Writing is Also Artificial

Last week I was invited to speak at an event called Artifice: Questions of Language, Technology, and Devotion in Contemporary Classrooms hosted by a program called Teaching Religion in Public (TRiP) at Indiana University’s Center for Religion & the Human. It was hosted and organized by Constance Furey and Hannah Garvey (who did such a brilliant job curating the event, down to the smallest details). The impetus for the conversation was to think about AI as artifice, and to think about artifice more broadly. I had many good conversations, with many interesting people. And I’m left with a lot to think about. I’m especially struck by the fact that, while I’ve been thinking quite a bit about technology and devotion, over the last couple of years, I haven’t been thinking about language. Not really, or not explicitly. I plan to work out some thoughts about that, over the next couple of weeks. But, for now, I wanted to share with you an abridged and lightly edited version of the talk I gave.

FOR MANY PEOPLE, and I would include myself among them, OpenAI’s 2022 release of ChatGPT drove them back toward older cultural forms—writing chief among them—as refuge from the artifice of AI. Writing created by real, living, human people has felt like something authentic as opposed to the artifice of AI. It’s led people back to their humanist roots. It’s sent them ad fontes, or back to the sources that AI has gobbled up and digested as its data, back towards the physical matter of the printed book, back towards old texts written by human hands, towards practices of writing with our own hands, on paper that we can smell and touch with our animal bodies. In this emerging intellectual environment, writing and reading can feel like very authentic and human things to do, against the anti-humanist methods of an AI (which doesn’t really go ad fontes but instead chews sources up and makes them into a kind of hamburger for the mind).

I will admit that I’ve found myself, since November of 2022, more deeply invested in writing and reading than I’ve ever been. I’ve always considered myself, above anything else, a writer. So, I’ve always cared about writing. But I’ve found myself becoming profoundly emotional about things I never thought I cared about. I have shed actual tears for The College Essay. It’s a cultural product that (as a college professor) I’ve often resented. Nothing fills me with as much dread as a massive stack of college essays that I’m charged with grading on a deadline. It’s really one of the worst feelings because you know that you won’t be able to fight off the boredom inevitably embedded into so many of them. And yet, I’ve found myself literally grieving and weeping over this hated thing. Because I have felt, on some level, as if it’s a thing that keeps me rooted in a world, and a set of practices. And it’s been one reliable tool I can use to get to know my students, as real and living humans. At best, it can be a window into the way that they’re processing the course, internally. That’s valuable.

But, at least for a moment, I want to ask some critical questions about this reaction. To be clear, I am very critical of what AI is, and how it’s been rolled out in our society. I maintain a lot of trepidation about how it will inevitably be used (and is already being used) to amplify the ongoing enshittification of our educational and democratic political systems. And I want to be part of whatever movement there is to preserve reading and writing practices as essential to these systems, rather than watch them become luxury or boutique educational products that only a few learn to employ well. But I’m beginning to think that if we genuinely want to preserve some cultural centrality for writing and reading, we might actually be better off if we regard these as the (ambivalent) works of artifice that they are.

Writing and reading are technological products that we have figured out, in essence, how to hack for human (rather than antihuman) purposes. This wasn’t inevitable, and that achievement can show us a lot about who we are.

IT WAS LAST SUMMER, I believe, when I started to think a little differently about this whole matter. I’d spent a couple of years immersed in copious, perhaps even excessive, negative feelings about AI. I wrote bitter things about it on Substack (which, I’d discovered, had become a treasure trove of commentary and debate about AI). And I spoke about it, resentfully, with anyone who would listen to me in my actual life. I worried about books. I worried over the practice of writing, both for myself and for my students. And then I read an essay, on Substack, that began to shift my thinking: Jac Mullen ’s “How to Do Soul-Craft With State Tools”.

In it, Mullen is speaking directly to people like me. “We seem to be grieving literacy in public lately,” he writes. So much is at stake in this grief, he notes: our memory, our sense of interiority, even the sense of self that we’ve learned (in our literate culture) to develop. He understands the sorrow. But Mullen’s deeper concern is that, as he puts it, “grief alone cannot safeguard what it mourns.” If we want to safeguard something we have to understand what it actually is, where it comes from, what its roots look like, and why it appears to be so endangered in the first place.

What Mullen offers, in that essay, is a sort of ecological history of writing. He recounts the fact that writing itself began as a state tool. It was used, administratively, to keep track of economic records. It was part of a method that states used to enrich themselves. It took centuries, and waves of cultural and political change, for anything like mass literacy to develop. And all of this helps him make the point that writing is fragile, and unnatural.

