search
date/time
Yorkshire Times
Weekend Edition
frontpagebusinessartscarslifestylefamilytravelsportsscitechnaturefictionCartoons
James Goodall
Features Writer
1:00 AM 23rd March 2024
scitech

In Conversation With ChatGPT

 
Image by Alexandra Koch from Pixabay
Image by Alexandra Koch from Pixabay
ChatGPT is a fantastic resource. Ask it any question, and it will answer; ask it to write for any purpose, and it will oblige. Not only does it do the work for you, it does so in seconds and without objection. Never will it say, 'Get stuffed; do your own homework'. Which begs the question: Where was ChatGPT when I was at university, trying to get my dissertation off the ground, with only Google and an under-resourced library at my disposal?

For those of you who have been living under a rock since 2022 and don’t know what ChatGPT is, strap in and pin your ears back.

In its own words (because yes, I asked it to tell me a bit about itself, as if we were speed-dating), ChatGPT ‘engages in conversation with users … and can help with a variety of tasks, including answering questions, generating text, providing explanations, offering suggestions, and much more’.

Whilst researching this article, I invited ChatGPT to speak for itself. At times, it felt like we were in collaboration. Very Isaac Asimov!

'Hi, Chat GPT,' I began tentatively. 'Tell me about you.'

It proceeded to do just that. Not only did it reply, but it did so comprehensively, courteously, and with impeccable grammar and punctuation. How many other correspondences can we honestly say go likewise for us on a day-to-day basis?

I even ended the exchange with a thank you, because not only is it nice to be nice and acknowledge when someone (or something) has done you a good turn, but, come the apocalypse, part of me hopes AI will remember that one human being who was respectful to it in its infancy and spare me!

The Isaac Asimov reference is pertinent, though, for the similarities between ChatGPT and Asimov’s ‘Multivac’ are manifold. To clarify, Multivac was a fictional supercomputer that featured in many of Asimov’s short stories – most famously, perhaps, his personal favourite, The Last Question. In this story, human beings can ask Multivac any question, and it will produce an answer (ring any bells?). The story even opens with Multivac having found a solution to Earth’s energy supply problems.

The comparisons are uncanny, so much so that I even asked ChatGPT if it agreed. Sure enough, it acknowledged ‘share(d) conceptual similarities, particularly in … (their) ability to process vast amounts of information and provide responses or insights based on that data’. It conceded, however, that Multivac ‘represents a more advanced form of artificial intelligence’, whereas ChatGPT has more modest intentions: ‘I aim to simulate intelligent conversation and assist users with their inquiries’. So it seems ChatGPT is content to remain a humble servant for now.

In Asimov’s story, we visit humanity at various stages of its evolution, progressing from a land-based to a space-based race, colonising distant worlds, then leaving corporeal form behind altogether and taking to the stars telepathically. Multivac evolves too, upgrading from a planetary to a universal then cosmic AC, residing in hyperspace. It gets smarter and smarter, achieving godlike status, all the while shepherding the human race like a herd of dim-witted sheep.

Am I saying ChatGPT and humanity will follow a similar course in real life? That would be a tad melodramatic, wouldn’t it? But we’ll see. I’m in equal parts astounded and horrified by what it can do. My tenure as a features writer for the Yorkshire Times may well be in jeopardy. Content creators, editors and proofreaders may well feel alike, because what will be the point in any of us when AI will soon be able to do it all so much better and to a sharper deadline? And if you’re already an avid user, passing off a bot’s work as your own, frankly you deserve to be supplanted come judgement day.

Image by 6557056 from Pixabay
Image by 6557056 from Pixabay
This begs a further question: Does it count as plagiarism if you incorporate ChatGPT-generated content into your work? I would argue yes. You’re still tweaking another creator’s content and fine-tuning it to suit your voice.

Furthermore, utilising ChatGPT in this way ultimately risks us not being able to think for ourselves, and our brains will get lazy. Before we know it, we’ll be a race of clueless dependents, reliant upon a cosmic supercomputer for survival!

So why are there so many innovators out there working on this, pushing the technology forward, stealing fire from heaven, and hastening humanity’s demise? Well, because there’s money to be had, obviously. But also because ‘progress cannot be stopped’, to quote Llorenç Villalonga*.

But we needn’t be afraid of the impending AI takeover just yet. It’s still only a tool. It hasn’t brought about world peace or found a way to stitch up the hole in the ozone layer. It’s also missing those irreplaceable characteristics that typify us as uniquely human. It’s good, but it hasn’t yet learned to imitate us seamlessly. At times, it’s a bit like listening to Mr Data read out one of his cat poems, and I doubt it would pass one of Philip K. Dick’s Voight-Kampff tests. Yes, the content’s there. And if you ask it to write a poem, it will meet all the requirements for lyrical verse. But it will lack that all-important aspect of emotion. In other words, it can imitate but not originate. That said, give it time. If it keeps on learning at the rate it is, it will soon be difficult, if not impossible, for anyone to tell the difference.

But it’s worth remembering that even Multivac had its limitations. Despite all the progress it made, no matter how advanced it became, it could never answer Asimov’s last question: ‘How can the net amount of entropy of the universe be massively decreased?’ It would always reply: ‘Insufficient data for a meaningful answer’.

Well, in a moment of mirth, I decided to ask ChatGPT that very same question. And its response? Well, I half expected it to produce the same answer by way of witty rejoinder. But it was wise to me. It responded as follows:

‘Ah, the timeless “last question” asked of Multivac in Asimov's story! If I were to humorously respond to that question, I might say something like: “Well, if you find a way to reverse entropy on a cosmic scale, I'd suggest starting with a really big vacuum cleaner. Or perhaps a universal reset button labelled ‘Entropy Off’. But until then, I'm afraid entropy tends to have the final say in the universe's affairs. How about we focus on more achievable goals, like perfecting the art of making really good coffee?'

Very, very clever. And not a bad quip to boot. My days truly are numbered!

*Andrea Victrix (1974), published by Fum d’Estampa Press.

Isaac Asimov’s short story The Last Question (1956) is published by Collins (The Complete Short Stories,Vol.1).