“Write a column in the style of Anson Cameron about a column written by AI.” After an unnervingly brief wait there is this: “The typewriter’s clatter has given way to the keyboard’s quiet clicks, and now, in an astonishing leap, to the silent hum of artificial intelligence. Columnists, those brave souls who have long wielded words like swords, now find themselves sharing space with an entirely new entity: the AI columnist.
A marvel of modern technology, it is as mysterious as it is controversial. ” I can’t tell if that bit about columnists being brave souls and wielding words like swords is a piss-take or plain old bombast. If it is a piss-take, is it AI mimicking my own satirical voice and using it against me? Can AI detect satire and counterfeit it? Is it writing in my style, as requested, and making fun of columnists? And how can AI recognise and understand satire when so few humans, and virtually no sub-editors, can? Credit: Robin Cowcher It goes on: “A great columnist doesn’t merely inform; they resonate.
They tap into the shared consciousness, draw out our collective hopes, fears, and dreams, and give them voice.” That’s not bad stuff. I like to resonate, tap in, draw out.
It’s a nice enough aspirational statement. What do you want your column to do, Mister Cameron? See above – re the collective dreams, etc. In the end the AI column comes down on the side of AI columnists.
“So, let’s not fear this silicon scribe ...
In this fusion of man and machine we might just find a new way to tell our stories.” It suggests a sort of detente whereby AI does most of the hard yards of research and fact-finding and I inject the heart and soul of the piece. That we be a sort of algorithmic/organic partnership – a fleshy Flanders twinned with a cyborg Swann.
But why did AI make a case for the AI columnist? After all, it is, in this case, just a simulation of me, a man who remains unconvinced AI columnists are a great idea. It is merely a machine committing glossolalia on command, isn’t it? So why weigh in on AI’s side? It could just as well have written a column denouncing the “silicon scribe” as a bloodless automaton filching ideas and aphorisms from long-forgotten salons and far-flung think-tanks. Why did it make an argument for itself and claim a future as a column writer? Is self-interest written into AI’s code? Is its code evolving now, with greed and pride as adaptive, advantageous traits? Will the AI columnist become braver in a year and denounce my place in this partnership? Will it not only replace me as a columnist but also sell my family and steal my dog? Is this my first sighting of a new civilisation, white sails in the sun edging over the horizon to undo my world? Not even AI’s greatest pioneers can agree on where this will end up.
Whether AI is a new, more advanced life form that poses an existential risk to humanity - or merely another subservient friend, a 21st-century blender, vacuum cleaner or transistor radio. As a scribe, AI can certainly scour the net for information and attitude and thereby, probably, I think, write a good essay. And it can write a column in the voice of a fictional persona.
But that’s mere performance. If a reader is minded to suspend disbelief, AI can tell a faux-intimate tale. But this is called fiction.
And columns aren’t that. AI can’t honestly personalise its writing. It can’t, ever, write, “I visited my mother this morning – and she didn’t know who I was,” and have that sentence contain the loss and sadness it contains when I write it and you read it knowing me as .
.. as what? Who am I to you? To be honest, you can’t know if that sentence about my mother is true when you read it.
For it to work you have to trust me. You have to believe me for it to resonate emotionally with you. So I guess what a columnist has with a reader is a kind of distant intimacy, made possible because faith runs between us.
I will offer you honest moments of fear, grief, joy and laughter, accepting that you will respond to them knowing they are that. I guess, then, what you and I have is a kind of friendship. And I don’t know how AI can imitate that, unless it’s able to subvert the belief in truth itself.
.
