You might have heard that a fairly large board game publisher recently crowdfunded an entry in a well-loved series using AI-generative illustrations. Now, here at Space-Biff!, it’s my policy to pass on any titles that use generative AI in place of human craftsmanship. There’s some nuance there; I’m talking about image and text generation, not assistive tools, although plenty of people talk about those two things like they’re identical.

But I get that many people may not understand such a stance. So let’s talk. What follows is an explanation of why I’m not interested in covering games that use generated images and text.

Wee Aquinas vs. a semantically dead version of Wee Aquinas.

I. The Art of the Chair

Many of us have learned a lot over the past year about how a generative artificial intelligence produces images. The best metaphor I’ve heard is that it’s a learning map. We tell the AI to learn about what a “chair” looks like, and it trains on countless images, learning the gist of a chair’s shape and dimensions. After a few million cycles, it now knows that when it receives a prompt for a chair, it should produce an arrangement and selection of colors that replicates the concept the machine has had tagged as a chair.

The defenders of AI will point out that, hey, this is also how human brains work. We create a cognitive map of a thing’s necessary and sufficient features and then replicate them. This does some heavy lifting for their argument, because when humans recognize chairs, we’re also drawing on countless images and descriptions of chairs. If humans also draw inspiration from other pieces of artwork, then it must be okay for AI to do the same.

Except that’s not actually how art works, and the fact that someone would use such an analogy to defend AI generative images or text exhibits is rather telling. Because they aren’t explaining why AI generation is art. They’re explaining that they don’t know what art is.

When I say “Draw a chair,” we can divide the outcome into two broad results. The first is that our test subjects draw a chair. Four legs, a sturdy back, probably wood, whatever. That’s because we’ve engaged in a cognitive experiment and confirmed that, yes, human beings do indeed create cognitive maps around definitions and then adhere to those definitions.

But something else can happen. Let’s say we ask Pablo Picasso to draw a chair. Or Tim Burton. Or G.R.R. Martin. Or Dr. Seuss. Or any number of other artists. Those chairs will conform to our cognitive map of what a chair ought to be. But they might also defy it. Instead of producing the “necessary and sufficient” cognitive version of a chair, the one that’s as close to the ideal of what a chair ought to be, they will produce something we recognize as a chair that also tells us something else. Maybe the chair will be a single peg sprouting from the soil only to terminate in a low-backed stool. Maybe it will be a throne of swords that threatens to spill the blood of its inhabitant and remind them of their grim responsibility. Maybe the chair will be drawn from every angle at once and usher in an entirely new style of perception. Maybe the chair will be a wide nimbus of femurs and rib bones. Maybe it will be pornographic, or regal, or a cozy barstool.

But it will not be just a chair. It will be art.

You’ve probably heard about this guy. (This is not an AI image.)

II. Roger Ebert Hastily Backpedals

I was fresh out of high school when Roger Ebert stated that video games were not comparable to the artistic efforts of other mediums. Like many others of my age and inclination, my initial reaction was a pimpled rage. To my credit, there were some unexpectedly good games out there. I’d cut my teeth on Planescape: Torment; surely the extent of Ebert’s experience with the medium was, I dunno, Mario Kart.

Over the next months, Ebert’s position became increasingly clear. In response to one thoughtful query, he wrote:

I did indeed consider video games inherently inferior to film and literature. There is a structural reason for that: Video games by their nature require player choices, which is the opposite of the strategy of serious film and literature, which requires authorial control.

I am prepared to believe that video games can be elegant, subtle, sophisticated, challenging and visually wonderful. But I believe the nature of the medium prevents it from moving beyond craftsmanship to the stature of art. To my knowledge, no one in or out of the field has ever been able to cite a game worthy of comparison with the great dramatists, poets, filmmakers, novelists and composers. That a game can aspire to artistic importance as a visual experience, I accept. But for most gamers, video games represent a loss of those precious hours we have available to make ourselves more cultured, civilized and empathetic.

