I'm a photographer, and I think AI art is in the same infant stage that art photography went through in the 1800s. Photography was initially seen as a passive, mechanical, soulless reproduction of...
Exemplary
I'm a photographer, and I think AI art is in the same infant stage that art photography went through in the 1800s. Photography was initially seen as a passive, mechanical, soulless reproduction of reality. As such, it lacked the necessary qualities to be considered as art. Photographers responded by trying to make their photographs more painterly. They used soft focus, staged scenes, elaborate manual processes in the darkroom that required many steps and lots of craftmanship, all in search of elevating the mechanical to Art. But this movement, although it produced some fine work, was eventually abandoned.
People didn't stop making photographs that resemble paintings; some artists still do. But later work in that vein was a lot more self aware. Early art photography tried to emulate an existing art form in order to borrow its cultural cachet. But this would only ever be seen as a pale imitation of true art by most critics and the general public. Photography didn't really come into its own until artists began to explore its unique characteristics. Any new medium that aspires to be art needs to identify what makes that medium better suited to certain forms of expression than other media. Why make a photograph when you could make a painting? "Because I suck at painting" might be an honest answer, but it's not a good raison d'etre for a form of art that aspires to be taken seriously.
Art photography became accepted as art when photographers demonstrated that the medium has qualities that are different from, and difficult to replicate in, other forms of art.
Today, AI generated art is also seen as a mechanistic, soulless reproduction. Not of reality, as early photography was, but of other, real, human artists' work (or of whatever else was in the training data, including, funny enough, photographs). A kind of wonky, inferior, unskilled reproduction of another medium that possesses real value. If AI art is to have value, beyond the mere convenience of not having to pay a real human who spent years perfecting the craft of illustration, it needs to figure out how to be itself. What sorts of expressions it's uniquely suited to. What kinds of techniques it enables to express human thought and emotion in ways that cannot easily be replicated in other media.
In short, AI art needs to find itself a soul. I expect in the coming decades as the technology itself matures, with fewer wonky artifacts and limitations, human artists will start to approach it as a novel form of expression. Not just as a cheap way of replicating existing styles, but as a way of creating something that could not previously have existed.
That sort of thing won't go away. Just like most people who take photographs with their iPhone just want a naive replication of reality, possibly with a rose-tinted-glasses filter on top to make it look prettier, but still conceived of as being basically a replication of reality slightly improved. People will still use AI to generate fanart that closely mimics an artist's original style, or fake Picassos, or Generic Corporate Stock Photographs But You Actually Don't Have to Pay a Human Being to Create a Photo of a Middle-Aged Dude Holding an Arthritis Medication (tm). But I'd be surprised if, in a few decades, we aren't also seeing artists exploring AI-assisted art as a unique medium chosen because it enables unique forms of expression.
A lot of what makes art art is that it's a human process that involves work, struggle, and life experience. And whether the end consumer understands it themself, the circumstances of the art...
A lot of what makes art art is that it's a human process that involves work, struggle, and life experience. And whether the end consumer understands it themself, the circumstances of the art piece's creation (i.e. the "behind the scenes", the story of how it came together/got funded, the minds behind the piece, etc.) will influence how the art is experienced by influencing how and among whom it's talked about, how it's critically analyzed, and how and to what extent it's disseminated to the general public. Art is inherently a social process in both its creation and consumption, and I think the use of AI to confuse how and by whom art is being created and to undervalue the craftsmanship of those who have studied the form and serve to preserve its history is truly a slow descent into madness.
These are interesting and thought provoking points. I've been thinking a lot about this since that piece made by Midjourney AI was used to win the Colorado State Fair Fine Arts Competition. They...
