Arts and Ai



Original Source Here

Arts and Ai

Are We Even Asking the Right Questions?

One of my first experiments bringing Ai into my digital painting process for The Fallen Cycle.

Over the past several weeks, there has been a wave of reactions to recent developments in the world of Ai and the arts, which seems pretty well typified in this article from the Verge. I’d like to get into why I think this line of thought is asking the wrong questions for the wrong reasons, but you’ll have to bear with me a little.

Let me start with a little about me. I’m a “multi-hyphenate” artist, which is to say I’ve done a lot of work in a variety of mediums: music production, visual art, writing. I’m not famous, but I’ve developed a methodology already, and didn’t have much interest in the early Ai apps of yesteryear, until one came along that seemed to present some utility as tool rather than replacement.

In talking to several artist friends of mine, I heard that MidJourney had an approach to Licensing that seemed to have artists in mind. My curiosity was piqued.

My next question was: is this a tool that I can find use for in my existing workflow, or which might save my joints from a little extra time scribbling?

As I’ve tried to answer that question over the past couple months, it’s been a source of inspiration, excitement, and even absurd humor, although as I see it — at least at this stage — nothing it produces is out-of-the-box production ready. It’s halfway there. (Midway, you might say).

However, it is hardly a one-trick pony. Once you get a handle on how it parses text prompts, the in-development beta test version of Midjourney can already imagine clay bas relief designs to later be worked with in real clay, or a particular style of oil painting, or the production design element in a set — such as paintings on the wall.

Raw MJ output

More broadly it seems much better understood as an imagination midwife — or middleman, if that’s your gender preference — for artistic production. It is not an artist. A machine learning algorithm doesn’t have agency, and I think most of us would like to keep it that way. For its sake as well, given our track record.

I have found that it fits quite naturally into a visual workflow that I’ve been developing since the late 90s, playing with methods like the Gysin / Burroughs / Bowie approach to cut-up and randomization, that is, a means of riffing and generating unexpected combinations that is still very much dependent on the involvement of human agency.

This is true for many of the music projects I’ve worked on as well, but visually it started with collaging, and then digital collaging, and then photobashing, and then digital painting with photobashing or pencils as a monochromatized underpainting, and so on.

At first, so-called “traditional art” was done in tandem, something I did at the same time but in a very different mode.

As time went on it has kind of all dovetailed, and the differences in methodology seem increasingly less important. It’s become a question of matching method with style. The process for me is always a dialogue, a back and forth between trying to get it to conform to your intentions and letting your own intentions be guided by what you actually get. Found art and collage teach you similar skills.

Maybe it’s a dialogue with the piece itself, maybe it’s a dialogue with a bunch of pieces of source and found objects, maybe it’s a dialogue with your collaborators or team members. Or with the artwork that inspires you and that you’re trying to explore for yourself.

It should be obvious where the tire meets the road here. However, there seems to be a persistent idea in the forums that Ai is a servant that should deliver you exactly what you ask for, or that the end goal is for it to deliver a finished image at a moments notice. I’d like to challenge that idea.

What it seems to do instead is a great deal more fun, and confounding — attempts at communication and interpretation, and subsequent sometimes creative misunderstandings. “This isn’t what I thought I was asking for, but it’s actually a quite interesting direction” is a common reaction. Discovery is mixed in throughout the illustration process, not provided upfront. At least for me, this is nothing new.

In the testing I’ve done with some fellow artists, programmers, and the like, it’s already proven quite valuable to have a shared channel where we can each riff off of each other’s visual ideas, mediated by the MidJourney bot. Its utility for activities like Roleplaying Games is obvious, and also at this extent mostly unexplored.

It “isn’t the thing, it’s the thing that gets you to the thing”, as Lee Pace says on Halt and Catch Fire. A tool meant to accelerate the imagination process between humans, not an end destination in itself.

My sense from listening to the MidJourney weekly talks on Discord is that the Devs and Mods are willing to deal with the absolute nightmare of community moderation specifically for this reason. My only complaint so far stems from this choice, although I think I’ve come to understand why they are willing to take the risk. They are using a rather brute force attempt at controlling output by banning an ever-increasing list of words, many of which also have innocuous meanings. It is a problem they’re well aware of, but at the scale they’re dealing with, brute force measures seem the way forward at least until they have the moderation staff and methods to attempt something more subtle. This is probably a story for another day.

I said earlier that my process still keeps my own choices pretty central. There is however an equally valid aesthetic tradition of removing the human element partially or even entirely from the creative process, usually by randomization from sources as varied as mathematics (like the 12 tone rows of Schoenberg), paint spatter, or the forces of nature. So even removing the artist isn’t a deal-breaker, although that methodology has always been anathema to my own.