In a later conversation, Ebert further clarified. When asked if all films were art, he responded that, no, they are not. Raging Bull is art; Pirates of the Caribbean: The Curse of the Black Pearl is not.

Now, there are two ways to interpret that statement. The first interpretation, and the one I held for a very long time, is that Ebert was making the classical error of confusing “high art” for “all art.” In this perspective, art only becomes art once it’s good enough. I’ve never liked this argument myself. I’m not much a fan of essentialist thresholds. Besides, I’m one of those weirdos who parrots aphorisms like “The world would be very quiet indeed if only the prettiest bird sang.” Much of my favorite art could fairly be characterized sentimental trash.

But I don’t actually think that’s what Ebert was saying. I think, rather, he was speaking about intentionality. The reason Raging Bull is art but Pirates of the Caribbean is not is because one of them was crafted by an author while the other was crafted by a machine. A corporate machine, to be precise, with six principal writers, a director not exactly known as an auteur, and who knows how many studio and property-holder notes. Raging Bull, on the other hand, is Martin Scorsese through and through. It has authorial vision and a deliberative nature.

Now, this is to say nothing of quality, and certainly nothing of subjective tastes. I doubt that Roger Ebert would disagree that you are free to prefer the swashbuckling adventures of Elizabeth Swan, Will Turner, and Jack Sparrow over Jake LaMotta’s bouts with his personal demons. Nor, really, is this about gatekeeping. Ebert was unabashed about enjoying trash.

Rather, I think Ebert’s theory of art was, at its heart, a very simple one. Art is something humans produce. It is not something machines, social or otherwise, cobble together from mathematical formulae.

One type of machine. (This is not an AI image.)

III. Modeling Art

Of course, this raises a rather sticky question, one that plenty of very smart thinkers have taken a stab at with very little consensus: What, then, is art?

Now, I want to be clear that Ebert’s perspective — and really, this is my perspective on what I think is Ebert’s perspective — is only one model, one definition, of what “art” might be. There are plenty to choose from.

For Plato, a work of art was an incomplete and flawed imitation of a perfect conceptual form, and was thus an inadequate method for acquiring real knowledge. Georg Hegel went the other direction, arguing that art was a perceptual gateway to truths that could only be arrived at rationally. Immanuel Kant is less interested in defining art than he is in examining what happens when we make aesthetic judgements, but he arrives at the idea that art is the purposive expression of beauty; that while a landscape may trigger the part of our mind that appreciates aesthetics, it doesn’t really qualify as art because it wasn’t deliberately crafted to trigger that emotion. Meanwhile, Cartoonist Scott McCloud famously declared that art is anything humans do that isn’t directly related to survival or reproduction.

Now, these are all quick and dirty descriptions of some rather heady aesthetic philosophies, and I’m not trying to come down on one side or the other. This might surprise you, but according to my own rubric I actually don’t believe that board games qualify as art. Of course, given that I’ve dedicated thousands of hours to trying to elevate the way we discuss this medium, it should be obvious that I don’t mean this in a way that portrays board games as somehow “less than” art. Rather, it’s that I conceptualize play as a function of culture that runs parallel to artistic endeavors. Whether we’re talking about sports, gambling, friendly competitions, or formal rules-based activities like board games, these pursuits satisfy a different need than, say, a story. That said, I’m very much aware that boundaries are not always useful, and that some board games might cross the line from Play into the domain of Art. Or vice versa! The point is, there are lots of ways to think about the function of art in society.

But there are commonalities between those theories. Brush away the competing perspectives on capital-t Truth and you arrive at a shared kernel that’s more or less intact across all of these models. That common thread is purpose, intent, or deliberate creation. Art and play are things that humans do. They are produced by humans, shaped by humans, given meaning by humans. Art doesn’t happen without human input.

“Sophistry!” cries the AI Bro. “Art is a pretty picture! You can see it right there! With enough careful selection and upscaling, you can’t even tell the difference between the produce of generative AI and pathetic meat-made art!”

