Question: does this program only produce digital files? If so, then it’s missing an essential element in painting… brush strokes. The quality of those on a canvas are an important element in art. Real art
AI exists in an abstract, disembodied world. We often do that, too: look at Putin, or Liz Truss, who naively expect the world to conform to their schemata. But then, so do scientists, who as a group tend to be right. Artists generally do not put stock in abstract schemas; the very point of their work is to find an opening no one else has noticed and to make something unpredictable with it. If art doesn't surprise it cannot delight. Approaching art always means not seeing, being mystified, have to solve some kind of puzzle, which if done lifts you to a higher, more demanding mystification. The unknowing never departs from art; at the very least, it consists of how we have changed in the time since we first encountered a work. AI is never going to tell us who we are, and for the same reason it cannot create the species of communication we call art.
Very interesting article, well thought out. One request: when linking to something outside the article make sure the link goes to a freely available site. I don't subscribe to the NY Times and don't want to do so. The image of the winner of the Colorado State Fair is available from the Smithsonian site, which is free to all. It should have been used instead.
This article expresses nicely an opinion on a subject that can never really be settled- “what is art”. Even before AI there were artists like Duchamp presenting “found art” and discussions about how much of meaning could be ascribed to the artist who created the signifier and the audience member who determined what was signified.
WK provides a compelling article for ignoring the hand wringing about AI and machine learning. Qoholeth said “Meaningless meaningless, everything is meaningless.” Art is derivative until it isn’t, profound until it isn’t, skillful until it isn’t, you get the drift. No one, including MK has defined art in a non-trivial way. There is a reason for that. The attempt is self-referential and recursive. We can’t even adequately define what is real, let alone figure out art.
1 - Much of art is in what the viewer brings to the piece. AI is more than capable of presenting a canvas for the viewer to bring themselves to.
2 - Certainly you are entitled to your opinion as I am to mine but who are either one of us to pronounce what is and is not "art" definitively for others...
Art has two types of value: the value it gives to the audience, and the value it gives to its creator. Humans create art because it's fun and rewarding for them as the individual. The process of creation is itself the reward. The only value AI "art" provides is value to the audience. People can't enjoy the process of making AI art because they're not the ones making it -- a godless, soulless machine is.
Mike Solana has a great piece about DALL-E called "Demonic" that tackles this same issue.
Art is a touchy subject. Some people consider Jackson Pollock's work to be "art". It's just random drippings of house paint onto canvasses -- something anyone can do. Emphasis on "random". What makes his drippings more valuable works of "art" than mine? Pablo Picasso famously quipped that it took him five years to learn to paint like Rafael, but a lifetime to learn to paint like a child. Think about that for a minute. That means that, in Picasso's opinion at any rate, any child is a good as Picasso.
I've going to be contrarian because I've been using Midjourney for a while which I prefer to DALL-E. Though you make some good points, this is also a bit of a "get off my lawn" article. Let me ask you a question, what exactly is art? Does it require a specific artistic process, or is it in the eye of the beholder? The AI piece that won the prize. If it were in a museum and spectators admired it along with other art pieces, is it art? I would say yes but it should be (and will be) put into a separate category of art just as photography goes into a different category than a sculpture. You bring up music and I have the same question. I'm a musician. Is music about the process of music, or is it about what listeners think is music and like to listen to? Structurally it is music regardless of the process behind it. On that note, many musicians these days use samples and pre-set beats and other such things. Furthermore any musician who goes into the studio knows that the end product may be entirely different than the beginning product due to the production process and a good producer/engineer.
You put up some very poor examples of artwork done and people are saying things like, "that doesn't look like a Picasso!" On Midjourney you would start with 4 versions of the same idea, all very different. You can then hone in on one if you like where it is going and redo 4 more based off of the one you pick. Sometimes they are all terrible. For a musician like myself who is horrible at art, it's a lot of fun. It's also fun to see what in the world it comes up with off of cues. It's fun to try to stump it. Furthermore, with Midjourney there are very complex commands if you learn how to use it that when combined with other commands can be quite interesting.
