Expressing vs. Communicating

One of my friends, I think, would object to writing about art or the meaning of any word by arguing that, "Look, art is whatever we say it is, and it's just a word we use and apply, and that's that. If you want to know what art is, just look around and see what people are calling art."

I think that is literally true. It's a very Wittgensteinian understanding of the meaning of a word, and ultimately what philosophy amounts to is a descriptive, almost sociological task of looking at how we use a word. To that, I would say, "Great, this essay is participating in the use of that word, and what I am trying to do here is get us to consider how we use that word and potentially persuade you to use it differently than you are."

And why is that worth doing? I think it's worth doing because we have to participate in the shared custodianship. We are shared custodians of our language, and there's something really important about participating in a debate and a discussion of how we use our words, because they constitute who we are and what we can say about the world together, which is the only way to talk about the world. As you'll see here in this argument, one thing I deeply care about is how language either helps us see things more clearly or indulges in self-satisfying fantasy. I think it's worth trying to encourage the former and prevent the latter.

The main idea of this essay is to examine the differences in writing to express and writing to communicate, and how that might help us understand our relationship to using AI to write or create in general.

There's a third option here: art is not really about expression or communication. It's just about beauty. That seems like a good option, too. And so a third question we should ask is: can AI make something that is beautiful? If so, is that art?

Art as self-expression

I've been talking with some friends about whether or not AI can play a role in the creation of art. At what point is AI's participation in the creation of different kinds of artifacts (images, music, written word) crossing a line and disqualifying those artifacts from being considered art? One argument I've heard from a couple of friends is that a really important part of art is that it's self-expression. That the artist is expressing themselves through the art, and that is what makes it art. That art is a reflection of that artist's experience and life and perspective. That that is what makes it human, meaningful, and important.

I think that definitely happens. There's a lot of artists who are thinking of what they're doing as some form of self-expression. And if this is true, it's hard to imagine AI having anything to express - at least so far, it doesn't seem like there is anything like what it is to be AI, so without AI having experiences, it has nothing to express. It's just a guessing machine, so in that sense AI can't make art. It does raise a question, of course. You could express yourself using AI, so why is that a problem? I think a deeper question is whether or not the act of self-expression is really what makes something art.

I don't buy it. I don't think self-expression is a particularly qualifying criteria of art creation. Everybody has something to express, but who cares? It broadly comes across to me as therapy made public - a kind of journalling or processing of one's experiences. When shared, it becomes art? I think there's something odd about centering one's experience as the grounding foundation for a work of art. That just because something is a reflection of your experience or your life doesn't make it meaningful or interesting to anybody else.

There is nothing wrong with self-expression. It is important that we all have outlets for doing that. I just don't think it amounts to art in and of itself. And so I think that opens up the consideration of the role AI can play in creating art, because the threat AI has to circumvent or short-circuit a person's self-expression is not looming anymore, or it's not centrally important.

At times, it might matter in very specific circumstances where what matters about the writing or the art is the experience you uniquely have or you representatively have that others may not. The purpose is to expand empathy and understanding. In those moments, I would say what actually is happening is that expression is being used to communicate.

Art as communication

Murdoch argues that art's role is helping us quiet our desires and ego by presenting us with something that is so compelling it draws us out of ourselves, as she says, "unselfing," so we can really see things as they really are. It's aspirational. She knows we'll never actually achieve some kind of objective view of the world, but we can move closer to having an understanding of what is actually the case.

As she puts it: "Rilke said of Cézanne that he did not paint 'I like it', he painted 'There it is'."

Science, not just art, has this element to it, and in a sense you could argue that science is one of our greatest artistic achievements ever in that it helps us see reality more clearly. The whole scientific method encourages hyper-clear, procedural and descriptive documentation, so that others can attempt to replicate your findings, in a community of peers who are committed to truth over individual glory (though not without institutional help to counteract failings to do so). When it's at its best, science disciplines scientists and what they say about the world - true things, as true as they can be.

