This piece was made possible by Substack contributors; consider becoming one!
___
The AI photo editing app Lensa experienced a surge of publicity over the last few days as social media users dumped photos of themselves into the chute and got illustration-like images of themselves out. In a similarly streamlined process, artists and passersby dumped those images into their eyes, opened their mouths, and let the hot takes flow.
Lensa, according to my timeline on many a social media platform, is a nefarious plot against all that is art, an unfeeling mechanism designed to rob the skilled of revenue and the unwary of authentic aesthetic jouissance. Get thee behind me, Lensa! You shalt not feast upon my image or my soul!
As it happens, I am not in fact going to feed Lensa my image (or my soul I guess) because I hate having my image online and the thought of sharing a bunch of tweaked artsy photos of myself in impressionist or expressionist or comic book style with you or anyone triggers my fight or flight reflex and then some. I’d rather hit my foot with a hammer. Maybe more than once!
But! My own particular neuroses aren’t really here or there. The question is, “Is Lensa a force for evil in this veil of tears?!” And the answer is…eh. Not really.
There are two common arguments for why Lensa and AI are bad for art. The first is that the AI creates art that is “soulless,” inhuman, and a betrayal of all that is good. The second is that AI is exploitative.
The Robots Are Going to Eat Your Soul!
Let’s address the charge of soullessness first. The argument generally goes something like this: AI creates art without human intervention or human touch. There is no intention; you put in a prompt, and it spits out an image sans deliberation. There is no vision and no brain involved; therefore the product is not art as we understand it. It is a pallid, uncanny imitation of art, and a grotesque parody of the human spirit.
Okay, so maybe people don’t put it in quite such gothic terms. But that’s the gist.
The central problem with this argument is that it’s nonsense. There is absolutely an artist involved in creating the art that pops out of Lensa. Or more than one artist, really.
First, there’s whoever codes the AI. And then there’s the person who supplies the images and the prompts.
Coding and computer manipulation is an intimate part of a lot of art-making at this point. Unless you think that video-game developers, animators, digital painters, and electronica musicians aren’t artists. And you shouldn’t think that.
It’s true that the AI itself has neither brain nor soul; it’s inert. But you could say the same of a synthesizer, a drum, the laptop I’m writing this on, or a paintbrush. Virtually all human art is created using non-sentient tools. That doesn’t mean the art lacks a human touch or human intention. It just means we’re tool-using critters, and part of art is playing with stuff.
Some tools provide a more direct link between creator intent and final product. The laptop here for example reproduces my intent fairly directly for the most part, if you discount the occasional typos. “The robots will devour us!” The laptop delivered that sentence to you, but it started in my brain before, you know, the robot devoured it.
But tools don’t have to be straightforward. Lots of art incorporates randomness and even, at this point, algorithms. John Cage is one pivotal figure here; his famous piece “Silence” is not in fact silent; the musician doesn’t play, which means you end up very aware of people shuffling, breathing, existing. The musician’s “tool” is the world; the lack of intent is the art.
There are numerous other examples; automatic writing, surrealist games like Exquisite Corpse, the entire subgenre of algorithmic art. The fact that AI is collaborative and stochastic just means it’s aligned with a lot of the norms of current visual art practice. You can like the outcomes or dislike them, but there’s little ground to claim in some categorical way that the output isn’t or can’t be art.
Nonetheless, many people claim that Lensa isn’t art—just as many people claimed that photography wasn’t art, or that electronic music wasn’t art, or that disco wasn’t art, or that hip hop wasn’t art, or that dada wasn’t art, or that comics weren’t art, or…well, you get the idea.
New artistic tools or new artistic genres are often dismissed as soulless, as inauthentic as a betrayal of the human spirit. Sometimes those arguments are about preserving hegemony—the anti disco fervor was fueled by homophobia and racism for sure; anti pop art sentiment was also often homophobic. Painters opposed photography because they (reasonably) worried it would cost them commissions and money. The charge of soullessness serves as a shorthand to express economic and cultural anxieties. It’s a mini moral panic; people or groups or ideas seen as dangerous are cast as impure, uncanny, undesirable.
Over time, people usually come to accept the art in question, though often caveats remain. Ben Shapiro still doesn’t think hip hop is music because he’s a racist piece of shit.Comics artists still froth when you mention Roy Lichtenstein, maybe possibly because of lingering homophobia. Elvis Costello still whines that he doesn’t like 80s music because he’s a cranky butthead. And of course people can have different preferences; no one’s telling you that you have to like what Lensa does. But claiming it isn’t art, or that it’s inhuman, is silly. And, given the history of art to this point, indefensible.