I would also disagree with you on creativity. It's very good for the brain to think of creative things and it becomes a bit of a collaboration between the user and the AI. So while I get that a person doesn't go through the formal art process (whatever exactly that means) it can bring a lot of joy. Again, you seem to imply that art is only the process of the artist and not in the eye of the beholder. As a musician I understand and love the writing process but I can't put down someone who has used a great deal of technology and "loops" to create a song that inspired the public and goes top 10. A lot of the comments on here... I would suggest people give Midjourney a try and make your own conclusions. I find it extremely therapeutic. Again (I say again), it will be and should be put into a separate category as far as art goes, but it can still be considered art. IMO.
Do we as artists , as makers, just give up? That is my basic instinct and yet my second instinct , a more persistent instinct is not to give up but hole up. Nestle into one's hermitage, plug one's ears with Odysseus' wax and carry on. This is an incredibly difficult article to read and yet one I will be sharing broadly to my fellow artists who toil away , often in obscurity.
I am amazed by the comments here, many of which border on the Luddite.
First, the author (and many comments) seem to confound the process of generating the art with its value to the creator or its aesthetic worth or evaluation. This is a mistake. As several authors pointed out, every technological advance has come with critics who claimed that the “cheat” rendered the results “not really art”. What of the camera obscura? Photography? Engraving? Silk screening? Better artificial pigments? Everything is suspect, then it becomes mainstream.
Second, the author’s analogy between DALL-E and the Kimball piano is, frankly, ridiculous. On those old pianos, etc., I never once felt like I was playing something that **I had in my mind**. I’m not a very good artist. Kirn opens this essay by saying that he has always had trouble envisioning the underworld. I’ve always had trouble translating the vision in my mind to the page (not a good artist). These AIs (there are several) allow people like me to do so! Well! With practice, I can generate a visual image that is very close to what I was envisioning in my mind. The best analogy to this would be a police portrait artist, but 100,000x more versatile and better. On top of this, the systems are also quite surprising when given less input and direction. The other day, my kid and I gave it an input and got back something both excellent, haunting, and completely unexpected.
Third, in the next several years, these AIs will almost completely eliminate the need for most paid commercial artists in corporations. Why pay logo designers? 100% these systems will do such tasks, if not better, then vastly cheaper. What about things like advertising borders, images, etc.? All this will be replaced. The idea that fashion will be immune? Risible. 33% of all retail runs through Walmart. Cost is of paramount importance. If they can shave a few cents off the design cost, they absolutely will. New design for Hawaiian shirt? Done. New pattern for paisley? Done. Will AI take over at Versace? No. That is not necessarily because it will never be able to, but because that is brand incompatible.
Fourth, returning to “creativity”. Seems to me that the Turing Test should still apply. Right now, the computer AI still aren’t quite there… maybe? From the limited testing that I’ve seen, many people can still discern AI generated from human generated art. However, humans have no right to their hubris. That is exactly the fallacy of AI. So said everyone about everything that computers currently exceed humans at: first chess, then Go. Then people said that computers wouldn’t be able to play something like Poker, because computers wouldn’t be able to handle bluffing, etc. And then someone promptly designed a computer that beats the best poker players at poker.
Prediction: The machines will come up with things that are every bit as creative as human artists. Maybe not in the next 5 years. But in our lifetime? Absolutely. Computers can generate art faster than any human artist. If someone wants to train the system to produce “fine, creative, art” how does a human learn to do that? Through critique. That is *exactly* how a typical ML/AI learns, too. Produce art. “Critique” by providing feedback: is it “creative” (as judged by experts), “good” etc. System learns from this, and produces more things that are judged as creative, etc. It will develop a “style” as well. The best systems will be as good as the people who trained it, more versatile than any human artist, and faster at producing art than the fastest human.
Great article and raises the right arguments. But let’s be honest: there are so many human artists who think they are creating art but really just follow expectations of what is cute. They never take risks nor experience devine inspiration. And their skills may be inferior to those of AI. DALL-E is a logical step in the world that values safe places and thought conformism. Taking risks and following inspiration would get you banned and ostracized.