We find ourselves, now, at a similar or resonant sort of crossroads. It’s one where AI is rendering us vulnerable to state power and control in new ways. And it’s genuinely disconcerting. We are vulnerable to the surveillance that it makes possible, and to the economic destabilization that it can also make possible. But (as I read him) Mullen sees, in the transformation of literacy from a method of state coercion to something that can shape interiority itself in incredibly creative (we might even say transcendent) ways, a kind of parable that offers hope for our “cognitive commons”. Writing can remind us how brilliant humans can be at hacking the systems that threaten to dominate and control them.

This ecological view of literacy, and the sort of frame it offered for the practices of reading and writing, helped to ease some of my bitterness. It gave me a little more hope for our cognitive commons, that not all would be lost. Perhaps this was, in part, because it helped me to denaturalize, or even deromanticize, writing. While it’s a technology that might create a particular sort of human it isn’t, itself, the foundation or fountain of that humanness. And, as a technology, writing too has had its own ambivalent effects on human cultures and societies. In creating the conditions of possibility for one sort of (literate) human, it also helped to erode the conditions of possibility for other (more oral) ways of being human. I’m not saying that this is good or bad. It just is. This ambivalent effect becomes easier to see, or recall, when we study the historical emergence of writing.

Writing is an incredible technological system for sharing signs and symbols, and so for shaping human consciousness. It’s a technology without which I would not feel like myself. I’m entirely dependent upon it for a sense of identity (and so, we could say, survival) at this point in my life and career, if I’m being honest. We could say the same thing about American culture, at this point in its own history. There are, as commentators have increasingly been observing, elements of our democratic culture that fall apart (that are arguably already falling apart) with literacy’s loss of centrality. It impacts far more than our educational system. It changes participation in the civic sphere—the democratic system—itself.

But writing can also be a technology for abstracting, or detaching, ourselves from the physical world we inhabit. Writing has helped people like me become very comfortable in isolation, often to their own detriment. It’s a technology that’s allowed generation after generation of monastically-inclined writers and scholars shut themselves off and commune with the dead. Personally, I love this form of solitude. I feed on it. And I will fight for anyone who wants to make it part of their life. But I will also concede that this sort of isolation has arguably laid the groundwork for us to engage with our computers and phones in the almost monastically isolated trance that I’m sure you are all familiar with. And it’s probably not unrelated to the perception that writing is an elite activity, most accessible to those who can afford that form of solitude.

If we want to protect the practice of writing, and if we want it to remain culturally central, I think that its ambivalence needs to remain at the front of our minds. Writing is something that I want to protect. But it’s not a neutral good. It’s not morally superior to technology, it is a technology. It’s something we must, as with all things (especially technology) engage with critically. Especially if we want to protect, or even revive it. The future of writing shouldn’t look exactly like its past. It has to live in a new world, and it needs a new kind of steward or defender. I think it needs to inhabit our cognitive commons in ways that it hasn’t quite, at least not yet.

IT’S NOT INCIDENTAL that writing practices can feel like a site of resistance, or a way to escape from the proliferation of generative AI. But it’s not because writing itself is more natural, or more human. Writing is also a practice that calls us into artifice. Rather, it’s because LLMs like ChatGPT offer us what I think Marshall McLuhan might describe as a useable hybrid of old and new technological forms. It’s because LLMs so seamlessly (almost magically) appear to produce what it is we humans produce, when we engage in the practice of writing. It’s because it’s not actually all that easy (despite the many claims and rumors) to make out that hard and firm line between writing produced by a human and the writing produced by an LLM. What we see, in the LLM, is arguably a hybrid form of human and artificial intelligence. It’s the meeting point of two modes of cognition.

We might experience the page that we write on (whether that be paper or a screen) and the prompting system of the LLM as two distinct media. The page is flat, blank, open, and high in friction (its openness calls us into it, but it’s only filled through our difficult labor). The prompt is engaging, low friction, and already feels full of data and information (it can make our labor appear superlative or unnecessary, as we know from so many economic predictions). We engage with the page, cognitively, in a different manner than the way we engage with the prompt. We might feel like a different sort of being, as we face into each medium.

