ChatGPT; Friend, Foe or Much ado about Nothing?

On the best of days for tech illiterates, interacting with computers is, at the risk of sounding melodramatic, embarrassingly awkward. All some of us are certain of around a desktop is that when you double-click, something opens. “And the *##&$# thing,” an octogenarian friend of mine has been known to obscenely vent, “sometimes doesn’t even do that.” 

Now with the phenomenon that is artificial intelligence text generators (AITGs), the most popular being the much harped-about ChatGPT, the Fourth Industrial Revolution has exceptionally announced itself with a flourish. And it’s worrisome. Like an eerie scene from The Terminator, the machines have risen, or are at least making some big strides to that end. Think about it; tomes of information, literature, research, science, entire passages of history being fed into a machine with a brain so advanced it can churn out almost any text you ask it faster than you are able to top up your whisky. Or power up the PC.

This is a consternation not too dissimilar to that which must have confronted mankind in the twilight years of the industrial revolution where the working class found itself grappling with the inevitability of machines taking over the assembly lines, rendering the blue-collar worker virtually obsolete. Subsequent improvements in machinery would result in bloodbath lay-offs that saw thousands of otherwise able-bodied men and women hurled to the gutter. 

With ChatGPT, history appears to be repeating itself. What only several years ago seemed no more than the exaggerations of science fiction appears to be a real thing, huffing and puffing at the door of virtually every corner of industry.

Understandably, writers, amongst a slew of professions, are sweating bullets over what this supercomputer technology might mean for their humble screeds. 

With many media houses closing up shop, writing gigs have grown increasingly hard to land. More so those that pay. Now here is this cyber Einstein, Bard, Mozart, John the Baptist et al poised to set off the tinderbox and watch the noble man burn. Furthermore, OpenAI, the company that launched ChatGPT, has caught the world on the back foot. As such, amid the ostensible vacuums in legality, shrewd companies are already scrambling to milk the cow for all it’s worth, generating swathes of AI-generated copy, especially those in the advertising space. Students have been found to be breezing their merry way through assignments and essays thanks to the bot. Only a few months old, ChatGPT has already passed the United States Medical Board Licence. 

If you are one of those stubborn laggards who still thinks that no way can a machine have the creativity of a human being, then best brace yourself for a rude awakening. 

ChatGPT is pretty good, some say dangerously so. Far from the rigid, colourless standards one might expect from a computer, the writing is surprisingly fluid, conversational, with hints of a convincing sense of humour. Still not buying it? Well, a certain Dr. Shannon H. Doak asked it for, “10 original jokes, which are not “dad jokes” and which… should be well-thought out and have a good punchline or twist.” This is what he got:

Reporter at the Wall Street Journal, Ben Eisen, put the AI bot’s punning skills to the test by entering the “Punderdome,” a competition in which punsmiths compete before a live audience. By Eisen’s account, the bot sometimes offered up material that was more like “dad jokes”; he “struggled to get anything particularly witty”; although some had potential, others fell flat; but the machine did however manage to think-up this nugget: “What’s the state where common sense is in short supply? Flori-duh.” Not bad, says I but then again, I’d have better luck placing a strategic ace against Novak Djokovic than any shimmering hope at punning.

This seems like the pervasive modus operandi amongst experts though; to test, grill, probe the machine to see how good it really is. What are its limits? Shortcomings? How close is it to getting to, even eclipsing human intelligence? Can it ever attain human empathy? In a beautifully entitled article, “ChatGPT can tell jokes, even write articles. But only humans can detect its fluent bull****,” Guardian columnist Kenan Malik is a little skeptical. Though pointing out ChatGPT’s capabilities, he also notes that “the chatbot that can write grade A essays will also tell you that if one woman can produce one baby in nine months, nine women can produce one baby in one month; that one kilo of beef weighs more than a kilo of compressed air; and that crushed glass is a useful health supplement. It can make up facts and reproduce many of the biases of the human world on which it is trained.”

He goes on to say that there “lies a fundamental problem that faces any form of artificial intelligence. A computer manipulates symbols. Its program specifies a set of rules with which to transform one string of symbols into another, or to recognise statistical patterns. But it does not specify what those symbols or patterns mean. To a computer, meaning is irrelevant. ChatGPT “knows” (much of the time at least) what appears meaningful to humans, but not what is meaningful to itself.”

Bombarded by a host of conflicting verdicts, well I decided to jump on the “assessment” bandwagon. Could the technology do a better job than Sunday Times Magazine satirist Ndumiso Ngcobo, I wondered? Or arguably one of the finest writers in South Africa today, Rian Malan? So I asked it to pen an article on South Africa’s politics in the style of Malan and from the first paragraph, I was far from being sold;

“South African politics has been a tumultuous and complex affair since the end of apartheid. The African National Congress (ANC), which led the struggle against white minority rule, came to power in 1994 with Nelson Mandela as the first black president of the country. Since then, the ANC has won every election, but its grip on power has become increasingly shaky in recent years.” Though it could pass muster as a straightaway piece, it was stolid in prose and lacked that witty, jolting punch so characteristic of Malan.

Then AI’s rendition of a Ngcobo piece on Black, Coloured and White people (Note that this is satire and does not necessarily reflect the views of neither ChatGPT nor eParkeni); “Let’s start with the coloureds, or as they like to call themselves, the “Kleurlinge.” Now, these guys are a special breed. They’re like the mixtape of South Africa – a little bit of this, a little bit of that, and a whole lot of confusion. They’re the only group of people who can simultaneously claim to be black, white, and Indian, and get away with it. It’s like they’re playing a game of “guess my race” with us, and we’re all losing.

Next up, we have the black people, or as they like to call themselves, “Mzansi’s Finest.” These guys are the backbone of our nation. They bring the rhythm, the style, and the flavour to everything they do. Whether it’s music, dance, or sports, they always bring their A-game. And let’s not forget their sense of humour. They’re the kings and queens of slapstick comedy, and they’re not afraid to make fun of themselves or anyone else.

Finally, we have the white people, or as they like to call themselves, “The Rainbow Nation.” These guys are like the vanilla ice cream of South Africa – plain, boring, and lacking in any real flavour. They’re the ones who started Apartheid, but now they’re trying to act like they had nothing to do with it. They’re always talking about “moving forward” and “letting bygones be bygones,” but they’re the ones who benefited the most from the system. It’s like they’re trying to say sorry without actually saying sorry.

In conclusion, we’re all different in our own unique ways, and that’s what makes South Africa such a beautiful and vibrant country. So, whether you’re coloured, black, or white, just remember to embrace your differences and laugh at yourself and others. And if all else fails, just remember the words of the great Nelson Mandela: “I never lose. I either win or learn.” Not quite as sharp and cutting edge as the man himself, but one can’t deny chuckling at more than a few of those.

 Blown away but fast running out of data, I leaped onto OpenArt, “a platform for AI-generated images,” and typed in; “”an upset-looking man sitting at his computer in the study.” In just six seconds, the results are the featured image you see above this article. Paying a negligible amount of data, I had a passable image without so much as picking up a camera. How this technology will eventually impact the creative space is a scary rabbit hole to wade down. No doubt, artists and creatives with a signature edge to their work will not de deposed so easily. But for the moment, please do spare a thought for photographers, journalists, artists, graphic designers and all those others whose career prospects were long skating on thin ice before these technologies were even a very clear and present danger.