Okay, so let’s get into that. Because art serves an actual function beyond being something you can commodify and sell.

Not pictured: The tribe inventing games. (This is not an AI image.)

IV. Art Arises from Pareidolia, But Pareidolia Is Not Art

Imagine you are an early hominid living on the shore of a great salty sea, on the barrier between forest and desert and beach.

Your people selected this home for its safety and abundant resources, but dangers abound. There are small floating creatures in the sea that sting, leaving your kin ill for days. The young or the old rarely recover fully from such an assault, and sometimes seize and die. The nearby forest is rich with fuel — fire, you give it two opposable thumbs up — but it’s easy to get lost among the trees, and there are snakes and beasts that will gladly snatch a lone wanderer. There are reports that another family has settled nearby. Supplies have gone missing, and you suspect these newcomers are to blame. But another clan means opportunity as well, whether for trade or mating or the new stories unknown people can tell.

In such a situation, your survivability isn’t so much a question of raw strength and endurance. Even the strong break bones, and even the longest runners tire. Rather, it’s a question of pattern recognition. Can you spot a jellyfish among the rippling water, pick out a snake in the grass, or determine whether to trust or fear your neighbor?

These talents are potent. They’re the reason for your success. Fortunately, they come deeply ingrained in the way you perceive the world. You can see the stonefish among the coral, or pick out tiger stripes among the tall grass. This process, which we call “pareidolia,” extends not only to visual and auditory stimuli, but also to the way you understand intent. This is imperative, because you need to understand that certain creatures or people mean you harm while others are safe. And because you’re a communicative creature, the members of your family tell each other about what they’ve learned.

This is all excellent stuff. By recognizing patterns and telling your kinsmen about them, there’s a very decent chance your children won’t stuff their bellies with the wrong berries. At times, however, this talent goes wrong. Seeing patterns isn’t very useful if you only recognize them when they’re obvious. Noticing the saber-toothed tiger only once it’s falling on your head simply won’t do. So that part of your brain tends toward the active. You see faces in vines and creatures in clouds. When those newcomers mumble to one another in a tongue you only partially understand, suspicion stirs in your breast. When a storm blows in after some vigorous me-time, you wonder if maybe the sky prefers you’d leave your genitals alone, thanks.

How do you compensate for this overcompensation? You use the same talent to your advantage. You observe the world and then you talk about it. You bounce ideas back and forth with your family and maybe even the newcomers living up the beach. Over time, it becomes apparent that similar people can think very different things, and very different people can think very similar things. You start building those ideas into a conceptual framework that explains why the world operates the way it does. And you share those frameworks with fellow hominids, first as a matter of survival, later because it feels good.

And you keep doing it. For a million years and more. You do it with such enthusiasm that it becomes as much a part of who you are as eating and mating.

This is art.

Yet another type of machine. (This is not an AI image.)

V. The Parasite in the Bloodstream

Let me be declarative for a moment. Art is a human endeavor. Art is the process by which we human beings interpret and understand the world around us. Existence has no inherent meaning; at least I don’t believe it does. The swirl of motes that dance about us do not have feelings of justice. The swaying branches of the trees above do not in fact contain faces. When a tsunami kills a quarter million people, this is a senseless event. It’s when we examine the million data points around us and draw it into a shape that reflects our subjective reality in some way, whether through telling a story, painting a face, even shaping agency via play, we have created art. This is a deliberate process, but it’s one that we humans can’t help but engage in nearly every day, both as creators and recipients.

Just because this process is deliberate and interpretive doesn’t mean it can’t be whimsical, or absurd, or nonsensical, or whatever. I’m not saying art has any inherent hurdle it has to cross before it stops being a doodle and becomes art. It just has to be something a human has deliberately created because it gives shape to this chaotic world. It can be a doodle! It can be as small as a blown raspberry or a silly face, or as huge as a multi-million dollar production.