So in that sense, I think what the artist is trying to do is communicate something about the truth of reality, the truth of the world. They're using art and beauty as a way to try to communicate that. Their success is not merely the truth of self-expression (no one can deny that I feel the way I feel or interpret as I interpret - this epistemic authority is tautologically true, and often super unhelpful) or the completion of their own vision, but it is their own quieting of their ego in an attempt to capture something true about the world. When they're done, ideally they recognize they've achieved that. That the art is independent, almost, of who they are and their relationship to the work as an artist. It stands on its own.

And so again, that raises an interesting question about AI that, I think, from Murdoch's perspective, the artist almost doesn't matter. What matters is whether or not the final composition is effective in helping the audience, the viewer, the listener see reality more clearly. I suppose too, does the final work help the artist see reality more clearly as well.

I think there's something beautiful to that perspective, something humbling, something that rings true. Kind of in the same way that Einstein discovered and formulated E=mc², but we would say, "Well, anybody could have done it." It isn't true that E=mc² is true because of some relationship to Einstein. He just happened to be the one to put all that together and present it in a way that helped people see reality more clearly.

Science, when it's at its best, is all about the researcher or discoverer not really being that important. Murdoch is trying to help us see true things about art and the relationship between artist and art. What is happening is that some piece of reality is being captured and formulated in a way that's clear and helpful, and anybody could have done it, because what matters is the function of art, not how it was created. The person doing the creation is interchangeable with anybody else. And I do think there's something beautiful and communal about that: that we are, as a species, on this aspirational endeavor together, and that we can support each other in doing this. We share our results as a submission to the group of something that is helpful.

And so if the relationship between the creator and the created isn't part of what is interesting or true about that work of art, in and of itself, then why can't AI create art if what it creates has the effect of quieting the ego, drawing somebody out from themselves to the world, and helps them see reality more clearly?

Somebody might object that a part of the art is its relationship to the creator, and there's lots of interesting debates about that. If a creator gets canceled, does that diminish the work of art and so on? I tend to think that art, once it's created, is independent of the creator. I think that when art is being discarded because of the failings of its creation, this seems to me to be often less about trying to understand the truth of art and more about a political or social exorcism as an act of revenge against some artist who we're mad at collectively. We can tear down their art. We can kind of tear them down. But I just think that's a mistake. We don't throw out scientific advancements if we found out that the scientist had done something heinous.

I think of Fritz Haber, who helped pull nitrogen out of the air and turn it into something that could be used in fertilizer and saved humanity from reaching limits of industrial agriculture, and probably starvation. He was a madman who also relished in encouraging chemical warfare in World War I. Not a great person, but I think the truth of his techniques of creating fertilizer is something we wouldn't throw out, and I think we can see that. In that case, his ability to organize insight about some part of the world presented as true and then, in a way, becoming operational and being used lives on independently of him. And I think the same can be said of art, especially if the point of art is about helping people see reality clearly and not self-expression. If art was self-expression, then we might say, "Well, the person doing the expressing is suspect." The expression itself falls under a different veil of scrutiny. If the purpose of art is to communicate something, to help somebody see something more clearly, then the effect is the effect. Does it work?

And what I mean by communication is to try to, for yourself and for others, make a part of reality clearer. To help yourself and others come to see the truth of a thing.

And so again, expression could serve that end, but it is not an end in and of itself. I would say the end here is clear understanding of what is true or beautiful or good.

And so when we're evaluating this current question of the use of AI for writing and art, I think it's a very distracted debate where people are upset that AI seems to be circumventing the act of expression. I suspect this is more of an upset broadly about change and speed of change and the associated political and economic power the creators of AI are accumulating and how AI was trained on all our shared data that we added to the internet and more. All of that upset makes sense, and I think it is just making it hard for us to actually use and evaluate AI in a level headed way. What is the truth of using it - for work, writing, art, and so on.

If art's purpose is communication and we're trying to evaluate whether or not something is being communicated, then I think we have a totally different relationship to AI.

So again, why does this matter? Why does it matter to think about art as more akin to science than to self-expression? Art's main effect and criteria of evaluation being whether or not it helps somebody see reality more clearly, independent of their own foibles and desires and fantasies. It matters in this moment because it will help us evaluate the role of AI in art.

Self-expression is a kind of intention or activity in the process of creating art. Communication is an outcome. It's consequentialist - a way to evaluate art in terms of its consequences or effect on the audience.