These are interesting and thought provoking points. I've been thinking a lot about this since that piece made by Midjourney AI was used to win the Colorado State Fair Fine Arts Competition. They question I have to ask myself is "what is the point of art?" The answer I have come up with is that art in all its forms and in its broadest sense is a way to take stuff from inside your head and find a way to share it with other people. This answer has come to shape how I see this issue. In this way I cannot agree with you more art IS social. Right now, in my opinion the way we see the AI being used, diminishes the craftsmanship of art, but not art itself. And I emphasis right now, because I can see this changing in the future. However how I see these tools being used is to allow a greater pool of people access to the ability to share their ideas with other people. Before cameras only those with immense skill and craftsmanship could convert the beauty that they saw in the world into images that could be enjoyed by others. Now almost anyone can capture a moment of personal significance in a way that can be shared with others. Cameras took a slice of art that was previously locked behind years of skill crafting and gave it to the masses. Now we have AI coming in and allowing anyone to using only a few words, and a selection process the ability to construct an image from their mind into something that can be shared with others. When does art stop becoming art? Is there some level of mechanization in the process of creation that invalidates the results? In this way I go back to where I started, and to your words that resonant with me "Art is inherently a social process." I think as long as a person is involved, as long as it's human ideas being given form, as long people can be connected through it, it is art.
It isn't. The latest correspondence chess world championship (which allows the use of computer assistance) just finished. The winner had 2 wins and 14 draws. Of those two wins, one win was from a...
To use chess as an example, centaur chess (a.k.a., advanced/cyborg, a human+machine pairing) historically could beat the best AI or human grandmasters. My understanding is poor, but the synergy comes from a human handling the positional evaluation, and the chess engine the flawless play and deeper evaluation. A quick search didn't show any great evidence either way for if this is still the case, but regardless, it won't be one day.
It isn't. The latest correspondence chess world championship (which allows the use of computer assistance) just finished. The winner had 2 wins and 14 draws. Of those two wins, one win was from a guy who resigned nearly all his games in drawn positions. The other was from an opponent who accidentally inputed the wrong move into the chess server. The entire tournament had maybe one or two "legitimate" wins as in one player actually outplayed the other. "Centaur chess" is dead.
I believe the question "what is art?" is not a productive one. The category "art" is historically broad and inherently inclusive. Whenever the question "is this art?" is made, odds are that we...
I believe the question "what is art?" is not a productive one. The category "art" is historically broad and inherently inclusive. Whenever the question "is this art?" is made, odds are that we will eventually arrive at yes.
It may be more productive to ask, instead, is this fair?. From a legal, social and economic viewpoint, is it fair for someone to profit from the faithful reproduction of a living artist's style? If I had the ability to perfectly reproduce the likeness of an actress (in visual, audio, and performance), in a manner that is somewhat automatic, requiring much less effort, would it be moral to do so without their permission? Wouldn't I be profiting from their life experiences, their body of work, their actual body in and of itself?
I think it's up to the viewer how to approach it. You can certainly try to judge art based solely on the results, going in "cold" without knowing any more about it. (You still bring your cultural...
I think it's up to the viewer how to approach it. You can certainly try to judge art based solely on the results, going in "cold" without knowing any more about it. (You still bring your cultural knowledge and preconceptions with you, though.) A simple example of this might be wanting to watch a movie without spoilers.
Alternately, you can use knowledge of its creation and related history to put it into context. You can judge it as human performance, assuming it was made under certain conditions.
It's possible for this to be entirely faked and for it to still be art. Hollywood movies are largely fakery; that's how the game works. Or, consider music that's made using studio wizardry; it's possible to make songs that couldn't be performed live.
I think it comes down to having some ground rules that are understood by both the creators and viewers. A fictional movie about a soccer game just doesn't operate by the same rules as a real soccer game. It's not wrong in some cosmic sense to lip sync during a live performance, as long as everyone understands what kind of performance they're watching.
When there's an easy way and a hard way to get similar results, provenance matters for appreciating what you're looking at. I expect that some artists are going to take video of their creative processes more often, as a way of helping people better appreciate how it was done.
