Luddite would be a strong word with inaccurate implications, but I have spent the better part of two decades fighting the incessant march of digital technology into every corner of my waking life. Like almost everyone else, I am hooked to the digital world, which I do not enjoy. I hope to become less connected but find it nearly impossible to completely extricate myself from the internet, computers, smartphones and now Ai.
As a rural letter carrier, I am required to keep a digital scanner and a cellphone on me throughout the day. As a writer, the pressure to be online, posting, writing, publishing, and self-publishing is like the grave and barren womb, never satisfied. For writers of Wendell Berry or Cormac McCarthy’s magnitude, there is an ability to be completely divorced from the digital world, but for those of us who have published enough to exist but not enough to exit the rat race, acquiring new publishing opportunities without a digital presence poses challenges.
Regardless, Blood Meridian and The Unsettling of America like this self-published essay will ultimately all find a home in the scrapheap of a generative Ai language model. It is seemingly unavoidable that all writing, regardless of quality or status, will be used to teach Ai.
In 2017 I gave up my smartphone, a device I reluctantly first purchased around 2013. In the wake of that experience I wrote a series of four essays. If interested, they can be read here; Dear Smartphone I Quit Part 1, Dear Smartphone I Quit Part 2, What Is Your Attention Worth and culmination piece, One Year Without A Smartphone. In the years since many other pieces have been written about the plight of smartphone use, some of which even with identical titles. But as we face the encroachment of Ai, an innovation that has already decimated many professional writing careers, including my own, it is as good a time as any to address artificial intelligence.
In 1987, as Wendell Berry aired his grievances towards computers with, Why I Am Not Going To Buy A Computer, President Ronald Regan was sharing his wholesale approval calling the rise of the computer age “The greatest force for the advancement of human freedom the world has ever seen.” It is not hard to find similar quotes today extolling the power of Ai.
Unlike Berry, I often use a computer to write, but I also remember a time before smartphones when I carried a moleskin and a pen in my back pocket. Whether paid or for leisure, for profession or for hobby, the job of writing, simply enough, is writing.
Writers write. Painters paint. Dancers dance. It’s actually that simple, everything else is just that, everything else.
I strongly resent the assertion that I or anybody else could write better using Ai. Writing is the act of acute thinking, and I do not see any sturdy connection between the sort of quick search functionality of Ai and the deeply personal and intensely gratifying act of writing–the clear expression of acute thinking. While Ai may be able to replicated writing, it inherently, and unequivocally voids the very process of writing, the act of wrestling with the blank page that is so essential to writing.
“The future doesn’t need faster humans. It needs deeper ones. So go ahead. Be wild. Be useless. Be gloriously human. That’s the revolution.” -
When somebody has used Ai to write work that is demonstrably better than Berry’s, and when this better is demonstrably attributable to the use of Ai, then I will speak of Ai with a more respectful tone of voice, though I still will not use it.
While I’ve worked as an editor it is not my true vocation–and it’s appeal grows less enticing as the task of editing increasing becomes identifying and improving the recycled work of Ai. How could I write consciously about the working class, culture and art while using a tool that diminishes opportunity for workers, commodifies culture and recycles the creative work of artists without payment or accreditation?
I do not admire the inventors of Ai or their ability to harvest the work of others anymore than I admire the corporations who strip mine coal, clearcut forests and pillage the oceans. Like so many other intrusive technologies, Ai is now unavoidable. In every search, social media post, and text message, Ai has interjected itself into every aspect of our digital life. Like the smartphone and computer before it, it has arrived without invitation, forcing us to opt out where we once would’ve had to opt in. And for this rude intrusion, like sour cream in guacamole and corn syrup in ice cream, we are forced to passively accept or aggressively fight back degradation in all of its forms. For its presumptive intervention into our lives, I will not use Ai.
“My wish simply is to live my life as fully as I can. In both our work and our leisure, I think, we should be so employed. And in our time this means that we must save ourselves from the products that we are asked to buy in order, ultimately, to replace ourselves.” ― Wendell Berry
What would Ai cost me? It’s free for now but like every other technology will either require a subscription with price increases or be overrun with advertisements once users are hooked. Or in the case of Disney, it’ll require a subscription and be riddled with ads. But the real cost is so much more. What is the cost of only receiving information in summary form? Or using the most accessible answer rather than the correct one? What is the price of turning over our decisions to a curator?
The price that our environment pays for even the most banal applications of Ai is catastrophic. Reportedly a single prompt uses about sixteen ounces of fresh water. What is the value of this prompt? What is the value of a summary that could’ve been found just as easily from a Google search or better yet, a book? With the growing climate crisis and water shortages, it is beyond absurd to devour what little resources we have left for the sake of novelty.
The electrical demands of Ai are already straining power grids, increasing the demand for coal, natural gas and ever nuclear power depending on the regions and power grids in question. Giant data centers use water to cool their systems, which has put an additional strain on a world that is already experiencing drought and water shortages. This has also led tech companies to set their sites of fresh water resources like the Great Lakes and other large bodies of water as potential data center sites. Multiple Rust Belt cities, including Flint, Michigan still have toxic drinking water but tech companies have no problem getting their hands on fresh water. For the environmental impact of a service that is absolutely not required, I won’t use Ai.
