A Different Aftermath

September 10, 2022 1:47 PM

I made another weird little comic thing, hopeful and a little bittersweet, about conservation after the apocalypse. A topic near and dear to my heart, Lord knows.

Your favourite and mine Wombat/Kingfisher Ursula Vernon has released a short comic about the post-apocalypse on Twitter, using AI generated art.
posted by MartinWisse (21 comments total)

I needed that, thanks.
posted by signal at 2:43 PM on September 10, 2022 [1 favorite]

Threadreader version
posted by cozenedindigo at 2:47 PM on September 10, 2022 [2 favorites]

I saw this without knowing it was AI generated art and blew my mind.

Sure gives the lie to some of the intemperate comments in Sep 7's discussion from people claiming there was no creative input at all to making AI generated art. I wish the artist here had a longer post about their process; the second linked thread is great but I want to see a visual diary! The silhouettes they drew and work prints from the process of iterating to the final art.
posted by Nelson at 2:48 PM on September 10, 2022 [5 favorites]

This was lovely, thanks for posting.
But-- kudzu is edible, everything but the vines. The young leaves are like spinach, use fresh or cooked; the roots can be pounded and powdered to make starch. Yes, yes, it's invasive, but in a world depicted like in this comic, I'd love to have fresh greens available, and malls can fuck right off anyway.
posted by winesong at 3:31 PM on September 10, 2022 [1 favorite]

That was really lovely! Cheered me up no end after a rather bad day. Humans may be fucked, but lie on earth will probably be ok. (I hope.

Also, I fell for the person wityh the lemon balmm. Somebody stop them before they plant mint!
posted by Fuchsoid at 4:01 PM on September 10, 2022 [2 favorites]

But-- kudzu is edible

I've read some jaw dropping comments on this site in my many years here - this one is way up there.
posted by MillMan at 4:22 PM on September 10, 2022 [6 favorites]

It doesn't look that different from some commercial comics... I think because tracing from photographs, or using a photo reference and painting over it, or even just using a straight up photo with a filter over it are all pretty common in comics already. Some of the panels have that photo-with-a-filter-on-top vibe, especially the ones with plants in the background.

My friend who's in comics says it's much harder to use these tools for sequential art, VS for illustration where you can use the image right out of the box (sometimes). It takes more work to make everything look cohesive and to get the images to support the story you're trying to tell. This is more like using AI as a tool to assist with something (detailed city backgrounds) most artists don't draw by hand anyway.
posted by subdee at 4:59 PM on September 10, 2022 [4 favorites]

@subdee
Yeah, the author makes exactly that comment in the thread about their process. This was much easier since it was all silhouettes over backgrounds.
posted by Inkoate at 5:05 PM on September 10, 2022 [2 favorites]

This is more like using AI as a tool to assist with something (detailed city backgrounds) most artists don't draw by hand anyway.

Just being honest, this comment immediately made me think of something I had just read.
posted by jeremias at 5:26 PM on September 10, 2022 [1 favorite]

The history of technology is people adapting themselves and the world to facilitate the use of tools. Cars need lots of roads to be useful, along with people both learning how to drive and how to not get run over.

So it's easy to imagine the ai making a lot of parts of the process much easier, but still needing help across the finish line or with a good amount of human work to fill in the parts that are unique to the current project.
posted by kaibutsu at 6:08 PM on September 10, 2022

Well, goddamn if that wasn't the best thing I've seen today. Thank you!
posted by rp at 8:47 PM on September 10, 2022

I’m going through some things, and the story gave me some wracking sobs at the frogs. The story is beautiful, but somehow I’d glossed over the part about using the AI, and damn if I don’t feel a little hoodwinked, like my feelings were bought cheaply.

I know I’m supposed to be open minded about all of this stuff, and I’m sure there’s some bright sunny world where it will contribute to a better place, and not be commercialized and used to put animators and artists out of work. You know, just an art version of the internet! Surely it would never be twisted to create a better profit margin by allowing companies to just stop paying artists at all, reducing them to freelance touch-up specialists competing for Fiverr grade piecework.

AI don’t need to pay rent, eat, or sleep, and undoubtedly there are people looking at their payrolls right now who are ecstatic about that.