Also, whenever AI-generated art is shared, often people want to know the prompt. If you know what they asked for then you have a better idea of how good a job the AI is doing.
I understand the outrage with an AI used to reproduce the style of an existing artist. However, things like this are going to revolutionize fan-made content. Imagine having infinite public domain...
I understand the outrage with an AI used to reproduce the style of an existing artist. However, things like this are going to revolutionize fan-made content. Imagine having infinite public domain art for fanfictions, comics, independent games, RPG campaigns...
Is there a list of such generators that are available to the public?
It seems inevitable that animation studios will adapt to use AI in spite of short-term backlash from various sources, just like they've done with 3D and computer generated/computer assisted...
It seems inevitable that animation studios will adapt to use AI in spite of short-term backlash from various sources, just like they've done with 3D and computer generated/computer assisted animation.
CG in anime is still not 100% seamless, as in you can probably tell which action scenes use CG, but it's become ubiquitous. Meanwhile the backlash has more or less gone away as artists and technology have simultaneously improved to make the CG higher and higher fidelity. Simply put, CG lets artists/directors realize their visions for larger and more complicated scenes with less human labor.
There is so much tedious gruntwork in animation that AI might someday be recognized as a holy grail of cost cutting. Animation studios have already been outsourcing to cheaper countries for a long time, and big-time manga artists similarly often have apprentices and ghost writers that they delegate illustration tasks to. Anime outsourcing is notorious for having hugely unpredictable quality—recall the Higurashi no Naku Koro ni anime with some episodes of hilariously bad character art and warped faces where they look like a completely different person.
Imagine if the original studio could both cut costs by not having to pay for outsourcing, and on top of that, guarantee a consistent style every time a character is rendered. Only needing a quick once-over by an experienced human artist to fix the details. Game changer.
Now as for some pessimism, I worry what will happen to artists in training. Great artists often come about by working alongside great artists from a generation before. What if the industry is able to cut two-thirds of the less experienced artists from the workforce, focusing on retaining the high-level creatives? Then would the industry as a whole only produce a third as many master artists as it does today?
It would be too ironic if AI, hailed as a tool to free up humans' time on unimportant busywork and allow for creative work and leisure, had a hand in stifling that creativity by cutting off a knowledge transfer pathway that people need to reach that high level in the first place. Unless we can teach AI to teach us?
I don’t really get the pessimism. People start drawing because it’s fun. There are plenty of books on how to draw and people start learning it as children. The people who are more serious have...
I don’t really get the pessimism. People start drawing because it’s fun. There are plenty of books on how to draw and people start learning it as children. The people who are more serious have plenty of resources available. Amateurs will be able to take on more ambitious projects with new tools.
It’s hard to imagine any techniques being lost because nothing is hidden, It’s all there on the page. Even if you don’t know exactly how it was done, someone who cares enough should be able to figure it out, or come up with something equivalent.
I'm a photographer, and I think AI art is in the same infant stage that art photography went through in the 1800s. Photography was initially seen as a passive, mechanical, soulless reproduction of reality. As such, it lacked the necessary qualities to be considered as art. Photographers responded by trying to make their photographs more painterly. They used soft focus, staged scenes, elaborate manual processes in the darkroom that required many steps and lots of craftmanship, all in search of elevating the mechanical to Art. But this movement, although it produced some fine work, was eventually abandoned.
People didn't stop making photographs that resemble paintings; some artists still do. But later work in that vein was a lot more self aware. Early art photography tried to emulate an existing art form in order to borrow its cultural cachet. But this would only ever be seen as a pale imitation of true art by most critics and the general public. Photography didn't really come into its own until artists began to explore its unique characteristics. Any new medium that aspires to be art needs to identify what makes that medium better suited to certain forms of expression than other media. Why make a photograph when you could make a painting? "Because I suck at painting" might be an honest answer, but it's not a good raison d'etre for a form of art that aspires to be taken seriously.
