new draft legislation on AI the jokes write themselves, mainly because all of the writers have given up and gone into public relations and everything is now being produced by ChatGPT.

At the very least, the apparent decision by staffers at the department of communications & digital technologies to get AI to spit out its policy for it, replete with fabricated references, shows how well the government of national unity (GNU) is bonding: for years, work-shy know-nothings getting paid to half-arse their jobs was the preserve of the ANC, but the fact that this has happened under the DA’s Solly Malatsi suggests that there is more common ground within the GNU than we might have imagined.

Malatsi’s retraction and not-apology were also vintage ANC, as he implied that the real problem had been lack of human oversight and not, you know, attempted fraud, before promising that there would be “consequence management for those responsible” for drafting and signing off on the document. I don’t know exactly what consequence management is, but I suspect it sits on the scale of unaccountability somewhere between consequence suffering, consequence massaging and consequence evaporation.

To be clear, I’m not cross with any staffers who may have used AI to produce or embellish the draft policy document. After all, AI has been created in the image of politics as a thing that copies copies of copies with ever-diminishing returns, lies without flinching and tells you whatever it thinks you want to hear as long as you keep paying for it. I don’t blame those staffers for believing they had found their digital soulmate and they could trust it to do their jobs for them.

No, to me the startling part isn’t how the draft legislation came to exist but rather that it was necessary at all ― that our collective surrender to AI and the utterly monstrous future it threatens us with is now so far advanced that governments have to rush out policies on how to make sure it ends us at a sustainable and politically acceptable rate.

Of course, as a writer I concede that I’m a little bitter, though mostly what I feel is confused. For example, if the theft of other people’s intellectual property and livelihoods is now not only legal but integral to the global economy, then what exactly are the rules for someone like me? How many books do I have to plagiarise before I break free of the gravity of the law and ascend to become a captain of industry? A thousand? A million?

Wherever we end up, though, Malatsi and his department have shown us exactly where we are right now: in a country almost absurdly unprepared for the future envisioned by AI’s most powerful and dangerous proponents.

And if stealing books is OK, how about cars or houses? I understand the ultimate distinction between being in prison and being on the cover of Forbes is how cheaply you can produce killer robots for the Epstein class, but still, I’d like it in writing because right now I really don’t see the point of a lot of laws.

Fine, perhaps I’m more than slightly bitter. But I should also concede that all of this might change next week if aspiring fascists such as Alex Karp, the owner of Palantir and all of your secrets, get bored with AI and decide to tackle impossible challenges like living forever or travelling at the speed of light or asking a woman one question about herself during a date.

As a relative luddite, I also can’t predict how the technology will develop. So far AI seems to be very good at only two things — making human beings more stupid and revealing the broken souls of many corporate types who, having been trained never to publicly express their belief that the world would be much better if the working poor starved to death, can now say it in a socially acceptable way by enthusing about the cost-saving potential of automation — but perhaps that too will change.

Wherever we end up, though, Malatsi and his department have shown us exactly where we are right now: in a country almost absurdly unprepared for the future envisioned by AI’s most powerful and dangerous proponents.

But perhaps that’s not a bad thing. Perhaps, paradoxically, this bungle, with its skiving employees and stumbling boss, is an accidental act of resistance against the ultra-efficient, antihuman world we’re being told is inevitable.

If my choice is inept humans doing their job badly, or perfect machines carrying out economic eugenics and ultimately genocide on behalf of the 0.01%, I’ll take the rubbish cadres every time.

• Eaton is an Arena Holdings columnist.