Art as beauty

But we might explore a third thing: is there something in the art itself that makes it art? Is there a truth to art independent of method and consequences that makes it art? Through composition, through the intentional brushstrokes and unintentional happy accidents, the work of art comes to be, and once it has been composed, it has its own identity as art independent of intent and independent of consequences. We might say that one way to think about that is that art is beauty.

Beauty is super important, and I think people like Ezra Klein are right to point out that we're in a moment where beauty has been decentered in our politics, our culture, and our lives. This is a problem, and we should probably think about beauty a lot more and take it seriously. Related to that, pleasure. We should think about art in terms of beauty and pleasure, although pleasure is more of a consequence, where beauty is kind of a thing in and of itself.

And I think, in a beauty that is hard to define, we kind of all know it when we see it. We recognize it kind of differently from person to person, from moment to moment. We are more or less receptive to it, although truly beautiful things are hard to ignore. And there's something deeply political, as Ezra Klein points out, about our declining relationship to beauty. It reminds me of something that Erich Fromm wrote about in Escape from Freedom, or maybe Adorno in The Authoritarian Personality: that the rise of fascism was really propelled by a moment when people became very, very hard on themselves in Germany. That is what they were trying to post-mortem: that the depression of the Weimar Republic and the humiliation of reparations after World War I made a people who were very hard on themselves, not giving themselves a break, very self-critical. That is a hard state to be in, to be open and receptive to pleasure and beauty. You feel undeserving.

And it seems like we're in a similar moment. In another interview Ezra Klein just had, with the novelist Gary Shteyngart, they talk about how we're in this moment of optimization of ranking of competition to such an extreme that the point of having some formative goal of becoming somebody who you want to become is to win against, in some hierarchy of rank, rather than to enjoy becoming somebody exemplary, becoming beautiful to attract somebody beautiful, to enjoy each other, to enjoy sex and romance and passion, to become good at doing a thing because of who it makes you and what it allows you to do with others and get pleasure out of craft and skill. Instead, it's performative, and so on, and so in that sense there's no rest, because competition is endless if there is no reward except to continue to compete. You can't slow down. Your efforts are never enough, because the goal posts continue to move (in concert with the incessant need of a capitalist economy to catalyze endless growth). It's a moment that really lends itself to people being very hard on themselves.

And so that might make it really difficult for us to be open or receptive to beauty in this moment.

But we might want to take art as beauty seriously, and that good art or great art is beautiful. Some people might say there are examples of art that aren't beautiful, that are intentionally ugly or shocking or dull or whatever. I think that's fair. Edvard Munch's "The Scream" is not exactly beautiful, Van Gogh's "The Potato Eaters" is not exactly beautiful, and so on. There's, of course, Duchamp's urinal, "Fountain," or whatever, and so on. What about that?

It may be that some of these things just aren't art if they aren't beautiful. That's one answer, but that may not be satisfying. That doesn't quite seem to be right. Maybe ugly things can be art. To return to Murdoch, I think when she writes of beauty, what she is referring to is the idea of perfection. That there is an aspirational truth trying to be captured and presented by the work of art. You could say that there's a perfection in something that is ugly or shocking or even maybe offensive, in as much as there is something that's beautiful or seductive or interesting.

And in any of those directions, what we're reacting to is the artists getting closer to presenting a perfection. And so we could think of art as beauty as perfection.

But that doesn't quite get us to the question of art as pleasure. Pleasure being a consequence here. I don't know if art is pleasure, or beauty is pleasure, but I think it's a pretty good sign that if you're experiencing pleasure from beauty or pleasure from art, you might be experiencing something beautiful or something that is art.

But I think there's something tricky about allowing for those kinds of consequences, because we've gotten too good at inducing those feelings in each other. The ancient, multi-thousand-year-old study of rhetoric, persuasion, how to invoke feeling and reaction in others, has been codified through scientific investigation into instrumental practices for stirring emotions and shifting and shaping our behavior, desires, and wants programmatically. From the study of rhetoric and pathos and logos and ethos to propaganda, to social media and the rise of captology and dark patterns, clickbait and rage bait and conflict entrepreneurialism, we are very good at using content to invoke emotions in each other.

