My view, entirely without any evidence, is due to AI being everywhere there will be a backlash against AI art. People will prefer something that is verifiably human made. People are going gaga over AI art right now because it is new and something that AI has not been capable of previously, but I think the shine will wear off. A big part of art is it quite often says something about the human (or artist’s) experience.
Of course, the world will not have 8 billion artists, so I’m not saying that this solve the employment issue, just my thoughts on that specific part of the AI debate.
Mammalian brains evolved over millenia. Probably wishful thinking to think this could work. ChatGPT’s BS detector is likely non-existent, another problem.
Where have you been? A subset of people have been very angrily lashing out at it since the very beginning. Not just the “it’s stealing our ideas” people and the “it’s taking our jerbs” people, but also the “it is only art if it is made by humans” people. Meanwhile, AI art related Reddit and Facebook groups have grown to hundreds of thousands of members.
One thing that has kind of surprised me (though it probably shouldn’t have) is just how credulous, unobservant, and gullible people are over A I images. I keep seeing obvious AI images of things that couldn’t be real getting thousands of comments and shares by people believing it is a real photograph of a real thing.
Like this “log cabin”
and this “mobile library in Rome”
and this bird.
Scammers are going to make a shit-ton of money selling fake things with AI.
Do any of these images look impossible to you? Like they couldn’t possibly exist?
In the long run, it’s going to lead to massive distrust of photographs, and a host of other things found on the Internet. That’s going to be a huge issue and one I don’t see a real solution for. It’s going to be a mess.
I’m aware. Perhaps you missed where I said people are currently gaga over it, but I think in time there will be a rejection of AI art as AI grows to encompass more and more areas of our lives.
I see that treehouse as completely buildable. Does it exist? Probably not. Did the house grow organically as part of the tree? Of course not. Could it be built from scratch to look as it does? Sure, armed with enough money.
OTOH, the bird and the bookmobile are obviously fantasy garbage.
Could something similar that be built? Yes, it isn’t literally impossible. But look at details of the image (especially the guard rails and the top part with the “glass window”) and you will find a number of MC Escher failures in the image. As for practicality, either some of the stories are far too short for human use or some are far taller than they need to be.
Another one getting thousands of comments and thousands of shares from credulous admirers is an old woman knitting a giant cat. If details of the cat didn’t immediately expose it for an AI, the old lady’ feet definitely should have. But despite the wide press coverage of AI (and, hell, the decades of existence of Photoshop) apparently there are vast numbers of people not even primed for the concept of maybe a photo being unreliable.
Knitty Grandma has caused a flood of parodies, such as this one:
Bing image creator has always been very heavy on the censorship of what you can prompt (for instance, the names of every president all the way back to George Washington are banned) but tonight they have gone for the ultimate position and bannned all words. Literally everything anyone enters is getting a content warning.
ISTM there’s a basic disconnect here in this syllogism
I like and trust my friend Jane.
Jane wouldn’t lie to me.
Jane sends me stuff she found on the internet.
Therefore the stuff Jane sends me is reliable.
In effect each person outsources their credulity and fact checking to the person who sent it to them. Which is certainly a convenient and usually-valid heuristic honed over millennia of small village living where everyone knows everyone.
The problem of course is that Jane did the same thing, outsourcing her credulity and fact-checking to her source. As long as there is one troll or jokester in world, their entire “downline” of forwarders trusts somebody else to have done the checking.
Which results in an avalanche failure of rapidly flowing unstoppable BS.
None of us have the time to personally vet every factoid, every photo, every audio clip, every video.
I would hope that Wikipedia remains reasonably accurate in most important cases. If I want to check out Johnny Cash on YouTube, I want it to be things he actually did. Not fake crap. Photos I don’t know what you do, but real, actual photos only please. Jobs!, for bots anyway. I want to see the truth, things that actually happened. I guess there is going to be a marketplace for curating the actual truth. The fake crap disgusts me and can burn.
I’ve no doubts that there will always be a market for handmade bespoke art pieces. That’s probably the case for fine art even now – most AI stuff is either silly images, intentional trolls, “that’s neat” renderings that will be made and forgotten and some commercial art. Probably not too many people decorating their living rooms with AI renders printed off RedBubble.
Of course, even now there’s a difference between owning an original piece and having a framed print of a painting. Maybe people willing to have a reproduction will prove willing to have an AI generated piece that looks nice.
I had a few pieces of “computer generated art” back in the 70s. They were mostly geometric, kinda spirographic prints. I almost wish I’d kept them for the kitsch value, but they weren’t especially interesting. Now, my walls are covered with art, and as much as possible I like works that come directly from the hand of a human (or in one case, the paw of a dachshund).
