i see three main points people are making, which i think are fair: first the supposition that the products of people’s labor are being used somewhere in the process of art generation in an unfairly profitable way; second the supposition that the products of the artists’ labor are being stolen before they can sell it; third that the other two result in a chain-reaction down the line where nobody’s an artist any more because neural networks are adjective-er than humans.
for what its worth the way AIs combine images is just not the way you think AIs combine images. machine learning is complicated. neural networks are complicated. read up on GANs; watch a few Two Minute Papers videos (like this one, which is particularly relevant: https://www.youtube.com/watch?v=FCf8OA4GPvI). the way they combine im its not forgery (AI are not perfectly replicating your piece and selling it as it is for their own value; at some point it might involve copying/“looking” at it), its not stealing (you still have your art with you), its not even Youtube Poop remixing. generally the hypest way to do neural networks right now is to take a collection of items the AI knows to contain similar concepts and creates an understanding of that shared concept; a good dataset will be big enough to understand all the full variations of that concept, or accurately deduce the rest of the variations, all through the natural inconsistenties that a bunch of pictures of the same thing might produce. an AI knows how to do an oil painting style because it probably knows quite well the actual qualities of an oil painting. it also knows what the sun looks like, how cloth drapes at different weights, how holes in the ceiling of a building might imply shafts of light illuminating on the floor, and other incredible stuff. beats me how it does that, and you likely won’t get it either. but the point is the way people are talking about AI on here seems solely informed by imagination.
for unfair and stealing bit, there’s something to be said about how certain online communities, websites, people, blogs. etc are naturally private due to obscurity while also being technically completely public; like caves hidden in valleys that take quite a bit of looking to get to and see. in the past we just assumed they would stay hidden: the art on this blog will only be seen by its followers, or people i link the blog to, surely; ill teach you the way to my little cave. unfortunately the internet has been filled with extremely informed and skilled valley-treading and cave-finding robots, and unless you have big ol’ signs that say “DO NOT FIND THIS CAVE!” (i remember for forums you could stop scrapers from coming to your website, i think on tumblr there is something like a “do not show in search results” button too), or somehow otherwise avoid their path using strange internet tricks or complete luck, they will totally find your little cave and rummage through it. what i mean to say is this: your art is already part of the publicly available data, and you’re in a tough spot to do anything about it. i have no solutions.
the objectionable part is not in this collection of references for analysis, nor in how the AI is usually interpreting what objects are what, but the misuse of the AI to produce things you don’t like (which is part of the point i wanted to make in my first post, besides arguing for the artistic integrity of what the AI is doing with its sources). and i see how it seems unfair – someone smarter and more knowledgeable about art, labor, intellectual property, and value could articulate this more – that a customer of a neural network could so freely replicate someone else’s work with the original stylist having nothing to show for it. it wouldn’t be any better if a sufficiently skilled person was requested to draw original pieces in your style. so in that way this is less about AI and more an objection against the violation of someone’s intellectual property, if that exists.
for the second point on losing artists, i understand the reasoning as following: nobody wants to be a starving artist, and to avoid that in capitalism one needs money, and if someone is dead set on being an artist (or for some reason can be nothing else) they must sell their art to make money. if no one is going to give them money for art, they will certainly be a starving artist, which we don’t want. the conclusion here is, if we want artists (starving, at the very least), to restrict anything else that makes art which might compete with the artist’s. okay, that might work; we could stop having AI make art. but what if, despite our best efforts, we can’t fully restrict such a thing? what if AI generated images are here to stay? what if artists are actually already partnering with AI to make art, and you’ve just been none the wiser?
naturally we would think at least some artists could no longer be paid for what they currently do. if companies can afford a concept artist they likely could better afford a subscription to DALL-E. but im simply not jumping to the conclusion that companies specifically will be replacing ALL artists with AI. some will certainly try for the novelty of it, two or three might succeed, most will probably find out negotiating with an artist or two for long term projects is much easier than talking to the team of computer science geeks to see if the damn art-generating program will churn out something slightly more [quality], more palatable to today’s changing tastes – if “AI generated images” is even the zeitgeist in whatever hypothetical time period this is happening. maybe by the time the tech is developed people will think handmade stuff is popular and companies will follow suit, with all their logos being drawn on glass panes and scanned.
i mean, who are we even talking about, considering artists? the fandom artists definitely aren’t disappearing because of neural networks. all those millionaire furry artists simply won’t lose business because people are going to Midjourney and entering in “shirtless anthropomorphic tiger, [famous furry artist]-style”. people who want oil paintings want some physical oil paintings and usually not a digitally-generated oil painting printed on canvas. people want a human intermediary for basically everything, and if you’ve ever worked retail with an older customer base you’d know that’s true. id certainly prefer human-made art to neural networks if i was gonna spend money (and i apologize to our future robot overlords for such a statement).