I am torn, though, because the writer was very open about it: if they had had to do it all by hand, they wouldn’t have done it, and we’d never have this honestly poignant story (fuck, how long has it been since I read something hopeful that, being confronted with the possibility of hope, it made me break down and cry?). Having said all that above, sharpened my pitchfork, renewed my membership in my local Luddite chapter, I’m also thinking about the armful of story ideas that I’ve never worked on because I never really had the art skills, or the drive to improve them.

And, sorry to go back and forth so much, but isn’t there something to be said for the work that is done? The effort that is spent, the craft that is mastered through hours and years of trying, failing, and trying again? Hurrah! AI is here to help even the lazy to tell stories! At least, until, the AI tells our stories for us, and we just WALL-E ourselves.
posted by Ghidorah at 6:25 AM on September 11, 2022 [3 favorites]

This is more like using AI as a tool to assist with something (detailed city backgrounds) most artists don't draw by hand anyway.

I take your point but most of these panels are largely detailed city backgrounds, with very little anything else on them. Many of the panels are "background" only. In others the silhouette is hand-drawn but not the primary visual impact.

So while even if something isn't hand drawn there's definitely art in selecting the image, maybe color grading it. And in the case of the AI technique also coaxing the AI to produce the image. At this level folks aren't just typing a few words and getting lucky; there's a detailed iterative process of adding more and more phrases to coax out the look you want. My impression is the artist here worked a long time to get a specific art style out of the generator. It's creative and unique. I just wish they'd write more about how they did that!

(I'm also intrigued by their aside they used DALL.E for the carpenter bee. Interesting choice and also fraught, because now you have to convince two of these systems to draw in a compatible style.)
posted by Nelson at 7:53 AM on September 11, 2022 [1 favorite]

Thank you, that was lovely.

It brought to mind The Man Who Planted Trees, by Jean Giono, narrated by Robert J. Lurtsema. It's a short, beautiful fable, one of my all time favorites. (Theres a free audiobook version with a different narrator, and of course the print version. But I haven't listened/read those so I can't speak to the experience.)
posted by Winnie the Proust at 10:40 AM on September 11, 2022

(It seems like not everyone in the thread knows that Vernon is an artist as well as a writer? Have a look at her comic Digger, which started in 2007 and won a Hugo in 2012, if you like humane stories with wombats, baby demons, archivists and deep feelings about hyena culture.)
She's been playing with AI generated art for a few months now- here's a post from her Tumblr in July with a process video of her editing a Midjourney image into a digital painting. Part of her commentary there is "This is not a composition I would necessarily have come up with by myself. On the other hand, it’s not an image the computer could have come up with by itself either."
posted by Shark Hat at 12:12 PM on September 11, 2022 [11 favorites]

The conversation around AI feels very gatekeep-y sometimes and even ableist. An undertone of "If you haven't spent years refining your skills, or if you're not able to do things by hand before relying on a tool, your art is not valid, your way of expressing your inner world is not valid."

I'm not saying there aren't valid arguments against the technology, the ethical/copyright issues with model training are probably going to delight philosophy enthusiasts for years, but that particular gatekeeping feels very weird and toxic to me.
posted by simmering octagon at 12:56 PM on September 11, 2022 [3 favorites]

Simmering octagon, that was in my mind as I wrote my post, and I can definitely how AI tools like these can certainly be used by people with disabilities in positive and important ways. And I also understand the idea that, as this artist said, that this wasn’t plug and play for them either, that using these AI requires it’s own skill set, and I can accept that as an accomplishment worth noting in itself.

That doesn’t change my feeling that the work people are putting in on all of these public facing AI arts tools is essentially crowd sourced free beta testing for a way for businesses to move forward without having to pay artists. It’s no different than the initial idea of cashless payments being a new level of convenience, and how that has morphed into a way to fire all the cashiers, and for businesses to make money off of all the balances held in their various cards.

It’s a fantastic and interesting advance, but it’s only as advanced as it is because it will allow higher corporate profits at the expense of labor.
posted by Ghidorah at 1:58 PM on September 11, 2022 [4 favorites]

I'm annoyed at the comic using AI art to tell a story about the death of artificial intelligence and basically my entire livelihood, and how the survivors are all happier and better off with subsistence farming, which is totally not that hard you guys despite everything my ancestors did.