Art photography became accepted as art when photographers demonstrated that the medium has qualities that are different from, and difficult to replicate in, other forms of art.
Today, AI generated art is also seen as a mechanistic, soulless reproduction. Not of reality, as early photography was, but of other, real, human artists' work (or of whatever else was in the training data, including, funny enough, photographs). A kind of wonky, inferior, unskilled reproduction of another medium that possesses real value. If AI art is to have value, beyond the mere convenience of not having to pay a real human who spent years perfecting the craft of illustration, it needs to figure out how to be itself. What sorts of expressions it's uniquely suited to. What kinds of techniques it enables to express human thought and emotion in ways that cannot easily be replicated in other media.
In short, AI art needs to find itself a soul. I expect in the coming decades as the technology itself matures, with fewer wonky artifacts and limitations, human artists will start to approach it as a novel form of expression. Not just as a cheap way of replicating existing styles, but as a way of creating something that could not previously have existed.
That sort of thing won't go away. Just like most people who take photographs with their iPhone just want a naive replication of reality, possibly with a rose-tinted-glasses filter on top to make it look prettier, but still conceived of as being basically a replication of reality slightly improved. People will still use AI to generate fanart that closely mimics an artist's original style, or fake Picassos, or Generic Corporate Stock Photographs But You Actually Don't Have to Pay a Human Being to Create a Photo of a Middle-Aged Dude Holding an Arthritis Medication (tm). But I'd be surprised if, in a few decades, we aren't also seeing artists exploring AI-assisted art as a unique medium chosen because it enables unique forms of expression.
A lot of what makes art art is that it's a human process that involves work, struggle, and life experience. And whether the end consumer understands it themself, the circumstances of the art piece's creation (i.e. the "behind the scenes", the story of how it came together/got funded, the minds behind the piece, etc.) will influence how the art is experienced by influencing how and among whom it's talked about, how it's critically analyzed, and how and to what extent it's disseminated to the general public. Art is inherently a social process in both its creation and consumption, and I think the use of AI to confuse how and by whom art is being created and to undervalue the craftsmanship of those who have studied the form and serve to preserve its history is truly a slow descent into madness.
These are interesting and thought provoking points. I've been thinking a lot about this since that piece made by Midjourney AI was used to win the Colorado State Fair Fine Arts Competition. They question I have to ask myself is "what is the point of art?" The answer I have come up with is that art in all its forms and in its broadest sense is a way to take stuff from inside your head and find a way to share it with other people. This answer has come to shape how I see this issue. In this way I cannot agree with you more art IS social. Right now, in my opinion the way we see the AI being used, diminishes the craftsmanship of art, but not art itself. And I emphasis right now, because I can see this changing in the future. However how I see these tools being used is to allow a greater pool of people access to the ability to share their ideas with other people. Before cameras only those with immense skill and craftsmanship could convert the beauty that they saw in the world into images that could be enjoyed by others. Now almost anyone can capture a moment of personal significance in a way that can be shared with others. Cameras took a slice of art that was previously locked behind years of skill crafting and gave it to the masses. Now we have AI coming in and allowing anyone to using only a few words, and a selection process the ability to construct an image from their mind into something that can be shared with others. When does art stop becoming art? Is there some level of mechanization in the process of creation that invalidates the results? In this way I go back to where I started, and to your words that resonant with me "Art is inherently a social process." I think as long as a person is involved, as long as it's human ideas being given form, as long people can be connected through it, it is art.
It isn't. The latest correspondence chess world championship (which allows the use of computer assistance) just finished. The winner had 2 wins and 14 draws. Of those two wins, one win was from a guy who resigned nearly all his games in drawn positions. The other was from an opponent who accidentally inputed the wrong move into the chess server. The entire tournament had maybe one or two "legitimate" wins as in one player actually outplayed the other. "Centaur chess" is dead.
I believe the question "what is art?" is not a productive one. The category "art" is historically broad and inherently inclusive. Whenever the question "is this art?" is made, odds are that we will eventually arrive at yes.