Raw MJ output

“What is art?” “Why is art?” “How is art?” The bulk of popularized modern and post-modern art has interrogated these ideas for close to a century. Warhol, De Kooning, Duchamp, Cage, all those guys they teach you about in art school. There is nothing new under the sun.

Are we eternally stuck in this cul-de-sac, recycling the same questions? We are often inured with a type of preciousness in regard to what we create.

Tibetan sand paintings might provide a worthwhile exercise. Or, more improbably, something I picked up from the art departments of several companies I’ve worked for, where we would all toss our comps on the table or on the screen, but without any discussion of who had made what, and discuss the work. You learn one another’s style, and might infer who created what, but what was important was that it didn’t matter. The work, and what it does and doesn’t do is what’s important.

Ironically, I’ve been in many art classes with a similar methodology, except critique was often pointedly personal. Students frequently fall into attacking or defending art as a sort of proxy or effigy for the self. As an artist, you don’t have much to fall back on. You have to legitimize your work, and if it isn’t to be found in dollars, with the extent you’re willing to go for your art.

Digital painting in progress working with MJ

I think there’s an insight to be found here. There can be so much anxiety about proving your ability as an artist, and if you have any success at all, plenty of opportunities to turn the work into that proxy ego — the structure of ownership and control is very much tangled up in this.

Art is personal, and I don’t intend to contest that, but it doesn’t occur in isolation, even if we’re cloistered atop a mountain. It is a conversation beginning with inspiration, negotiated with reality, and finally let go of, like sending a child off to college.

This is always my sense when I’ve finished a major project like a book, though in an even more dramatic sense now that I think of it. For better or worse, it’s out in the world now. Hopefully it’ll make some new friends.

The investiture of our ego within our own work is something I’ve spent years trying to untangle for myself. With some success, but it isn’t easy. If you’re trying to work as an artist, I’m not sure if there’s ever much hope in severing the knot cleanly and entirely, but it still seems far better than trying to tie new ones.

Raw MJ output

This is at the heart of my 2020 book MASKS: Bowie & Artists of Artifice: art is already artifice. But it is artifice to employ an effect, and which is often tied up in our very personal internal experience, or famously, “a lie that tells the truth.” The intention of the deceit isn’t purely to fool but rather to stand in, one way or another, for reality. A daub of paint tricks the eye and suddenly it is a shadow or a tree stump, depending on its context. It ceases to be a daub of paint, while clearly it is still just that and only that.

Or in Brian Eno’s words, in regard to the “meaning of art”:

My notion is that art does something, not that it means something.

Its meaning is what it does.

I will admit, I have a lingering sense of sadness that working by hand might eventually become obsolete in the way that illuminated manuscripts or papermaking are today, maintained by a small, often dwindling niche of craftspeople. But it is also no different than that, in this regard. The wheel turns.

There’s no denying that wheel keeps turning faster. While clearly still in Beta development, I’ve watched MidJourney progress in noticeable leaps and bounds in the short months I’ve been working with it. It is learning from our inputs, while we learn how to better communicate with it.

Who knows where this technology will be in five years? Even the developers themselves don’t really know.

Nevertheless, in each of these cases, it’s part of a historic continuum, and not in the least bit unprecedented. A lot of the initial concern spreading through journalistic publications and social media seems shockingly unaware of either of these histories. Or perhaps this is a bit like that “vintage” meme of Mohammad Khatami, where he assumes every American clearly must have read their De Toqueville.

I’d like to suggest that we should be concerned, but not because this technology is a threat to creativity. In that regard, if anything, it is a boon.

The very real dangers posed by Ai, whether when applied to the creative process or facial recognition, come from the method of their use, and the biases they might reinforce. Not to put too fine a point on it, but the fail points are obvious when we consider that corporations and governments are going to use them.

Raw MJ output

The risk within the arts is probably quite benign in comparison to, for one example out of many, the Israeli government teaming up with Google to provide Ai driven facial recognition and behavior interpretation tools for police and military use, but that isn’t to say no one’s jobs are at risk. I imagine the distribution of these effects will be asymmetrical at least for a while, for example, many new shoe-string budget independent project which might otherwise not exist without the assistance of Ai, whereas Corporations very well may use Ai tools to slash in-house art departments down to skeleton crews in the name of profit. Corporations are also likely to see it as an artist replacement rather than a tool to be used by artists, which is a mistake on pretty much every level. A predictable one.

I don’t want to get too much further into the weeds about the broader issues regarding ethics and machine intelligence, because that is not as much in my wheelhouse. For that I’d recommend checking out the work of scholars like Damien Williams, who are attempting to tackle this broader issue headlong. Whether or not corporations pay that any mind is probably another story.