That doesn’t change my feeling that the work people are putting in on all of these public facing AI arts tools is essentially crowd sourced free beta testing for a way for businesses to move forward without having to pay artists.

Honestly I'm getting the whiff of moral panic off this. Businesses weren't paying artists anyway; it seems far more likely that we as a society will become more sensitive to the specific quirks of AI art, and associate those quirks with corporate soullessness the way we already do with certain kinds of music, or purple people with noodle arms and tiny heads.
posted by Merus at 5:59 PM on September 11, 2022 [2 favorites]

I mean, as a musician and especially as a percussionist/drummer, I've been watching technology replace musicians like myself for almost 40 years. Now, one person can write, perform, and produce music in their home studio in ways that reduce maybe eight people doing creative, artistic work to just themselves.

I have mixed feelings about this, but I think on the whole it's a big net positive. Lots of performers and other creative people have fewer specific jobs in making music, and a lot of people dedicated exclusively to those jobs have lost their careers. But on the whole, making music is available to more people and more music is being made. And some skills become more valued or exist when they didn't before, while other skills become less valued or defunct. That sucks for some people.

As a formally trained percussionist and drummer, I mostly hate the idea of people who can't play those instruments making these tracks electronically because there's a whole universe of skill and nuance that is missing. But, much of that isn't important to the popular audience. And creative people who aren't drummers use drum machines and their equivalent in novel creative ways that are interesting and enjoyable in their own right. And more music is made by more people because there are fewer barriers to entry. So some things are lost, but other things are gained. And, again, I think it's a big net gain even as I mourn the loss of a number of things, some of which are dear to my heart.

I think that history has repeatedly demonstrated, emphatically demonstrated, that art is not defined solely by technique and that various technical skills arise and are refined and more greatly valued, then challenged, then decline in importance and are replaced. All of the work in learning a technical creative skill can become socially worthless within a single person's lifetime. That really sucks for them. But it absolutely doesn't indicate a general decline in the creative art itself — it frequently precedes or corresponds with a creative revival.

If we think about static visual art, with or without sculpture, we see that what it is, what is necessary, can and does radically change over time. It doesn't have to have perspective, it doesn't have to be representational, there is not one single technical skill that is required.

An AI recapitulates both technique and concept from existing art, but so do human artists. And, to be frank — and this is very true of musicians as well — a great many of them (probably including myself) are not adding anything or saying something new — they're not being creative in that fundamental way that we imagine is what distinguishes an ideal human artist from an AI.

I think that our romantic ideas about art both overvalue or exaggerate what "art" is while also undervaluing most art that is actually created. In other words, an AI cannot, at present, create new art in the romantic sense that we imagine "art" to be Art. But then, neither can most artists. On the other hand, an AI can already do some of what most artists do almost as well as most of them do it, and that art that is made, by a person or an AI, has value and utility that we ought not disregard. Sturgeon's Law says that 90% of everything is crap, and that's true if our standard for non-crap is bona fide excellence. But that shouldn't be our standard and, in practice as a culture, it is not our standard. Wholly derivative stuff rearranged in various ways is perfectly acceptable to most audiences most of the time — in fact, mass audiences almost always dislike real innovation.

AI art — visual art, in this discussion — isn't a threat to art itself, but it's certainly a threat to artists who specialize in the specific technical skills that an AI can utilize. We're more familiar with one technical skill replacing another technical skill, but there's also a long history of creative technical skills being subsumed into other creative technical skills and becoming forgotten/unneeded as things that exist independently in their own right. They might be mechanized, or they might be implicitly recapitulated as part of a comparable technique.

Again, a good example would be drum machines. A creative musician need not have any of the manual skills of playing a percussion instrument, but they can understand a great deal of percussion's function in music and create percussion tracks. As I said, some things — many things — that I treasure are lost and not included, but percussion itself, as an aspect of music, certainly isn't lost and creative musicians realize new expressions of percussive rhythm while entirely lacking all manual percussive skills. And that's okay

I think it's a net win, assuming it allows more creative musicians to create more music, which I think it does. And, honestly, there's still a place for skilled manual percussionists and drummers — including writing/creating electronic percussion tracks as well as, of course, playing traditional manual instruments.