so who is really completely losing out here? many artists are, quite frankly, not big enough to get fried. concept artists for triple-AAA shooters might take a hit, i imagine; its decently easy to generate generic cities and alien landscapes with AI. but frankly, artificial intelligence is a tool, and not a scooch-along robot replacing your cubicle in the office. you still need people who know about art and art willing to put in the work for it to produce anything of commercial value. even if it gets so good that the CEO of a business can log on to ARTIFICIAL-INTELLIGENCE-WEBSITE.com and type into the text box “I wuld like to buy a compny logo for $50 please”, im sure there will always be more status in having that “human touch” to your designs, and less status in those robo-packs of ugly slightly mismatched placeholders. besides, artists are already using AI to help create their products in the first place; i refer to that Two Minute Papers video again (https://www.youtube.com/watch?v=FCf8OA4GPvI).
your art is publicly available and you really don’t have much control over what happens to it. unless little intranets for families and towns and cities and such become more popular, unless you start password protecting websites like forums used to do, your shits getting collected. its not a good reality and im definitely an advocate for internet privacy. for many artists the internet is unfortunately the only place they know how to get their stuff out there. its not a good situation for artists to begin with. the thing is, what’s happening with modern AI art generation isn’t… really that bad. unless some specific human wants to make it bad. if from this moment on we decide to push for legal-social-whatever punishments for not compensating the artists whom AI was trained on, okay, i guess. companies developing these AI probably have too much startup money anyway and most artists make far too little for the years of wrist pain. but i would say the customers requesting your art style or whatever are certainly at more moral fault, because it almost certainly wouldn’t just pop out a given art style unless your prompt guided it to.
i will stand by the fact that the art of modern neural networks is more complicated, less hurtful, and way cooler than ripped off collage nonsense, and it will be exciting and scary to see where we go from an AI winning an art tournament – for the moment you can be scared if you want, i guess, but if you didn’t already lose your job to a guy utilizing a neural network then you’re probably already safe. (and you might as well get on the train before you’re left behind.)
i’m on the sea! i’m on the sea! / i am where i would ever be / with the blue above, and the blue below / and silence wheresoe’er i go
↪ a commissioned expression study of my swashbuckler tiefling, Ocean, by the extremely talented & wonderful @m00rkas. you’ve given me everything w these, mo. thank you so much again!
It is really important to me that all of you learn about Al Bean, astronaut on Apollo 12 and the fourth man to walk on the moon, who after 20 years in the US Navy and 18 years with NASA during which he spent 69 days in space and more than 10 hours doing EVAs on the moon , retired to become a painter.
He is my favorite astronaut for any number of reasons, but he’s also one of my favorite visual artists.
Like, look at this stuff????
It’s all so expressive and textured and colorful! He literally painted his own experience on the moon! And that’s just really fucking cool to me!
Just look at this! This is one of my absolute favorite emotions of all time. Is Anyone Out There? is like the ultimate reaction image. Any time I have an existential crisis, this is how I picture myself.
And then there’s this one:
The Fantasy
For all of the six Apollo missions to land on the moon, there was no spare time. Every second of their time on the surface was budgeted to perfection: sleeping, eating, putting on the suits, entering and exiting the LEM, rock collection, setting up longterm experiments to transmit data back to Earth, everything. These timetables usually got screwed over by something, but for the most part the astronauts stuck to them.
The crew of Apollo 12 (Pete Conrad, Al Bean, and Dick Gordon) had other plans. Conrad and Bean had snuck a small camera with a timer into the LEM to take a couple pictures together on the moon throughout the mission. They had hidden the key for the timer in one of the rock collection bags, with the idea being to grab the key soon after landing, take some fun photos here and there, and then sneak the camera back to Earth to develop them. They had practiced where they would hide the key and how to get it out from under the collected rocks back on Earth dozens of times.
But when they got to the moon, the key was nowhere to be found. Al Bean spent precious time digging through the collection bags before he called it off. The camera had been pushing their luck anyways, he couldn’t afford to spend anymore time not on the mission objectives. Conrad and Bean continued the mission as per the NASA plan while Dick Gordon orbited overhead.
Fast forward to the very end of the mission. Bean and Conrad are doing last checks of the LEM before they enter for the last time and depart from the moon. As Bean is stowing one of the collection bags, the camera key falls out. The unofficially planned photo time has come and gone, and he tosses the key over his shoulder to rest forever on the surface of the moon.
This painting, The Fantasy, is that moment. There have never been three people on the moon at the same time, there was never an unofficial photo shoot on the moon, this picture could never have happened.
“The most experienced astronaut was designated commander, in charge of all aspects of the mission, including flying the lunar module. Prudent thinking suggested that the next-most-experienced crew member be assigned to take care of the command module, since it was our only way back home. Pete had flown two Gemini flights, the second with Dick as his crewmate. This left the least experienced – me – to accompany the commander on the lunar surface.
“I was the rookie. I had not flown at all; yet I got the prize assignment. But not once during the three years of training which preceded our mission did Dick say that it wasn’t fair and that he wished he could walk on the moon, too. I do not have his unwavering discipline or strength of character.
“We often fantasized about Dick’s joining us on the moon but we never found a way. In my paintings, though, I can have it my way. Now, at last, our best friend has come the last sixty miles.” – Al Bean, about TheFantasy.
