laika
Loitering Member
- Messages
- 1,415
It was only a matter of time: first it was self-checkout, now it's art.
AI-Generated Art Wins First Prize at State Fair
- Thread starter laika
- Start date
ZenDruid
Supporting Member
- Messages
- 1,038
It's not the death of art. They said the same thing about photography. In this case the artist used his own apps, algorithms, gens, whatever, to direct AI to produce what he wanted. This will have it's use and demand in the commercial art world, corporations may want it for modern status, but it won't be cheap. I think most collectors will still want and value handmade art by a recognized skilled artist that they can touch and hang on the wall.
Bartc
Well-known member
- Messages
- 1,462
I disagree with Wayne here and hope that my agreement with Zen isn't just wishful thinking.
My definition of "art" is a process, not the product, which I consider more an "object of art". The process is whatever the artist does to try to express a reaction to a situation, scene or other stimulus. The object is what results in our concrete version of that process; it's performance in other art forms such as dance, etc. So anything you deliberately do that wasn't intended as strictly utilitarian can be an art form, whether facile or elaborate, regardless of genre, and irrespective of taste and style. That makes "art" the digestion process and the object, well you know.... LOL How's that for ecumenical?
I can see Zen's point. After all, though I chafe at how easy it is now for any child or idiot to make images in digital media with point and shoot AI assistance and Photoshop, having been a creative and pro photographer in the old analog chemical days, I have to admit it can still be art. Whether I appreciate either the process or the result is another issue entirely. But by my own definition, an "artist" using AI to create is still making art. Hoist on my own petard! Merde!
In some Asian countries, Japan and Korea for instance, hand making traditional items via traditional methods makes some artists and artisans "national treasures". Take a look at the post about the Korean inkstone maker, for example. There was one Japanese guy who was treasured for how he hand made indigo cloth, something in the West we would consider just raw materials for a pair of Levis.
Are you really less of an artist because you just paint on store bought linen cloth instead of growing the flax and weaving it by hand the old way? I think not.
But in the end, I have the same nagging distaste for AI generated objects of art that the chess master would have for the computer. And I do increasingly worry about a world in which computer generated images, writing, etc., can fool us, making "truth" so much harder to ascertain. YMMV
My definition of "art" is a process, not the product, which I consider more an "object of art". The process is whatever the artist does to try to express a reaction to a situation, scene or other stimulus. The object is what results in our concrete version of that process; it's performance in other art forms such as dance, etc. So anything you deliberately do that wasn't intended as strictly utilitarian can be an art form, whether facile or elaborate, regardless of genre, and irrespective of taste and style. That makes "art" the digestion process and the object, well you know.... LOL How's that for ecumenical?
I can see Zen's point. After all, though I chafe at how easy it is now for any child or idiot to make images in digital media with point and shoot AI assistance and Photoshop, having been a creative and pro photographer in the old analog chemical days, I have to admit it can still be art. Whether I appreciate either the process or the result is another issue entirely. But by my own definition, an "artist" using AI to create is still making art. Hoist on my own petard! Merde!
In some Asian countries, Japan and Korea for instance, hand making traditional items via traditional methods makes some artists and artisans "national treasures". Take a look at the post about the Korean inkstone maker, for example. There was one Japanese guy who was treasured for how he hand made indigo cloth, something in the West we would consider just raw materials for a pair of Levis.
Are you really less of an artist because you just paint on store bought linen cloth instead of growing the flax and weaving it by hand the old way? I think not.
But in the end, I have the same nagging distaste for AI generated objects of art that the chess master would have for the computer. And I do increasingly worry about a world in which computer generated images, writing, etc., can fool us, making "truth" so much harder to ascertain. YMMV
E.J.H.
Well-known member
- Messages
- 821
I thought the question "is it (good?) art" was less the point here than "is it okay to let purely AI generated art compete with humans in an art competition".
Although I must say I love living in the present times, always been a scifi nerd, and I am actually experiencing a reality in which such questions are not in the realm of fantasy anymore!
Let's enjoy it, it could quickly turn dystopian: http://smbc-comics.com/comic/golden-2
Although I must say I love living in the present times, always been a scifi nerd, and I am actually experiencing a reality in which such questions are not in the realm of fantasy anymore!
Let's enjoy it, it could quickly turn dystopian: http://smbc-comics.com/comic/golden-2
Enyaw
old man
- Messages
- 8,094
art is the expression and stirring a stick in crap and painting a banana is art. The prep and painting of the banana is the artistic expression and then the painting becomes art. The sum of the whole. You can't have a process without a result ..... they are one and the same. Least that's how I understand it.
- Messages
- 2,146
I'm still undecided about it. On one hand, I do like the image. The guy who created/ generated (?) it probably still spent a lot of time on it, chosing the right image, making adjustments etc. On the other hand, some AI images just make me feel uncomfortable. Particularly where human figures are depicted, it's a bit too close to the uncanny valley for me, so I'm a bit apprehensive looking at it. But I would agree that it is art. Maybe deserving of its own category, and not just included with digital art.