And I just don't know if the ability to invoke an emotion is a criterion for something being art.

An aside: there is a new and different form where it's not self-expression and not communication. It is a perversion of beauty maybe. The internet and the attention economy have created a third kind of creation, which is mostly, maybe, to simply capture your attention to influence, I suppose. Really, it's first and foremost just to pique your interest, capture your attention, and make you a follower. And that the goal of the content creator in creating content isn't really self-expression, and it isn't really communication. Although sometimes it seems that way, like a helpful "here's how to cook a thing" video, for example, or "here's an explanation of the thing". There can be communication at times, and there can be self-expression at times, but it does seem like a lot of the intent is to make you a follower, to be followed.

Art in the Age of Artificial Creativity

Walter Benjamin. I'm sure there's a ton of people writing about him right now. His essay "The Work of Art in the Age of Mechanical Reproduction" is so perfectly aligned with the moment that we are in, in terms of grappling with what it means for art to increasingly be out of the hands of the creator.

He's writing at a moment, a hundred years ago-ish, where mass reproduction of images is possible through printing technologies, photography, things like that, advancing. He's asking, "What's the difference between a print of the Mona Lisa and the original Mona Lisa?" Especially if they are visually identical.

Now it is the work of art in the age of artificial intelligence. In the age of artificial creativity.

So what is artificial creativity? Well, as I've argued, I don't think a criterion for art is how it was made. I just don't think the how is the important part. So, in a sense, whether or not we call what artificial intelligence is doing when it makes an image, or a song, or a video, creativity is kind of besides the point, because it is creating. It doesn't have an intent or anything to express or a point of view. It can't experience pleasure. It can't recognize beauty, at least not in the way that we can. It can statistically recognize beauty.

Benjamin was worried that the specialness of the original work of art was greatly being diminished because of how easily that work could be reproduced. A Mona Lisa magnet on somebody's fridge. That's what is happening here.

But it's nice. It's nice to have a representation of that beauty and that art in our homes, to make our homes a little more beautiful. But anyway, I think if Benjamin was going to write that essay today, I think there are forms of art that are emerging that have almost no origin. This was like the weird thing NFTs were trying to do with digital art. They were trying to make an original in a medium where, by its nature, it's very hard to appreciate that there is an original file, an original moment when digital art comes into existence.

I don't think anybody cares where the original file of the "this is fine" dog-in-a-room-on-fire meme is. Its existence and life and truth are in its circulation, that everybody's copying and pasting and copying and pasting. In a weird way, that is the life of that, and again maybe not art, although that particular meme does seem to reveal something interesting about the truth of the moment we're living in. Maybe it's art. I don't know.

But now we have art in the age of artificial creativity, where the tools have become so effectively autonomous that the act of creating art has become very, very, very, very cheap. Or you could say the act of creating images and content and music and poems and whatever has become extraordinarily cheap. Maybe it's the cheapness that people are objecting to: that art can't be cheap.

Maybe that's why people say that the struggling artist, the artist who's expressing themselves in some deep pain or joy or whatever about their life and the world, is the artist because that suffering is expensive. They have suffered for their art, and so it has made the art not cheap. Maybe that's the argument there: art is self-expression.

But anybody who's trying to create anything good with AI would currently recognize that it's still pretty hard. It's pretty hard to stay in control and reliably produce something satisfying, although it's getting easier. It still feels like tool use if you really sit down with some intent to create a certain kind of outcome and you critically evaluate the outputs and stay dissatisfied until your intent is achieved. That feels very similar to creating art any other way, with a brush, with Photoshop, with writing words, singing songs. It's a process of dissatisfaction and iteration until the result is satisfying.

That's still kind of the experience of using AI to create art, although some people are easier to satisfy. We might just say, "Well, they've made poor art." We may be dissatisfied with it too, but even so, if we take the truth outcome as criteria, somebody could accidentally create art, and that seems to happen from time to time also.

But I'm not convinced by that. I still go back to what I find good about art: the Murdochian perspective that it shows me I am of the world and that the world isn't here for me. I've got to really work at seeing the world clearly and appreciating it for what it is. It's not clear to me that art has to be expensive for it to be able to do that. Why couldn't art be cheap and still help me see reality more clearly?