One way of looking at it is that this change clears out a lot of mediocrity in manual performance, leaving behind a relatively clear space for the most skilled and talented to exercise their creative skills in works where they are especially valued. Also, as sometimes happens, later some of the value of these manual skills will be rediscovered and be back in vogue. This is the way of things.

It's a mistake to take comfort or refuge in the fact that these tools as they exist now are inadequate for what the OP writer/artist initially hoped to do — most if not all of the additional visual work they did to create this will be within the capabilities of these AI tools by next year.

For the visual artist, these developments are extremely disruptive because they implicitly devalue a whole bunch of technical skills used to create visual art. That's bad for them and we shouldn't be dismissive of this. But it's not bad for "art", per se, nor is it even bad for the broad commercial production of visual art. It expands accessibility and democratizes commercial and popular art specifically because it devalues a specific cluster of technical graphical art skills.

Again, that sucks for those people who have dedicated themselves to mastering those specific skills. Not all of them — as in my example of percussionists/drummers, there will remain a niche for the most highly skilled and creative. Everyone else who is creative and wishes to express themselves in these ways — certainly most of those who want to make a living at it — will have to switch to a different set of skills. This kind of disruption has always happened, over and over, in the arts. There's no small amount of difficulty and pain involved, but it almost always heralds a burst of creativity and genuine innovation.
posted by Ivan Fyodorovich at 10:49 PM on September 11, 2022 [8 favorites]

Honestly I'm getting the whiff of moral panic off this. Businesses weren't paying artists anyway; it seems far more likely that we as a society will become more sensitive to the specific quirks of AI art

What is “moral panic” even supposed to mean here? Concern for an entire group of people seeing their livelihood not only endangered, but also handwaved away as being not worth worrying about is a thing, even if you don’t personally want to acknowledge that “disruption” as used in Silicon Valley/business speak is almost always a way of saying “we will find a way to pay people less for their labor, which will make us rich” and that this, given most of the technological advances we’ve made recently have been applied to make it easier for businesses to employ fewer people, in worse conditions.

As far as “businesses weren’t paying artists anyway” please do tell this to the artists who have jobs in design, in advertising, and all throughout society. I’m not sure what your definition of art and artists is, but I’m thinking of the people working in visual mediums. Every product you buy has someone’s artwork on it, even if it’s just a label on a piece of fruit. If you can’t see that AI tools will be used to replace any and all in-house or freelance graphic designers as soon as they’re ready for the big time, then just maybe look into the increased speed of automation and shadow labor disguised as convenience and self service. Corporations don’t adopt new tech and new ideas unless they save money, and they are quite good at packaging the loss of jobs as great for customers.

And, as for humans being able to tell something is or isn’t AI, well, that’s why these incredible expensive to create AI tools are being made publicly available. It’s beta testing, and, as above, shadow labor. Public betas are a way of getting the general populace to do the work that used to be in house, that used to be paid labor. We’ve shown we are perfectly happy to do this work for free. The goal of this level of public access is to hone the tools to the point where it is cheaper to pay someone to touch up the AI output so that it’s close enough that most people won’t notice. And, just like in most animation these days, those jobs will be farmed out piecemeal across the world, to whatever country has the lowest possible wages.

This is the ongoing direction of nearly all business adoption of new technology, and it has been for years, and there have largely been two responses to it: the people who haven’t yet had their livelihoods threatened/erased, thinking, “that sucks, but luckily my job is safe” and the people who used to have, or never could find employment in their field because it’s been made redundant. You say moral panic, I say empathy for people who will have the rug pulled out from under them.
posted by Ghidorah at 4:38 AM on September 12, 2022 [4 favorites]

The use of AI undercut the comic's message--all I could think of is the upcoming apocalypse in commercial art. And it does have that "AI" look--that long shot of city scenery is already a cliche, isn't it? Could be the apocalypse, could be the opening panel to your superhero comic, could be the end of a big city romance. There's nothing in it that makes me think, oh, this is someone's unique vision of the world.

I know it's inevitable that we'll see more of this but hopefully, like the frogs and birds in the comic, some hand drawing will survive.
posted by kingdead at 4:56 AM on September 13, 2022