Question: does this program only produce digital files? If so, then it’s missing an essential element in painting… brush strokes. The quality of those on a canvas are an important element in art. Real art
AI exists in an abstract, disembodied world. We often do that, too: look at Putin, or Liz Truss, who naively expect the world to conform to their schemata. But then, so do scientists, who as a group tend to be right. Artists generally do not put stock in abstract schemas; the very point of their work is to find an opening no one else has noticed and to make something unpredictable with it. If art doesn't surprise it cannot delight. Approaching art always means not seeing, being mystified, have to solve some kind of puzzle, which if done lifts you to a higher, more demanding mystification. The unknowing never departs from art; at the very least, it consists of how we have changed in the time since we first encountered a work. AI is never going to tell us who we are, and for the same reason it cannot create the species of communication we call art.
Very interesting article, well thought out. One request: when linking to something outside the article make sure the link goes to a freely available site. I don't subscribe to the NY Times and don't want to do so. The image of the winner of the Colorado State Fair is available from the Smithsonian site, which is free to all. It should have been used instead.
This article expresses nicely an opinion on a subject that can never really be settled- “what is art”. Even before AI there were artists like Duchamp presenting “found art” and discussions about how much of meaning could be ascribed to the artist who created the signifier and the audience member who determined what was signified.
I highly recommend the "America This Week" with Matt Taibbi & Walter Kirk.
It is both the funniest and most depressing thing you'll hear each week.
WK provides a compelling article for ignoring the hand wringing about AI and machine learning. Qoholeth said “Meaningless meaningless, everything is meaningless.” Art is derivative until it isn’t, profound until it isn’t, skillful until it isn’t, you get the drift. No one, including MK has defined art in a non-trivial way. There is a reason for that. The attempt is self-referential and recursive. We can’t even adequately define what is real, let alone figure out art.
Sorry :
1 - Much of art is in what the viewer brings to the piece. AI is more than capable of presenting a canvas for the viewer to bring themselves to.
2 - Certainly you are entitled to your opinion as I am to mine but who are either one of us to pronounce what is and is not "art" definitively for others...
3 - The lady doth protest too much.
Art has two types of value: the value it gives to the audience, and the value it gives to its creator. Humans create art because it's fun and rewarding for them as the individual. The process of creation is itself the reward. The only value AI "art" provides is value to the audience. People can't enjoy the process of making AI art because they're not the ones making it -- a godless, soulless machine is.
Mike Solana has a great piece about DALL-E called "Demonic" that tackles this same issue.
Art is a touchy subject. Some people consider Jackson Pollock's work to be "art". It's just random drippings of house paint onto canvasses -- something anyone can do. Emphasis on "random". What makes his drippings more valuable works of "art" than mine? Pablo Picasso famously quipped that it took him five years to learn to paint like Rafael, but a lifetime to learn to paint like a child. Think about that for a minute. That means that, in Picasso's opinion at any rate, any child is a good as Picasso.
I've going to be contrarian because I've been using Midjourney for a while which I prefer to DALL-E. Though you make some good points, this is also a bit of a "get off my lawn" article. Let me ask you a question, what exactly is art? Does it require a specific artistic process, or is it in the eye of the beholder? The AI piece that won the prize. If it were in a museum and spectators admired it along with other art pieces, is it art? I would say yes but it should be (and will be) put into a separate category of art just as photography goes into a different category than a sculpture. You bring up music and I have the same question. I'm a musician. Is music about the process of music, or is it about what listeners think is music and like to listen to? Structurally it is music regardless of the process behind it. On that note, many musicians these days use samples and pre-set beats and other such things. Furthermore any musician who goes into the studio knows that the end product may be entirely different than the beginning product due to the production process and a good producer/engineer.
You put up some very poor examples of artwork done and people are saying things like, "that doesn't look like a Picasso!" On Midjourney you would start with 4 versions of the same idea, all very different. You can then hone in on one if you like where it is going and redo 4 more based off of the one you pick. Sometimes they are all terrible. For a musician like myself who is horrible at art, it's a lot of fun. It's also fun to see what in the world it comes up with off of cues. It's fun to try to stump it. Furthermore, with Midjourney there are very complex commands if you learn how to use it that when combined with other commands can be quite interesting.