While employed as a copywriter for a time, I wasn’t surprised by Ai’s clear-cutting of “creative” jobs–most of which are tasked with curation far more than anything resembling real creativity. Every marketing project contains a proof of concept, a look/feel of similar work by other brands and a strategic deck that proves the creative direction, saying something like “Increase market share by becoming the Patagonia of iced tea.” For this type of work, which is largely applying proven approaches to a new category, Ai is sufficient–and admittedly hard to argue. As such, my last salary job and several freelance projects thereafter were lost in what many on LinkedIn referred to as the job–apocalypse.
As the saying goes, the only thing worse than having a job is looking for one. Having to switch directions and find new work is never fun, but truthfully, the demand for content has outgrown human capability, and the switch to Ai content is par for an industry that values ROI over quality or even usefulness. The question remains, who is all of this content for? A human uses Ai to generate vast amounts of content to appease an Ai algorithm with the hopes that it’ll ignite action in humans. This equation barely requires human interaction as is, in the coming months and years, it is sure to stop needing human direction altogether.
Which brings me to my final reason for not using Ai. it is displacing workers while lowering quality. Hiding in plain sight, behind trite statements like “Ai will not replace workers, a worker who uses Ai will,” it has already rendered many of my peers redundant, and it is just getting started. In reality, the billion dollar problem that Ai is being used to resolve is the issue of having to pay wages. The job–apocalypse is merely the first shot over the white collar bow of a capitalistic effort to remove the need for wages–which is to replace labor.
There is no such thing as free labor. Even Ai labor has a cost, although it is a bill that the bourgeois, once again, will not pay for themselves. We are paying it every time Ai scrapes our work, every time we train a language model, and every time we choose an Ai shortcut. Ai is a fence, laundering labor from the working class to the ruling class. Ai illegally obtains work, cleans it like a digital Al Capone, and serves it up as legitimate. By obscuring the origin of the work, Ai launders labor, washing it for corporations to use without repercussion.
It is also not lost on me that when Ai was supposed to render blue collar jobs obsolete culture as a whole called it innovation, when it attacked white collar labor it became a job–apocalypse, a threat to our very humanity. In reality, everyone is ok with the use of Ai for the things they don’t value or do not wish to do themselves, but outraged at the use of Ai for the things they value or wish to pursue. Regardless of your individual preferences, the integration of Ai at any level is the acceptance of Ai at every level, and for that above all else, I will not use Ai.
To make myself as plain as I can, for my love of writing, for the sake of our environment, and in solidarity with workers everywhere, I will not use Ai.
Excellent piece, Noah. As an artist / writer married to an academic, this obviously resonates. Like you, I have already lost quite a few copywriting jobs and I expect to lose more. My wife was laid off from her position as an English professor, along with 60+ other faculty members, last year.
I think a lot of what you write here is self-evident, or at least I would have thought so a year or two ago. However, as with my other bugbear, Spotify, my greatest concern is not the dominance of the shameless tech billionaire philistines who birthed this catastrophe, but the apologists and accelerationists within my own tribe.
Just as I resigned myself as a young punk rocker to the reality that it would be necessary for me and mine to somehow coexist in an uneasy peace with the overwhelming Goliath of corporate entertainment, I assumed the same would be true for things like Spotify and AI: sure, it was the dominant normie paradigm, but my friends and I knew better, and our egalitarian niche of self-determination would sustain us as we continued to seek alternate paths. But as I watch many of the people (and the kind of people) I once respected as allies and peers so readily capitulate to something that seems to me so clearly vulgar, destructive and wrong, I no longer feel any confidence that there are enough of us to sustain such a niche.
I can’t really express the level of disappointment I feel when People Who Should Know Better take this gullible if-you-can’t-beat-‘em approach to what is now occurring. Where is the resistance, Noah?
Basically, it not the existence of things like AI and Spotify that breaks my heart; it is the disillusionment of suddenly feeling like I am alone at the Alamo.
“When somebody has used Ai to write work that is demonstrably better than Berry’s, and when this better is demonstrably attributable to the use of Ai, then I will speak of Ai with a more respectful tone of voice, though I still will not use it.”
The example I often use is John Coltrane.
While I believe AI is still far from being able to passably replicate a Trane solo, I also think it will happen eventually, and maybe sooner than we think. Because of this, I dislike that so much of the critique I've seen of AI from artists thus far has consisted on dunking on the technology’s ineptness and its many uncanny mistakes. This won’t always be true, and when it isn’t, this is when the rubber will hit the road.
If you were to play me a previously unheard Coltrane solo alongside an identical Coltrane solo made with AI, I will always want to hear the former and not the latter. What it *sounds* like is beside the point. Because what I really want to hear is not a particularly excellent tenor saxophone solo—I can hear that anywhere—but the human agency that jazz music, in particular, makes audible. I want to hear the genius brain of a corporeal John Coltrane making decisions in real time and hear this great artist react spontaneously to his environment; I want to hear in his solo what books are on his nightstand, what he had for breakfast, and how much or how little sleep he had the night before. To try to hear this in an otherwise identical Trane solo made by an algorithm would require a suspension of disbelief of which I am incapable.
I have no doubt that soon I won’t be able to determine in a blindfold test the difference between an actual John Coltrane solo and an AI-derived simulation of a John Coltrane solo, and I bet many Coltrane scholars and former Coltrane bandmates and producers won’t be able to, either. The ability to distinguish between the two, however, is not the point, is it?