Okay, but can’t AI-generated art be deliberate? Maybe if we twist my meaning around enough. But it’s deliberate in the way a search engine is deliberate. Here’s a thought experiment. If I open a search engine and type in a term — perhaps “chair” — are the images that pop onto my screen works of art that I have created? Am I an artist?

Whenever I offer this hypothetical, AI Bros become gymnasts. “Of course you can be inspired by a search! People have always done art research!” Which, you’ll note, is nowhere near what I asked. I can stare at a patch of dirt and be inspired. That doesn’t make dirt into art. Again, that’s pareidolia. Just because a landscape can be beautiful doesn’t mean it’s art. It’s only in the act of purposely creating something that art is formed.

The same goes for generative artificial intelligence. You may have noticed that I decline to call it “AI art.” The whole concept is a misnomer. I might as well offer other mismatched terms by saying that a machine plays or digests or grows exuberant. By definition, art cannot be made by a machine. In Ebert’s case, it can hardly even be made by a corporate machine, and that’s still a structure dominated by feeling humans.

Even calling it “generative AI” triggers an admittedly persnickety part of my brain into engaging in an etymologist’s fallacy, because “generative” is a word we use in another context: generation, as in reproduction. Something is generative when it reproduces its own kind. When a human creates art, it will inevitably be inspired by the art they observed before.

Now, I want to be clear: at some point in our evolution, we invented art wholesale. We learned how to tell stories and sing songs and slap handprints on cave walls. We did that. Meaning from chaos. Order from the inchoate. That’s a beautiful and mind-blowing step in the story of who we are as a species. It’s written right into our helices that we are creative creatures who sing shapes into the cosmos.

Because of that, it’s entirely fitting that we draw inspiration from one another. The stories we tell are more meaningful and refined because of the thousand stories we’ve digested before.

An AI engine, on the other hand? It can only train on the things we’ve made. Once the system swaps to training on AI-produced images, it collapses into formless rubble. That’s because nothing about generative AI is actually generative. There is no semantic meaning behind the semiotic constructs it spits out. It’s a parasite; its host is us. Oh, it results in images that are very art-like; but our comprehension of those images relies entirely on our capacity to see meaning in static. The machine itself sees no chair. It sees a clustering tendency, a color palette, spaces that are usually empty when humans attach an image to the proper tag.

This is also why I’m suspicious, but undecided, about generative AI being employed as part of the creative process rather than short-circuiting it altogether. I’m an advocate for what Amabel Holland calls “bad” work, but which I suspect hews closer to “novice” work or something of that nature. Not every game is entitled to polished illustrations. There’s room for stick figures in art, or functional art, or art that struggles to express itself. I’m thinking not only of Holland’s work here, but also Splotters and 18xx and the dozens of rough but compelling gems I cover on Space-Biff! every year.

My current stance comes down to another thought experiment. If somebody could write a cowboy story, but found their prose lacking compared to that of Cormac McCarthy, and a generative AI could give their story a comb-over that would bring it within spitting distance of McCarthy’s descriptions, would that be a suitable substitute? Personally, I have a very hard time accepting such a tool as valid. This isn’t the same as a time-saving or assistive tool like a word processor or a spellcheck. It’s a generator that circumvents the human intentionality of art altogether. The words cease being the author’s and become the algorithm’s. Why should I bother to read something nobody could be bothered to write?

Perhaps there’s another requirement bouncing around my head when I think about art. It must be yours, flaws and all. That’s the only way it can be a full expression of your attempt to decipher the world around you. Otherwise you haven’t deciphered anything; you’ve offloaded that most human of efforts onto a machine.

Some “bad” art. (This is not an AI image.)

VI. Why Do I Care About Chairs?

Because I care about meaning.

Art is a form of communion. We create art, we share art, we consume art. We don’t undertake these activities because they’re commodities we can purchase. We do them because they are the lifeblood of who we are as creatures. As such, art isn’t mere semiotics. It’s inseparable from the semantic meaning behind the artifact itself.

As a critic, I’m interested in examining the shapes and meanings we impose on the world. When somebody presents me with a game, I’m immersed within a solid-state expression of that designer’s deliberate intent. Their language is built of systems, illustrations, textual statements, rules, limitations, possibilities. But even though these are unfamiliar organs of creation when it comes to more traditional artistic mediums, they aren’t all that different from a sculpture’s stone or a painter’s pigments or a storyteller’s narrative beats. They are the medium, the tools, the method of conveyance. By communing with someone else’s creation, I receive a glimpse — an imperfect glimpse, all too brief and limited, but a glimpse nonetheless — of that creator’s understanding of the world around them.

There is no intent behind a machine’s produce. None whatsoever. Sure, it might produce a legible image or readable text. But again, that’s an imitation of something a human has produced, but one that skips the deliberative, interpretive steps that a human mind undertakes when they produce art. The result is a formless mass. It’s the exact opposite of art, because it doesn’t imbue the world with form and shape and meaning. It’s more noise, not more meaning.

When an AI Bro says, “But you use a word processor to write text rather than writing by pencil, don’t you? So what’s the difference?” they betray their perception of art. In this mode of thinking, art is only a product. Content, nothing more. And because it is content, something meant to fill our screens without filling our hearts, its only objectives are efficiency and profit. Such an impoverished attitude may serve well enough when it comes to steel beams or bundles of hay or word processors. For all I know, it will prove well enough for most players of board games.

But it isn’t enough for me. Both as an artists and as someone who critiques a great deal of art (and play), the entire reason for engaging with this stuff is because I want to interface with this incredible thing human beings produce. We all produce art, or at least we did when we were children. We have all breathed meaning into the world around us. To turn that process over to a machine is to deaden the portion of ourselves that kindles meaning. Or, at the very least, it blurs the line between true creation and imitation.

A different form of art. (This is not an AI image.)

To explore creation is why I started writing Space-Biff! in the first place. I’m proud of what I’ve made. I’m even proud when I look back at my earlier articles, which were fumbling on every level, because it reminds me of how far I’ve come. That is something I accomplished because I stuck to something my grandfather always said: “If something is worth doing, it’s worth doing badly.”

That, to me, is an unmissable part of creation. We create because we cannot help but create. Our creations always start out bad. Always. But if we work at it enough, our creations become better. In the process, we become better, too. Had I leaned on generative AI to produce my site’s articles, the greatest loss wouldn’t be that Space-Biff! could not exist as it does today. I’m not so self-aggrandizing as to believe that my silly little website adds all that much to the world. For all I know, maybe a machine could have resulted in a version of Space-Biff! that’s every bit as good as it is right now. Maybe even better. But no. The greatest loss wouldn’t be that. It would be that I wouldn’t have become myself. I wouldn’t have shaped my tastes and skills. An article wouldn’t have gone from consuming six hours, sometimes even multiple infuriating nights, to taking an hour. I wouldn’t have become confident in my aptitudes. I wouldn’t have shorn my weaknesses to a finer stubble. Speaking entirely literally, I wouldn’t be me.

That’s the beauty of art. As we create, we are created. It’s something we do as effortlessly as breathing. We need no machines to create. And in the process of imitating it via an engine, we counteract the very reason to create in the first place. All that remains is something we can sell.

That’s why, as both an artist and a critic, and I have no intention of covering games that circumvent the human element. To evaluate the produce of generative AI is a dead end; there’s nothing there. I might as well critique a lump of clay that has never been touched by human hands. Art and play are human endeavors, and that’s where my interests lie.

Special thanks to James Matthias not only for designing my current Wee Aquinas avatar, but also for indulging me by drawing a pair of chairs. He is an artist.

This article was funded by the generous donors at my Patreon, who regularly receive early glimpses of difficult pieces. My next article, about the highlights from SDHistCon 2023, is already live there for donors.