I would also disagree with you on creativity. It's very good for the brain to think of creative things and it becomes a bit of a collaboration between the user and the AI. So while I get that a person doesn't go through the formal art process (whatever exactly that means) it can bring a lot of joy. Again, you seem to imply that art is only the process of the artist and not in the eye of the beholder. As a musician I understand and love the writing process but I can't put down someone who has used a great deal of technology and "loops" to create a song that inspired the public and goes top 10. A lot of the comments on here... I would suggest people give Midjourney a try and make your own conclusions. I find it extremely therapeutic. Again (I say again), it will be and should be put into a separate category as far as art goes, but it can still be considered art. IMO.
You'll have to forgive me, but until I've seen what an AI NYT art critic writes about the work, I'm not sure what to think.
Do we as artists , as makers, just give up? That is my basic instinct and yet my second instinct , a more persistent instinct is not to give up but hole up. Nestle into one's hermitage, plug one's ears with Odysseus' wax and carry on. This is an incredibly difficult article to read and yet one I will be sharing broadly to my fellow artists who toil away , often in obscurity.
I am amazed by the comments here, many of which border on the Luddite.
First, the author (and many comments) seem to confound the process of generating the art with its value to the creator or its aesthetic worth or evaluation. This is a mistake. As several authors pointed out, every technological advance has come with critics who claimed that the “cheat” rendered the results “not really art”. What of the camera obscura? Photography? Engraving? Silk screening? Better artificial pigments? Everything is suspect, then it becomes mainstream.
Second, the author’s analogy between DALL-E and the Kimball piano is, frankly, ridiculous. On those old pianos, etc., I never once felt like I was playing something that **I had in my mind**. I’m not a very good artist. Kirn opens this essay by saying that he has always had trouble envisioning the underworld. I’ve always had trouble translating the vision in my mind to the page (not a good artist). These AIs (there are several) allow people like me to do so! Well! With practice, I can generate a visual image that is very close to what I was envisioning in my mind. The best analogy to this would be a police portrait artist, but 100,000x more versatile and better. On top of this, the systems are also quite surprising when given less input and direction. The other day, my kid and I gave it an input and got back something both excellent, haunting, and completely unexpected.
Third, in the next several years, these AIs will almost completely eliminate the need for most paid commercial artists in corporations. Why pay logo designers? 100% these systems will do such tasks, if not better, then vastly cheaper. What about things like advertising borders, images, etc.? All this will be replaced. The idea that fashion will be immune? Risible. 33% of all retail runs through Walmart. Cost is of paramount importance. If they can shave a few cents off the design cost, they absolutely will. New design for Hawaiian shirt? Done. New pattern for paisley? Done. Will AI take over at Versace? No. That is not necessarily because it will never be able to, but because that is brand incompatible.
Fourth, returning to “creativity”. Seems to me that the Turing Test should still apply. Right now, the computer AI still aren’t quite there… maybe? From the limited testing that I’ve seen, many people can still discern AI generated from human generated art. However, humans have no right to their hubris. That is exactly the fallacy of AI. So said everyone about everything that computers currently exceed humans at: first chess, then Go. Then people said that computers wouldn’t be able to play something like Poker, because computers wouldn’t be able to handle bluffing, etc. And then someone promptly designed a computer that beats the best poker players at poker.
Prediction: The machines will come up with things that are every bit as creative as human artists. Maybe not in the next 5 years. But in our lifetime? Absolutely. Computers can generate art faster than any human artist. If someone wants to train the system to produce “fine, creative, art” how does a human learn to do that? Through critique. That is *exactly* how a typical ML/AI learns, too. Produce art. “Critique” by providing feedback: is it “creative” (as judged by experts), “good” etc. System learns from this, and produces more things that are judged as creative, etc. It will develop a “style” as well. The best systems will be as good as the people who trained it, more versatile than any human artist, and faster at producing art than the fastest human.
Like so much related to technology today, this piece makes a cabin in the woods with no internet look better and better. 😆
Great article and raises the right arguments. But let’s be honest: there are so many human artists who think they are creating art but really just follow expectations of what is cute. They never take risks nor experience devine inspiration. And their skills may be inferior to those of AI. DALL-E is a logical step in the world that values safe places and thought conformism. Taking risks and following inspiration would get you banned and ostracized.
If this is art, then all Hail Hunter the new Van Gough!
Also, get the elephants and dogs paint brushes as we can flood the market and make millions!
When will people wake up and become somewhat normal again?