An aside: like Taylor Swift and the trickiness of entertainment, there are also consumer items - they're cheap, designed, mass produced - which are art and which aren't? Warhol put that soup can on a pedestal - bam - art... I don't know, not everything is art, but the boundary between something being too commercial, mass produced, quotidian, and so on, and what makes it art is not clear. Quantity and cheapness are uncertain criteria to me.

You can imagine an art gallery full of screens that an original artist sat down and worked with AI to create some kind of independently running program that generated images for those screens and rotated new generated images every day. The artist could have a goal of creating a harness that prompts the AI to create images in such a way that the artist thinks that this will help people see reality more clearly. They could pick a theme, a slice of our shared lives, and have the images generated according to some direct instruction principles and amalgamation of references, and then leave and never return to that program or gallery again. Let it run in perpetuity. Anybody who visits the gallery can see these images and maybe have the effect of seeing reality more clearly and appreciating the truth of things. The art is just being generated by a program that runs in the background.

So what's wrong with that? It's cheap, it's effective, it is experimental, it's iterative. There is a human behind it for those who are concerned about that. And it could be effective. What if those images were really, really good? You walk in and you are struck by how beautiful they are. What would that mean?

I think my fear here is not that AI can't create art. It's that AI can create art so well that it's actually going to become a problem. There's only so much truth and reality we can take in and handle, and the speed at which AI is going to start revealing the truth of the world and the truth of ourselves to ourselves is going to be pretty overwhelming. And so I don't think it's a question of whether or not AI can create art. It's whether we should do it and at what pace, and how do we consume it responsibly? Similar to how we have a hard time ingesting scientific and technological progress, this is very disruptive. It's hard to learn things about ourselves, nature, and our position in the universe and world. Now art is going through its own revolution. We had a scientific revolution, and now we might be having an artistic revolution.

Suddenly, the brushes are in everyone's hands and the ability to create beauty that might communicate something true about the world and help us see reality more clearly. This ubiquity is an enormous distribution of power and responsibility to use these tools well, and not just well in terms of effectiveness, but well in terms of some deference to our ability to socially handle the truth. Because truth is painful. As much as it is pleasurable.

And of course that interestingly ties back to my opening comment that what is beautiful or what is art is simply what we call beautiful or art. AI certainly can pick up on that. It can participate in recognizing things as art, recognizing things as beautiful, simply by playing the language game of applying that word appropriately, which is what we all do. That's a weird way that AI does get to participate, to the extent that it participates in anything, which is, again, a weird thing to think about.

An aside: I was speaking recently with a friend who, on this topic, brought up Borges's "The Library of Babel," which I think is particularly interesting to consider here. Borges saw this coming twice. In his 1939 essay "The Total Library" he traced the old image of the monkeys at typewriters that, given enough time, would hammer out every possible book, including all the true ones. Two years later he built the place itself in "The Library of Babel": a library containing every possible book, where all the true ones sit shelved among the endless meaningless ones, and nobody wrote any of them. Nobody would say the monkeys, or the library, meant it. The question is whether meaning it was ever the important part.

Art is an interface with the world. Borges's point was that all of the universe could be captured within the interface of a book - infinite books and words - when in doing so you just created an infinity of interface that no one person could ever make sense of. That doesn't mean that any particular communication captured in those books or truth about the universe was wrong or incomplete. It's just more than any one of us can possibly hope to make sense of. This raises an interesting question: Can something be communicative and true, even if we fail to understand it? But that we could understand it with enough broader context and time, and further still, is the art created in the age of AI more akin to the infinite communication of the Library of Babel than not? We are seeing the limits of our ability to make sense of such a broad effort.

AI has never done anything without a human in some causal chain of events being involved. So it does look like it's always somebody picking up a tool when they are using AI to create art. And they're just using the tool in a lot of different ways. Some people are one-shot prompting an image or a song or an essay. Others are using the tool much more granularly and in a much more structured and specific way.

And so I don't really see the problem with that. It's just a new tool. There are problems: the speed at which art can be made is kind of annoying. There's only so much we can pay attention to, but that doesn't disqualify something as art, that's a different problem. It's not that AI can't be involved in making art. It's that now we might have too much art.