The Robots Are Going to Eat Your Income!
The stronger argument against Lensa is that it’s exploitive. The AI is programmed with a lot of examples of art. Many of those examples are not out of copyright. But there’s no mechanism to pay the people whose art goes into the system. The creators of Lensa alienate art from its original creators, rework it, and monetize it for their own purposes. That’s capitalist exploitation.
The way the AI uses other’s artwork to make money for its owners is disturbing. However, it’s worth pointing out that all art is made out of other art. Sometimes that’s less direct—every artist studies and is influenced by the work of other creators. Sometimes, though, it’s blatant. Hip hop uses samples of other recordings. Collage is composed of visual material from multiple sources. Memes repurpose existing images. Found text poems use…well, found text. Sherlock Holmes films are based on Conan Doyle’s stories.
In some of these cases, no one is making any money; people aren’t usually paid to generate memes. In others, copyright has lapsed, much to the irritation of the Conan Doyle estate. But hip hop artists did in fact use bits and pieces of copyrighted recordings to create their own work, which they then sold.
Hip hop artists were eventually sued by the giant corporations whose work they were using, with sweeping effects on the art form; de la Soul’s early records, which used a kaleidoscope of recognizable bits and pieces of other records, are still unavailable digitally because of sample-clearing issues. In general, the genre moved towards shorter, less identifiable snippets, and/or to only one or two samples a track.
I think (or hope?) most people would agree that using copyright law to kneecap hip hop was not great, even if hip hop managed to adapt and give us a lot more wonderful records. Preventing artists from using other art—and copyright law very often does just that—limits the art that can be created. It’s important that people be paid. But it’s also important that people be able to use culture to respond to, interact with, remix, remake, reimagine culture.
Copyright law is supposed to balance the need for artists to make a living with the need for artists to have access to art. I’d agree that that balance seem out of whack in the current iteration of Lensa. But I don’t think the goal should be to make Lensa illegal or to prevent it from functioning. Figuring out a royalty system might work. Or partnering with individual artists; you could train it on your own work, and then people could pay you to provide them with an AI portrait in your style. In any case, the goal should be allowing art to flourish (and yes Lensa is art) not shutting it down.
The other thing that a lot of the criticisms of Lensa ignore is that exploitation of artists is not a new thing, and certainly doesn’t require new technology to implement. Comic book artists for Marvel and DC continue to labor under exploitive work-for-hire contracts; when their characters appear in films that make 10s and 100s of millions of dollars, they receive pocket change. Spotify notoriously pays creators literal pennies—and Spotify isn’t even creating its own art or tweaking that art in any way. It’s just delivering you someone’s work and screwing the creators.
The conclusion here isn’t “it’s all good”; one form of exploitation doesn’t excuse another. On the contrary, the fact that the default is miserable exploitation for almost all artists with huge paydays for a handful is a massive injustice which is harmful to art as well as to creators—and is part of the broader problem of capitalist exploitation in every industry.
It’s not always clear how to address that exploitation, especially when new technology appears. Artists are still struggling to get Spotify to pay a decent rate per stream, without a lot of success. Efforts to get users to switch to different services haven’t been very successful. I doubt that shaming people for using Lensa is going to work either, though it’s possible that artists whose art is being taken for the system could sue it, as they sued hip hop creators.
Ultimately, though, artists need a social safety net, health care, UBI, and a society that gives artists much more space and freedom to create without having to worry about starving to death or dying of preventable disease. In such a world, new tools, and even new robots, could be seen as an opportunity to make new, weird, random art, rather than as inimical malevolences poised to consume us all, one photo at a time.
____
And hey! Again, if you think artists should be paid for their work, consider paying the person/cyborg who wrote this piece by contributing to this substack!
Today I learned, or inferred anyway from what you wrote, that Roy Lichtenstein was gay.. I promise you, I am no more annoyed with his career now than I was two minutes ago. :-)
I'm not annoyed that Lichtenstein made cartoonists into unwilling collaborators. I'm only annoyed that he didn't pay them or credit them. (It's possible that he would have preferred to pay and acknowledge them, but didn't dare for fear of being sued? In which case I should blame our dumbass legal system and not Lichtenstein.)
I'm not bothered by AI, and I agree with you there's still artists involved. But I wonder if maybe AI that is trained on the public domains should be uncopyrightable.
AI is also devastating to low-level artists - the people who do $20 commissions for people who want a drawing of their D&D character, and other low-level art jobs. It's a first step for many illustrators and cartoonists, and it's already being replaced by AI. But it feels pointless to me to try and stop that; it would be like trying to stop the typewriter industry from being devastated by computers.