At the same time, the prompt wouldn’t be what it is if the LLM hadn’t consumed billions of pages. And it needs native page producers to engage with it, so that it has more data. It’s dependent upon the page in that way. So, in this sense, the LLM is more of a hybrid form of media. The LLM isn’t just a prompter that contrasts with the page, it’s a page-prompter. McLuhan argued that the meeting point of media forms can be explosive. This encounter can “breed [a] furious release of energy and change,” as he put it. This can obviously be highly destructive, and even terrifying. It can destroy worlds (as literacy eroded oral cultures). But McLuhan also argued that this explosivity is ambivalent. It can also be revealing. “The hybrid meeting of two media is a moment of truth and revelation from which new form is born,” he wrote.

Crucially, perhaps, in order for this hybrid encounter of media to be revealing (rather than simply destructive) we may need to maintain an awareness of how the old medium is being rendered, by the new technology, into an art form. “Every new technology creates an environment that is itself regarded as corrupt and degrading,” McLuhan wrote. And, in so doing, it “turns its predecessor into an art form”. As art forms, they become “anti-environments” or “counter-environments” to the emerging technology that “provide us with the means of perceiving the environment itself”. As art, they become a form of resistance to the emerging technology, even as they are incorporated into the new technology.

The LLM has created an environment that is, increasingly I think, regarded as morally corrupt and degrading. In the face of this, writing (even the worst of it) now appears to many as an art form. I’m sure you’ve heard the celebrations, or perhaps you’ve even celebrated, bad writing in new ways. I’ve heard many of my academic colleagues claim that they even encourage their students to produce bad writing, because it’s a reminder of what a human actually sounds like. Bad writing feels, to them, like a form of relief. Any writing that takes place between a human and a page (even the most degraded and unreadable) now appears as a work of art.

I worry that a perspective like this threatens to make writing itself too precious. It risks becoming less of an art form to practice, and more of an over-valued consumer commodity. As a commodity, this bad form of writing becomes a precious relic that we can aestheticize and marvel over as a window into our human past. This sort of art, McLuhan argues, is “ludicrous and snobbish”. But when we remember that writing itself has always been a technology (one that many of us have learned to use artfully) to accomplish specific things, I think the perspective shifts. It becomes more possible, I think, to consider it a form of art that is what McLuhan describes as “a means of training perception”. This is art, as what he also calls “a radar” for “the function of indispensable perceptual training” rather than art as “a privileged diet for the elite”.

Writing, beautiful though it may be, may ultimately be of less social and cultural value to us as something to merely consume and more valuable as an old technology that can be used to help train our perception in a new technological environment. Writing as a form of art that trains our perception has the ability to help us think prophetically.

ART, FOR McLuhan, had the ability to serve an almost prophetic function. He believed that this sort of art (which, for him, was really nothing more than a means and method for training perception) had the capacity to anticipate future social and technological developments. And, as something older and more established than the latest form of technology (as something that might have become art only because of the newest form of technology), he believed that art also had the more conservative capacity to remind us of the “futility of changing our goals as often as we change our technology”.

I’m highlighting the almost religious (the prophetic) function of art, such as the practice of writing, because I think that religious traditions, perhaps more than any other archive that houses reserves of text, can show us how writing has functioned as a social technology not merely among elites but across social classes and across great stretches of time and space. I think that religious traditions (especially, but not exclusively, those that hold words and text to be divine in some shape or form) offer us a window into what it looks like when writing becomes something that shapes and trains perception.

As I see it, religious traditions have been experimenting with the technology of writing for a long time (for centuries and centuries). They’ve done incredible and terrible things with it. And I think that record can be of value for the world at-large today, in new ways.

In religious traditions we are offered a glimpse at various ways in which writing itself (and reading) has been preserved as a sacred practice. They show us (reading the Bardo Thodol over the bodies of the dead, for instance) how texts can cross the lines between life and death. Religious traditions show us how writing has functioned as a technology of power, control, and domination. Religious traditions have had to navigate the tension between using writing as a state tool versus using writing as soul-craft over and over again. For centuries. Religious traditions have experimented with democratizing precious technology (like text) and they’ve had to watch the good and the bad things that happen when they do. Religious traditions have worked to use writing to train perception across social classes (beyond the elite.)

The knowledge we might gain (about writing, about religion, about ourselves) in this understanding might be, on balance, ambivalent. But the better we can understand writing in its ambivalence, the better we will be able to understand what it does to, and for, us. The better we understand these ambivalent effects of writing, the better we will be able to sense and predict the ambivalent effects of the technology that’s consumed it. And that claims, now, to be producing it. The more we can see and understand this ambivalence, the more we can maintain a sense of agency, I think. And the more we can maintain a sense of agency, the more we can preserve a sense of interiority that feels recognizable. Ultimately, it was centuries of good writing that taught me, at least, to understand ambivalent things as a little more true.