Art In The Age Of Optimization

Art In The Age Of Optimization

When I was a teenager, a family friend saw my budding interest in film photography and gave me a copy of Saul Leiter’s “Early Color.” It hit me like a truck, as great art often does at that point in your life. I tore through the book, completely engrossed by the way Leiter managed to turn perfectly mundane moments into these evocative, painterly scenes. The drudgery of everyday life, something I feared more with each step I took towards adulthood, had never been so beautiful.

So imagine my surprise when I read that Leiter’s work had been strongly disliked by some photography purists. It was underestimated, unappreciated, and even called “vulgar.” This wasn’t because of his subject matter but his medium. Color was seen by many to be unnecessary to good photography, with true skill lying in the manipulation of blacks and whites. I was furious. How could these critics not see what I saw? How could they not understand that the transformative aspect of Leiter’s work wasn’t the color film or even the camera lens, but his eye? How could a group of people who ostensibly loved art be so blind to it?

A decade after Leiter’s death, the validity of another new artistic medium is being called into question: AI Generated Art. With the use of a growing number of neural network based platforms, users can generate elaborate images, and their artistic merit is being hotly debated. Luckily, the answer is simple: AI Art sucks shit and everyone pushing it is a fool or a fraud. Case Closed!

What, I’m supposed to make a coherent argument for why typing “Lady looking at moon, large bust” isn’t art? Absolutely not. It’s depressing that we even have to acknowledge this, let alone dissuade people from actually jumping on the bandwagon of a technology that clearly exists to remove the human element from the process of artistic expression.

The main reason that I’ve mentioned Leiter’s work at all is that I’ve seen a handful of AI defenders compare backlash against AI art to the sort of issues artists like him dealt with; gatekeeping critics reacting to a newer medium with hostility. What this take seems to completely gloss over is that people on both sides of that argument were still making something. The only reason the debate existed in the first place was because all parties involved cared deeply for the past, present, and future of their medium.

What we’re dealing with here, however, is artistic expression as understood by a group of people who value the optimization of processes more highly than the processes themselves. These are people that have always found art and artists to be more hassle than they’re worth. To them “beautiful” means technically competent and “moving” means “there is an astronaut in the picture” and quite frankly the world has enough filler art marketed to rich dorks without the complete removal of artists from the equation.

Now of course they can’t just come out and say this. In fact, fans love to tout AI art’s accessibility, saying that now anyone can be an artist. Unsurprisingly, this claim seems more focused on art as a product than it is on art as a practice. And that love of accessibility does not seem to extend to social services, public spaces, or anything beyond the automation of skill based professions.

So the company line becomes, “we want art to be for everyone,” while the obvious goal remains the same as every other big tech attempt at optimization: to make money. No one truly believes that the goal here is to make art better or more accessible, right? Are people actually looking at this stuff and feeling like they’re at the dawn of a new age rather than the beginning of the end? The ideal outcome for these companies is to provide a service that makes it so that when some tech guy needs an image of an astronaut looking at the moon to promote his new NFT, he doesn’t have to talk to (or more importantly: pay) anyone to get it. Like the vast majority of silicon valley’s latest contributions to the world, the only thing this seeks to actually optimize is exploitation. So why does everyone seem so excited about it?

I chalk some of this up to AI art’s innocuous beginnings. Our obsession with Dall-E Mini this past summer seemed pretty harmless. Who wouldn’t want to see Shrek on an improv team? I’m not blameless in this; I tried Lensa and, just like everybody else, it made me insanely fucking hot. We were all maybe a little overeager to aid in the advancement of an AI we knew next to nothing about, but thus far that hadn’t really bit us in the ass as we thought it would. It seemed like a novelty that had some potentially interesting applications for digital artists.

Then the tone shifted. Even as it became clear that the platforms were trained on the art of human artists without their consent, a small but very vocal group of AI art evangelists started doing everything they could to underline that what they were doing (typing some stuff and pressing enter) was Real Art. They attempted to assure us that the AI learns just like we do and that the art it’s trained on, stolen or not, is just as organic as the influences of any human artist.

If AI art is truly real art, if training a program on our ideas is the same as a person being influenced by other people, then the artist receiving the credit should be the AI, not the person typing in the prompts. These people hope to place themselves not in the role of the artist but in the role of the patron, hoping that we won’t be able to tell the difference. In a system where profit is the only priority, ordering up inspiration on demand is just as good as having it yourself as long as the check clears.

But aside from the fact that an AI cannot make art “in the style of Vincent Van Gogh” without the work of the notoriously human Vincent Van Gogh, the programs themselves are being trained on art that is not theirs to learn from. Shrek cannot join an improv team if there’s no Shrek. It’s stealing plain and simple. If the AI art guys want to spend their free time letting a computer mama-bird them pictures of spacemen looking towards the horizon, then so be it; but the fact that the medium is already this exploitative in its earliest iteration is deeply concerning. 

Art as a career has already been historically dominated by the wealthy. The last thing in the average person’s corner was that talent couldn’t be denied or replaced. But now the initial investment of time and passion required may not be feasible, let alone financially viable, in a world where we’re all working more and being paid less.

So then what problem is this solving? What existing issue with art does this AI fix? Because from where I sit, its primary function is to make these simplifiers, these great optimizers, who only know how to remove humanity from themselves and their experiences, feel like they’ve finally “solved” art. AI is just the latest symptom of a world that views art as content and views content as a public utility: a faucet that should always turn on, free of charge.

The great existential threat of our time isn’t AI, it’s our own natural desire for simplicity. There is not a single foggy corner of the artistic process that some visionless hack isn’t trying to pick apart like a lab specimen in an attempt to simplify the process of commodification. These optimizers have so internalized the shareholder-first business model, the religion of the profit motive, that they’ve managed to apply it to their own interior lives. They seem desperate to surrender their humanity out of some misplaced shame. Maybe they just can no longer bear its weight, but whatever altar they’ve begun praying at isn’t one I’m eager to find or name.

I guess the question I’m trying to get at here is, “to what end?” If we remove all the process, what remains? Do we just keep trimming and trimming and trimming, making edges smoother until we create the perfect art delivery system? A system where all art is exactly what the investors asked for, with no room for accidents, experiments, or chaos of any kind?

Lines only go up, profits only increase, technology only moves forward, the cute robot dogs have mounted guns on them now, and still no one seems to want to admit that maybe forward is only better when you’re not pointed straight at hell.

Algorithms already play an outsized role in what gets made and who sees it, and I don’t think we want to see what happens when we remove ourselves from the equation altogether. We are not simple creatures. We are made up entirely of hypocrisies and idiosyncrasies. They’re what make us who we are. They’re what makes us beautiful. If you attempt to simplify them away, you’ll find you and whatever it is you love doing discarded along with them.

Just this past week, someone used an AI chat bot to generate prompts for an AI art program. It was a fun little experiment that may just serve as a preview for a future in which these platforms are allowed to proliferate without resistance. A future where our last organic ideas are shredded ceaselessly into paste and served cold. A future where a program designed to mimic humanity without feeling can make our own art back at us, showing us a gaunt reflection of inspiration that was once our own. 

Give a gift subscription

New Subscribers: Welcome to Brain Worms! This is my periodic newsletter about internet culture. I’m no journalist, this is just a place for me to string some pretty words together about what it’s like to be alive online. The newsletter just recently doubled in size, so I’m likely to increase the scope of that writing for the sake of more frequent output and having a less depressing beat to cover.

Next week we’ll take a look at the pervading themes of 2022’s pop cultural landscape and talk about the best and worst the year had to offer. A nice breather!!!

A paid subscription here is much appreciated as it directly helps support my writing in a time when, quite frankly, it’s pretty rough out there.




via Stephen’s Web ~ OLDaily

December 7, 2022 at 09:39AM