This leads us to the next concern that is frequently raised. Copyright. I’m not a lawyer, and in any event the fact of the matter is that this technology is too young for there to be extensive precedent on who can “own” the output of an Ai. The standing consensus seems to be that no one can, at least until that output is reworked and utilized for something else, in which case your version is “yours”, but the source is not.

What I can tell you with some certainty is that Copyright law is already inherently fucked… like, in its fundaments. Our labor is what we’re seeking to protect as artists, and if we look at who has the power in dealing as an artist with major corporations, it’s already clearly very much not protected.

Copyright law exists within a context intended for value extraction. This sort of ownership is inherently non-collaborative, seeking to reduce or outright criminalize methods of sampling, remix, or even re-interpret a work, if you don’t want to take the risk that Fair Use won’t hold up. This is another very large topic that I’m breezing through for brevity, however, this Rolling Stone article gives one example of ways that copyright law is contrary to the creative impulse.

Artists need all hands on deck when it comes to better protecting the benefits of our labor, or better yet our ability to survive without having to extract value ourselves, and the ability to draw from and respect our priors, sources and inspirations while we create something unique from those common sources.

Raw MJ output

I’d like to conclude this short discussion of Ai and the Arts in what might seem like a strange place. (Or maybe not so strange if you know me). Tolkien, mythology, and my conception of art as intrinsically collaborative.

It is a fairly long quote, but I think an important one, from On Fairy Stories, wherein he says:

Max Müller’s view of mythology as a “disease of language” can be abandoned without regret. Mythology is not a disease at all, though it may like all human things become diseased. You might as well say that thinking is a disease of the mind. It would be more near the truth to say that languages, especially modern European languages, are a disease of mythology. But Language cannot, all the same, be dismissed. The incarnate mind, the tongue, and the tale are in our world coeval. The human mind, endowed with the powers of generalization and abstraction, sees not only green-grass, discriminating it from other things (and finding it fair to look upon), but sees that it is green as well as being grass. But how powerful, how stimulating to the very faculty that produced it, was the invention of the adjective: no spell or incantation in Faerie is more potent. And that is not surprising: such incantations might indeed be said to be only another view of adjectives, a part of speech in a mythical grammar.

The mind that thought of light, heavy, grey, yellow, still, swift, also conceived of magic that would make heavy things light and able to fly, turn grey lead into yellow gold, and the still rock into a swift water. If it could do the one, it could do the other; it inevitably did both. When we can take green from grass, blue from heaven, and red from blood, we have already an enchanter’s power — upon one plane; and the desire to wield that power in the world external to our minds awakes. It does not follow that we shall use that power well upon any plane. We may put a deadly green upon a man’s face and produce a horror; we may make the rare and terrible blue moon to shine; or we may cause woods to spring with silver leaves and rams to wear fleeces of gold, and put hot fire into the belly of the cold worm. But in such “fantasy,” as it is called, new form is made; Faerie begins; Man becomes a sub-creator.

Tolkien’s priors are no secret — Arthurian legend, Finnish folklore, the Kalevala, Wagner’s Ring Cycle, which itself mines similar mythology, etc. Without those things, the Lord of the Rings would likely not exist, and if it did, it probably wouldn’t make any sense to us. If Lord of the Rings didn’t exist, modern fantasy would either be very different, or quite possibly wouldn’t exist at all as a headlining genre as it does today. His stories are simply another piece of this mythopoeia, another voice carrying its own reinterpreted version of those old tunes for a while.

Tolkien meant something slightly different than I do, in terms of humans being sub-creator with the divine. My interpretation is more on the order of symbols, psyche, and society, but it seems salient here. Co-creator rather than sub-creator.

Maybe this doesn’t strike others as deeply as it does me, but for me it is a big part of the allure and mystery of being an artist. My own work both in fiction and non-fiction have always begun and ended with this specific sense of mythopoeia, and our participation in a chain with all the people and ideas we are changed by.

It changes us, and hopefully we change others in our fashion, after our own contribution has been added or taken away.

Speaking as an artist, yes, so long as we need money to survive I want to be paid for my work, but it isn’t why I do what I do. I do it to find my place in that conversation, both between the living and the dead. In dialogue with an Ai as well? Well, why not?

You’re still a drummer if you’re playing on V-Drums or decide to turn to programming MIDI or sampling loops. You’re still an artist if you’re working with the outputs from Ai. You’re still an artist if you hang a urinal in a gallery and call it a fountain. The question, as always, is what you do with it, and how that affects others.

Either that, or none of us are. Surely corporations will use Ai to promote their interests, but nothing I value is served by avoiding a useful tool because corporations will do what they do. Your right to call yourself an artist was never the issue.

For all the uncertainty that exists around this technology, the one thing Ai isn’t going to do is steal the pencil (or stylus) from out of your hand. Let’s be co-creators, and begin to shift from concepts of ownership to participation.

Raw MJ output

AI/ML

Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: