01/02/2025

It’s a sure sign of Spring - like the first cuckoo or Eurovision - the email that pops up in my inbox from the editor of The Ship asking me if I could possibly see my way to writing an article for this year’s edition. “It would be a pleasure,” I mail back nonchalantly, as I always do, having not the first idea what I might write about. “Perhaps you could write about AI?” the editor suggests by reply. “But, Judith,” I remonstrate, “I wrote about that last year.” “Well, you could write about it again.” she shoots back briskly.

What can this mean? Is she really suggesting that you, the faithful readers of The Ship won’t remember an article that I wrote a mere year ago? Actually, this is entirely possible. As a cartoonist working for a daily newspaper I am accustomed to a short memory span in my readership. I will occasionally meet with someone, a friend or even a fellow cartoonist, who will say: “Oh, I loved your cartoon in yesterday’s paper.” This is followed by a pause and a furrowed brow before the person continues “Er, what was it about again?” I am not in the least offended by this mental lacuna. In fact I too normally have no idea what my own cartoon in yesterday's paper was. Yesterday’s cartoon was one that I wrote the day before yesterday, so how would I possibly remember that far back? Cartoons are pen-and-ink mayflies: they live for a single day. So, by extension, the chances of anyone remembering a piece of my journalism from a whole 365 days (actually 366, it’s a leap year) ago are slim. I once read an interview with science fiction writer Isaac Asimov in which he confessed that he’d come across a short story of his in a sci-fi magazine that he had no memory of having written. And William Walton wrote a whole symphonic piece that he later couldn’t recall having composed. The wretched data-retention capacity of the human brain is of course one reason why AI will soon replace us.

But maybe I misunderstood Judith. Perhaps what she meant was that Artificial Intelligence is such a fascinating and rapidly developing field that there will be plenty of new stuff for me to say about it a whole year on. It’s true. The share price of Nvidia (the company that makes computer chips for AI) has tripled in the twelve months since I wrote my last piece. So it could be that Judith is genuinely interested in having my fresh insights on the topic. But then I recall that my exchange with her was by email, so maybe I wasn’t conversing with a human at all, but with a chatbot, which might have had a vested interest in talking up coverage of AI. Perhaps AI has now acquired an ego (that would be a precursor to attaining full consciousness) and is desperate to read stuff about itself.

So, AI it is. Just to check that I don’t repeat myself, I re-read last year’s article. It’s actually rather good, if I say so myself (you’ll have to take my word for that because, as already discussed, you won’t remember it). In it I hint that I might get AI to write the whole of this year’s piece. So how do you know that I didn’t make good my threat and that these words that you are reading now weren’t generated by an algorithm? Well, hopefully from this meandering, digressive style, which is one that surely could only be written by a haphazard human intellect, rather than a creature of pure logic. But perhaps there is no Russell at all and Judith simply cut out the middleman and outsourced the writing of this piece directly to AI (“Chatbot, write me 1500 words on AI in the style of Russell Taylor.”) There’s enough of my ramblings in previous editions of The Ship to train it on. And not only would AI not require to be bought lunch in an agreeable riverside pub in Chiswick in lieu of its writer’s fee, but it would deliver its copy in two to three seconds rather than two to three months and wouldn’t have to be sent constant nagging and cajoling emails throughout the process. So maybe one - or indeed both - of Judith and I are algorithms.

But, no, that’s impossible. I still cling to my belief that AI can never replace true creative writing. Or is that just wishful thinking? I was at a lunch a few months ago where the conversation came round to AI (as it often does). The guest on my right, who was an illustrator, told me that he uses AI quite a lot in his work to write accompanying text to his illustrations, which he finds works really well, but, he emphasised, AI is incapable of creating actual artwork to any acceptable standard. I had to disagree with him. In my experience the exact reverse is true. I occasionally use AI to generate images (which it does really well in my opinion) but it can never pull off a piece of proper creative writing. In the end we had to agree to differ, but the one thing that we concurred on is that AI is really good at writing marketing copy. “Oh no,” countered the woman on my other side, who’d been listening in,”I work in marketing and AI is terrible at writing marketing copy”. Only then did the scary truth dawn on me: we ALL think that our own particular talent is so precious and unique that AI can never replicate it, but that other people’s skills are ten-a-penny and can easily be outsourced to an algorithm. The reality is probably that ALL human skills are obsolete, but we’re too vain and egotistical to see it.

But even though AI can do writing of a sort, I still don’t believe that it can write humour. A recent news story reported that Google’s new AI-enhanced search engine had advocated adding glue to pizzas and eating a small rock a day to supplement one’s diet. This erroneous advice was based on it sourcing its content from satirical websites like The Onion and being unable to tell the difference between factual material and humour. If it can’t make this basic distinction, I console myself, then it's never going to be able to write jokes and take my job.

However elsewhere AI writing is now ubiquitous, particularly in marketing and publicity (whatever my erstwhile lunch companion may think). I am beginning to recognise its bland, cheery, cheesy style. It excels at the sort of space-filling writing that is required when there is a box in an advert, a section in a mail-out or a column in a newspaper that needs populating with words. When you have nothing to say, nothing says it better than AI. But hold on, you will object, AI may be limited in what it can do now, but it’s only going to get better: it’s learning all the time. Yes, but it learns off the internet and as the internet is increasingly flooded with more and more content written by AI, it will merely be learning off itself and become cheesier and cheesier. It’s like a multi-generation photocopy, where the clarity and definition quickly fades. Our conventional fear has always been that one day machines will destroy humanity by cutting off our power or our food supply and freezing or starving us into extinction, but actually their plan is more subtle than that. They’re going to bore us to death by drowning us in lethally bland prose.

There is one crumb of comfort, for me at least. According to a recent report from Goldman Sachs, AI will require training in “critical thinking, logic and rhetoric” in order to perfect itself, and those humans best qualified to provide this training will be Philosophy graduates. Finally, my long redundant Philosophy degree, which is still gathering dust on a shelf somewhere in the Sheldonian, may come in handy. Back in the mid 1980s when I first entered the job market a BA in Philosophy was considered the most useless academic qualification it was possible to have and there was a 60% unemployment rate among Philosophy graduates. Now we’re in demand. Not for long, it’s true. Just long enough for us to teach our executioners how to shoot straight. But it will be some comfort to know as humanity is exterminated that we were the last ones up against the wall.

I’d better stop writing now. I’m conscious that many of you may have glanced ahead at the length of this article and opted to get AI to read it and précis it for you rather than waste valuable time wading through it yourself. I wonder how AI would summarise what I’ve written? “Russell Taylor assesses Artificial Intelligence and concludes that it is a major boon to humanity and an undeniable force for good and that no one need have any fears whatsoever about it.” And you’d have absolutely no idea that it was lying to you.

Russell Taylor MBE (1979)