ZenDruid
Supporting Member
- Messages
- 1,038
I agree with Triduana, AI should have it's own category. I've seen some AI humanoid images that look like a psychedelic trip in an alien brain. Then we have Computer Generated Images (CGI) that are flawless, think of movies like Jurassic Park or Avatar. And then we have hundreds of drawing tablets and accessories.
rcleary171
Well-known member
- Messages
- 205
I have a friend who is merrily producing and publishing AI generated art. So I took his challenge and used Hotpot AI Art generator for free.
Key Words Entered: "Morning landscape with sheep and a red sun."
Results (low resolution without license but you get the picture):
My issue with AI is not the intrinsic value of the image. I produced a nice picture with no effort. My issue (and my on going debate with my friend) is that he is not the author of his publications. The AI Art Generator is not a "paint brush" but the actual creator of the art. It made all of the artistic decisions based on a simple 8 word request.
Here's my argument: two rooms with locked doors. One room contains an AI generator and one a recent Art School Graduate. You slip your single line art request under one of the doors and leave for the day. The next day you receive your art results. You may not be able to tell if the results were human generated or not. But that is irrelevant - you did not create the art. And claiming credit is wrong.
Key Words Entered: "Morning landscape with sheep and a red sun."
Results (low resolution without license but you get the picture):
My issue with AI is not the intrinsic value of the image. I produced a nice picture with no effort. My issue (and my on going debate with my friend) is that he is not the author of his publications. The AI Art Generator is not a "paint brush" but the actual creator of the art. It made all of the artistic decisions based on a simple 8 word request.
Here's my argument: two rooms with locked doors. One room contains an AI generator and one a recent Art School Graduate. You slip your single line art request under one of the doors and leave for the day. The next day you receive your art results. You may not be able to tell if the results were human generated or not. But that is irrelevant - you did not create the art. And claiming credit is wrong.
I've seen some AI art printed, both on canvas and on expensive paper. The printed AI art is no where close to the screen version - the colors, glow, texture become very, very flat when printed and doesn't compare to human made art, in terms of color, texture, light...
That being said, I think there's a place for it, in it's own category. There - my 2 cents!
That being said, I think there's a place for it, in it's own category. There - my 2 cents!
rcleary171
Well-known member
- Messages
- 205
I had seen some AI art rendered on fiberglass (have no idea how that was managed) and the results were stunning. If you are willing to pay you can capture that beautiful picture onto a physical media.
And there's the crux of it ... if you are willing to pay, AND have the $$$$, you can own just about anything, including an original masterpiece. Since I have neither the willingness or the $$$ to pay for such, it's a moot point for me
rcleary171
Well-known member
- Messages
- 205
That artist (who's name escapes me) posted prices over $4,000. I'm sure digital-to-media cost is a good part of that amount.
brianvds
Well-known member
- Messages
- 1,187
When I first saw that prize-winning piece I was impressed and alarmed. But then I got to see higher resolution versions of it, and ended up vastly less impressed. Like virtually all AI art I have thus far seen, the moment you look a bit deeper, you see the weird flaws. In the case of this particular piece, the "figures" are not really figures at all, and much of the image is covered in textures and stuff that make no sense.
Perhaps the technology will mature in due course. I notice something that I noticed some two decades ago, when 3D image construction software like Bryce became available. For a while there it was all the rage, and the web was flooded with imagery, including amazingly detailed and impressive-looking stuff made by non-artists. I myself played around with it quite a bit. It could be used to make this sort of thing (not my own work):
Ah, found some of my own pieces, made at the time with software that could be downloaded for free (thus less advanced versions!):
At the time I wondered whether such software would soon replace traditionally made art, particularly since you didn't actually need much actual art skill to use it. To some extent this did happen: one example I can think of is artist's conceptions of what new building projects will look like when finished. These used to be done with traditional pen and ink and watercolor, and they tended to be very beautiful. The software-made ones are dreary by comparison, but presumably they are much cheaper, so you see them a lot nowadays.
But by and large, the fad blew over almost in its entirety; it soon become apparent that while a non-artist could make quite impressive-looking things with the software, it had its limitations. Basically, one had to be a fairly skilled artist to begin with in order to really get all that much out of the software, and even then it could ultimately not compete with traditionally painted stuff. And then, of course, digital painting (with graphics tablets) became available, and it produced vastly better results - but required traditional art skills.
Now I see something similar happening with AI art: every nerd and his dog are producing their own masterpieces, but already the flaws are visible; without much special effort I can often (though by no means always) spot when a piece was AI-generated. The other day I saw a bloke on YouTube who had a clever way of going about it: he had AI generate an image, then corrected the flaws himself via digital painting - but then, he's a highly skilled artist, so that he can notice the flaws in the first place, and can then paint over them himself. His somewhat overblown opinion is that AI art will not kill human art but actually save it.
Personally, I think that no one can predict the future, particularly the effect of new technologies. We'd be very naïve indeed to think it will all really be just a fad; I think AI art is more solid than the 3D stuff I posted above, and the technology has yet to mature. As the 3D software did, some art jobs are going to go the way of the dodo. But it might be that AI is going to become an instrument used by artists, rather than something with which you replace all artists.