It may be more productive to ask, instead, is this fair?. From a legal, social and economic viewpoint, is it fair for someone to profit from the faithful reproduction of a living artist's style? If I had the ability to perfectly reproduce the likeness of an actress (in visual, audio, and performance), in a manner that is somewhat automatic, requiring much less effort, would it be moral to do so without their permission? Wouldn't I be profiting from their life experiences, their body of work, their actual body in and of itself?
I think it's up to the viewer how to approach it. You can certainly try to judge art based solely on the results, going in "cold" without knowing any more about it. (You still bring your cultural knowledge and preconceptions with you, though.) A simple example of this might be wanting to watch a movie without spoilers.
Alternately, you can use knowledge of its creation and related history to put it into context. You can judge it as human performance, assuming it was made under certain conditions.
It's possible for this to be entirely faked and for it to still be art. Hollywood movies are largely fakery; that's how the game works. Or, consider music that's made using studio wizardry; it's possible to make songs that couldn't be performed live.
I think it comes down to having some ground rules that are understood by both the creators and viewers. A fictional movie about a soccer game just doesn't operate by the same rules as a real soccer game. It's not wrong in some cosmic sense to lip sync during a live performance, as long as everyone understands what kind of performance they're watching.
When there's an easy way and a hard way to get similar results, provenance matters for appreciating what you're looking at. I expect that some artists are going to take video of their creative processes more often, as a way of helping people better appreciate how it was done.
Also, whenever AI-generated art is shared, often people want to know the prompt. If you know what they asked for then you have a better idea of how good a job the AI is doing.
I understand the outrage with an AI used to reproduce the style of an existing artist. However, things like this are going to revolutionize fan-made content. Imagine having infinite public domain art for fanfictions, comics, independent games, RPG campaigns...
Is there a list of such generators that are available to the public?
It seems inevitable that animation studios will adapt to use AI in spite of short-term backlash from various sources, just like they've done with 3D and computer generated/computer assisted animation.
CG in anime is still not 100% seamless, as in you can probably tell which action scenes use CG, but it's become ubiquitous. Meanwhile the backlash has more or less gone away as artists and technology have simultaneously improved to make the CG higher and higher fidelity. Simply put, CG lets artists/directors realize their visions for larger and more complicated scenes with less human labor.
There is so much tedious gruntwork in animation that AI might someday be recognized as a holy grail of cost cutting. Animation studios have already been outsourcing to cheaper countries for a long time, and big-time manga artists similarly often have apprentices and ghost writers that they delegate illustration tasks to. Anime outsourcing is notorious for having hugely unpredictable quality—recall the Higurashi no Naku Koro ni anime with some episodes of hilariously bad character art and warped faces where they look like a completely different person.
Imagine if the original studio could both cut costs by not having to pay for outsourcing, and on top of that, guarantee a consistent style every time a character is rendered. Only needing a quick once-over by an experienced human artist to fix the details. Game changer.
Now as for some pessimism, I worry what will happen to artists in training. Great artists often come about by working alongside great artists from a generation before. What if the industry is able to cut two-thirds of the less experienced artists from the workforce, focusing on retaining the high-level creatives? Then would the industry as a whole only produce a third as many master artists as it does today?
It would be too ironic if AI, hailed as a tool to free up humans' time on unimportant busywork and allow for creative work and leisure, had a hand in stifling that creativity by cutting off a knowledge transfer pathway that people need to reach that high level in the first place. Unless we can teach AI to teach us?
I don’t really get the pessimism. People start drawing because it’s fun. There are plenty of books on how to draw and people start learning it as children. The people who are more serious have plenty of resources available. Amateurs will be able to take on more ambitious projects with new tools.
It’s hard to imagine any techniques being lost because nothing is hidden, It’s all there on the page. Even if you don’t know exactly how it was done, someone who cares enough should be able to figure it out, or come up with something equivalent.