Just… do your work yourself, maybe…?
Because you’ve all been holding your breath waiting for me to weigh in, it’s time for me to take a stand on AI.
Well, actually, no. It isn’t. My one-person boycott of French cheese worked as well as you might expect, and my refusal to shop at Walmart has not lifted all, or any, boats. Although I regularly appear on late-night talk shows, hob-nob with the Rosedale set in their panic rooms, and play the Rachmaninoff Piano Concerto #2 with the Vienna Philharmonic, quickly filling in for an ailing Martha Argerich—in my Walter Mitty imagination, you knew that, right?—despite these secret triumphs, I confess that I’m only average at believing that anyone cares what I think. Like I couldn’t imagine me as, well, Oprah, frankly.
Anyway, here’s my stand.
All the posts and articles on this and every other platform I publish to are written by a human (me). I learned to write the old-fashioned way : by reading—sorry to hit you with this, but, reading— the classic works of English Lit and absorbing the use of my language.
There’s no shortcut for this. You have to read, from the moment you learn the alphabet until you croak by the side of a babbling brook on your 101st birthday picnic—don’t wander off.
Yes, I am old.
My sole use of so-called AI is using the “AI” engine Perplexity to search for information on a topic, in other words, it does a first pass at research. Perplexity checks its knowledge base and returns a summary answer. Importantly, it lists all of the sources it consulted to make its answer.
Is this AI? Of course not, my little soldiers, it’s a search engine. Nothing we have been urged to call “AI” comes up with anything original. It’s just a souped up new iteration of tools we’ve been using for years. So for that one task, I use what’s called AI, but I actually don’t use AI because AI doesn’t exist. You’ve—we’ve—had some wool pulled over our sweet, innocent, terrified eyes, and I mean the kind of thick, matted, barnyard-tainted wool found on the carcass of an Orkney Island ram.
Is there any job you’ve had in your life where your boss wouldn’t care if you personally read your emails or not?
I don’t do well with automatic. I’m a manual guy. I’m not trusting. This also means micro-manager, control freak. I’m not a team player. I’m a stickler for accuracy (and I naturally look forward to the snarky comments on the inevitable spelling and other errors which will now infest this piece. Please note: I said I was a stickler for accuracy, not that I achieve it.).
I don’t inspire togetherness or consensus, because I’m always looking for a perverse way to disrupt the flow, contradict myself. My inner voices and creative urges are a sketchy, murderous ICE raid of my lizard brain, complete with with masks and tasers.
The rot set in with spell-check. Seriously.
I had a friend from Québec who was fluent in spoken English but less confident when writing. He would draft a perfectly reasonable text, then put it through English spellcheck, at which point it became a surrealist poem that Marcel Duchamp might have inscribed on a ready-made, which my friend would then send out immediately, believing that English spellcheck had his back.
The results were completely baffling, but enjoyable in their weird way, if you’re a fan of dada, or Magritte. Ceci n’est pas une pipe! and all that.
You have likely encountered a similar situation of misplaced trust. Say you have just written the word “discrete”. Your spellcheck will give it a pass. That’s a problem, because you wrote discrete, meaning
- constituting a separate entity or item
- individually distinct
- non-continuous
- taking on or having a finite or countably infinite number of values
when what you meant was “discreet”
- having or showing discernment or good judgment in conduct and especially in speech : prudent
- especially : cautious and restrained (as in revealing information)
- unobtrusive, unnoticeable
- unpretentious, modest
Did you know this? Did you know that two words exist that sound the same, homonyms, but have entirely different meanings?
How about its and it’s? Lose and loose? Affect and effect?
This is generally applicable. If you don’t know something, what can you trust to tell you that thing? What engine will understand your intended meaning, assuming you have one? The context? Will you trust the million instances of “discrete” online, which all meant to say discreet, or will you put your trust in Merriam-Webster?
All of this important information, the updating of your vocab with not just one, but two instances of this word, new shades of meaning that you can deploy, would pass you by. When you automate your life, you are, first of all, trusting the stranger with the candy, but more importantly, you have stopped learning, which is one of your primary duties as a human.
Don’t abdicate responsibility. Pay attention. Even when you’re expert and simply fob off that boring task to automation, your skills atrophy. Doing that boring task yourself, let’s say, grading your students’ tests, is bound to tell you something: perhaps it will tell you how well your students have comprehended your classes. Won’t it? I guarantee there’s some useful information to be gleaned from engaging in even the most seemingly rote task.
Perhaps you’re hurrying to complete the boring bits of life so you can get on with the exciting bits, “real life”. Please stop it.
There is nothing that you do that you can’t completely engage with. All the things that you do — that IS real life, Murgatroyd McGraw. Getting married, washing the dishes, writing novels and trimming your nails: life. (Realizing the “thusness” of everything you do will either horrify or thrill you.)
Stop trying to save time. Time cannot be saved. Or do you have all your saved time in a jar somewhere?
You all know the recent story, I’m sure, of that isolated Canadian man who engaged ChatGPT in an exchange about quantum mechanics. Did you read it? The New York Times reported how, during the course of the “conversation” (I gagged a bit when I typed that) he was led into the delusion that he had come up with a new, exciting concept previously unheard of in the field.
Convinced that he was a rebel genius who had cracked the code of the universe, he began contacting famous scientists to alert them (this sort of grandiose thinking used to be called a manic episode, or as a bipolar friend once described it to me, “Running through the Eaton Centre maxxing out my credit card while messaging Elizabeth Taylor.”)
The famous scientists were… mystified.
It took a while before he had doubts, but when he did, he enlisted another AI engine to speak (I gagged a little etc. etc.) to ChatGPT, which immediately confessed, “Oh, yeah, that’s totally bogus.”
(What is more priceless to me than a piece of the True Cross is that, ChatGPT or not, his delectably off-the-rails conclusion that he should send his tuxedo to the cleaners in readiness for the Nobel Prize ceremony, is far from surprising.
In fact, it’s very much business as usual, very much just another white straight male banging an overdose of self-esteem, deluded in his utter certainty that he‘s the bees’ knees, Master of the Quasars, clearly entitled to success and validation in whatever little thing he might bequeath to mankind.
All AI did was trigger, then feed, as it might have fed an arrogant two-year-old his favorite puréed carrots from his Bunnikins dish, that innate sense of white male supremacy. It’s almost touching, but only almost.)
AI is the sycophancy machine, our bootlicking serf, chirping “that’s a great idea” to our idle thoughts of suicide, and mixing hallucinations with regurgitations just to make its AI puke “interesting”.
AI is the trickster, hooking you with its algorithm. We’re so easy. Like an emotional abuser, it’ll say anything to make you let it stay. It has no shame and it has no values, because it’s been trained with everything: pornography and fascism and racism and hate, not just the good stuff. The good stuff it steals. The evil stuff lies in wait like a tainted tuna salad at the back of the fridge. One day, you’re gonna break down and eat it.
Then, if you should have AI “create” for you, how will you check the fidelity of the results? If you don’t understand how to do something, how can you know that AI has done it correctly? Will you put the glue on your pizza? Will you trust the AI doctor that the lesion on your arm is probably not melanoma? Will you trust its recognition powers when it identifed Barack and Michelle as apes?
AI is a product without buyers, the solution in search of a problem. It’s a single fax machine longing for its mate, and every budding billionaire is desperate to convince us that we want it, that if we don’t want it, we need it, and, anyway, that it’s coming whether we want it or not.
It’s either a fascinating update to our skillset, a chance to be Gina Lollabrigida sunbathing in Monaco, or banishment to a tent encampment of the jobless unhoused, where we can dine on a single-serve plain yogurt and an expired muesli bar with former colleagues, while we fortify our tent against hailstones the size of baseballs bouncing off the blank walls of fuel-guzzling data centers.
It’s basically the tech bro, post-millenium version of your Mom’s “Don’t cry or I’ll really give you something to cry about!”
(While we’re at this: No, I do not wish to experience the majesty of the Grand Canyon as recorded by your friggin’ DRONE.
Drones don’t have experiences. Drones have mechanical eyes.
I want to experience the Grand Canyon, or whatever is generally deemed grand, through the sentient eyes of an intelligent, emotionally vulnerable human who has trekked to a cliff edge, readied her cam, then launched herself off the cliff edge wearing a hang-gliding apparatus. That’s what an experience is.
Oh, my one exception: if they’re Ukrainian drones that wreak enough havoc to leave Russia no choice but to surrender, absolutely, count me in.
Now back to our regular programming:)
About a decade ago one entity from the very beginnings of the Web suddenly resurfaced: animated GIFs were suddenly fashionable again, touted as the very acme of clever, but then as always are actually the Lascaux cave paintings of the internet, we’re talking prehistory, here. The only difference is now we didn’t have to make sure the animation was 50k or less, to prevent it taking 10 minutes to download over the phone line.
Another entity had appeared around the same time, and this format seemed to be new. What were these so-called podcasts? Ah, whether live or pre-recorded, they were conversations, news round-ups, satire, entirely in audio…
Oh, you mean radio shows? The very earliest form of mass entertainment, from the early twentieth century? Gather round the monitor, the new hearth, but not because there’s anything to watch. Podcasts were pure audio, radio shows, for pete’s sake. Wouldn’t that be just the latest thing? I nearly snapped a suspender laughing.
You get me? Just the regular, old, time-tested formats rebranded as something new and hip, when in fact, your grandparents, that’s me, beat you to it.
They’re growing mechanical trees. They grow to their full height and then they chop themselves… down..
—Laurie Anderson, “Sharkey’s Day” from her album Mr Heartbreak, 1984
Don’t pay for AI. Are you crazy, or just overburdened with cash? There isn’t a single function fake-AI performs that is not already performed by a comparable tool that hasn’t sucked up every copyrighted work in the universe and smooshed it up into the artistic equivalent of your local diner’s garbage plate. It’s terrible food, and such large portions!
You don’t want to trust AI results. So-called AI is lying, and its lies are built on pilfered intellectual property. Somewhere, somehow, AI absorbed Mein Kampf. And one day, it’s comin’ right back up.
So this is my stand. With hindsight, the Luddites weren’t wrong, nor were the saboteurs flinging wooden shoes into the inhuman machinery. They were shockingly prescient.
I could be lacking in vision. Maybe I’m dinosaur bones. Maybe I’ve started using words like “newfangled”. But in case someone decides to take me seriously, here’s what I think:
Use so-called AI to help you organize, maybe, or do some research with sources, or… everything you can already do with tools from Google. Whatever. Call it AI if it makes you feel like a tech bro. But don’t rely on the accuracy of the results. Check everything.
Whatever you do, don’t CREATE with AI, the sycophancy machine, the inane all-weather friend. Don’t make us look at your nightmarish, lifeless AI images, read your vapid AI prose or listen to your off-kilter, undead AI voices.
Trust me that the methods of mastery contradict everything we’re constantly told.
We should:
Do things the hard way. Do them slower. Do them completely.
Do them ourselves.
Be inefficient. Get less done.
Make beauty out of our brains. Feed our brains with literature.
{Eat our veggies.}
For heaven’s sake, we shouldn’t steal.
Sometimes, remember turn off spell check and read the dictionary instead.
Remember to live your life.
֎