AI doesn't magically produce masterpieces. It needs to be prompted, and I see YouTube is already abuzz with tutorials on how to write better prompts. Personally I have no desire to write prompts as opposed to actually drawing. But the fact that they are making tutorials on the matter tells me it turned out that AI has its own learning curve, and the more complex and sophisticated you want your picture to be, the more detailed a prompt you'll need to write. It may turn out that in order to illustrate a comic book with AI, you'll need to spend years learning how to write suitable prompts, i.e. the prompt writers will be every bit as expensive to hire as it would be to hire an illustrator. Or prompt-writing will become a subset of what illustrators learn.
But who knows?
What I foresee is that whether we like it or not, the web is going to be flooded with AI art in the coming years. Eventually, assuming it has any limitations at all, those limitations will become clear. Then both computer engineers and artists will get back to their drawing boards.
The larger implications are unknown. As it stands, they deliberately limited AI training sets, so as to prevent a flood of AI-generated porn and fake news. Alas, in the longer run, there will be no way to keep the floodgates closed: expect AI-generated porn movies starring popular child actors, politicians saying things they never actually did, ultra-violent fake snuff films starring long-dead celebrities, and whatever other horrors you can imagine. We have already seen such deep fakes, and they will become ever better and more difficult to distinguish from the real thing. We are about to tumble down a very weird rabbit hole.
None of this will prevent me from making art. Making a living from it might be another matter, but then, I have never managed to make a living from art in the first place. Perhaps this makes me one of the lucky ones who won't actually be losing anything.
Perhaps the technology will mature in due course. I notice something that I noticed some two decades ago, when 3D image construction software like Bryce became available. For a while there it was all the rage, and the web was flooded with imagery, including amazingly detailed and impressive-looking stuff made by non-artists. I myself played around with it quite a bit. It could be used to make this sort of thing (not my own work):
Ah, found some of my own pieces, made at the time with software that could be downloaded for free (thus less advanced versions!):
At the time I wondered whether such software would soon replace traditionally made art, particularly since you didn't actually need much actual art skill to use it. To some extent this did happen: one example I can think of is artist's conceptions of what new building projects will look like when finished. These used to be done with traditional pen and ink and watercolor, and they tended to be very beautiful. The software-made ones are dreary by comparison, but presumably they are much cheaper, so you see them a lot nowadays.
But by and large, the fad blew over almost in its entirety; it soon become apparent that while a non-artist could make quite impressive-looking things with the software, it had its limitations. Basically, one had to be a fairly skilled artist to begin with in order to really get all that much out of the software, and even then it could ultimately not compete with traditionally painted stuff. And then, of course, digital painting (with graphics tablets) became available, and it produced vastly better results - but required traditional art skills.
Now I see something similar happening with AI art: every nerd and his dog are producing their own masterpieces, but already the flaws are visible; without much special effort I can often (though by no means always) spot when a piece was AI-generated. The other day I saw a bloke on YouTube who had a clever way of going about it: he had AI generate an image, then corrected the flaws himself via digital painting - but then, he's a highly skilled artist, so that he can notice the flaws in the first place, and can then paint over them himself. His somewhat overblown opinion is that AI art will not kill human art but actually save it.
Personally, I think that no one can predict the future, particularly the effect of new technologies. We'd be very naïve indeed to think it will all really be just a fad; I think AI art is more solid than the 3D stuff I posted above, and the technology has yet to mature. As the 3D software did, some art jobs are going to go the way of the dodo. But it might be that AI is going to become an instrument used by artists, rather than something with which you replace all artists.
AI doesn't magically produce masterpieces. It needs to be prompted, and I see YouTube is already abuzz with tutorials on how to write better prompts. Personally I have no desire to write prompts as opposed to actually drawing. But the fact that they are making tutorials on the matter tells me it turned out that AI has its own learning curve, and the more complex and sophisticated you want your picture to be, the more detailed a prompt you'll need to write. It may turn out that in order to illustrate a comic book with AI, you'll need to spend years learning how to write suitable prompts, i.e. the prompt writers will be every bit as expensive to hire as it would be to hire an illustrator. Or prompt-writing will become a subset of what illustrators learn.
But who knows?
What I foresee is that whether we like it or not, the web is going to be flooded with AI art in the coming years. Eventually, assuming it has any limitations at all, those limitations will become clear. Then both computer engineers and artists will get back to their drawing boards.
The larger implications are unknown. As it stands, they deliberately limited AI training sets, so as to prevent a flood of AI-generated porn and fake news. Alas, in the longer run, there will be no way to keep the floodgates closed: expect AI-generated porn movies starring popular child actors, politicians saying things they never actually did, ultra-violent fake snuff films starring long-dead celebrities, and whatever other horrors you can imagine. We have already seen such deep fakes, and they will become ever better and more difficult to distinguish from the real thing. We are about to tumble down a very weird rabbit hole.
None of this will prevent me from making art. Making a living from it might be another matter, but then, I have never managed to make a living from art in the first place. Perhaps this makes me one of the lucky ones who won't actually be losing anything.
Last edited:
