aw man. I think AI is neat, but you're right that people oversell what it actually is capable of. I first got into AI as a way to facilitate more randomness in roleplay groups--so basically, as a glorified generator--and I still think that's what it's best for. It helps me with brainstorming when I'm stuck and I know illustrators who use it the same way, to quickly come up with thumbnails that they'll springboard off of.
I also know someone who creates AI art because she has health problems that prevent her from spending the amount of time she used to on painting and drawing. She's up front about this fact but she gets soooooo much hate and suicide baiting for it.
Which, of course, shows that you're right that people want art that's created by real humans. But still, if people want to be certain whether content is created by an AI or a human, they should probably not suicide bait the people who admit they use AI... that's not going to encourage more people to disclose, you know?
And your book looks right up my alley, oh no. Guess I know what I'm asking for for mother's day. hey, kids....
Yes, bullying people for using AI is petty at best, and vicious at worst. There's nothing immoral about using AI to create content. It's a commercial question. Let people try whatever they want and let the market decide what succeeds. If is actually works, and if its work actually satisfies, then it will triumph in the market. If not, it won't. No amount of bullying with change that one way or the other.
This might be my favorite of your posts so far. Thank you for writing it. It showed up in my inbox just when I needed it; I've been struggling to make my own art in light of the concerns you mentioned, among others.
Also, I have my copy of ISABEL already and am looking forward to reading it!
E.T.A.: This is also why the idea of chatbots instead of real therapists is really depressing to me.
I'm not much of a fan of therapists to begin with, but the idea of a robot therapist is utterly chilling. If anything could be guaranteed to drive a sane person to raving madness, it would be a robot therapist. Loneliness is the great curse of our age. What could put the cherry on top of our loneliness more than a robot therapist?
I appreciate your optimism about surviving AI but it seems quaint somehow. The analogy of watching sports and needing to know a real human is involved is a compelling thought, but since sport is form of competition, I found myself thinking about war. Once, to be a warrior required a certain set of abilities and qualities such as courage and valor. War was once a near ritual with armies facing off on "fields of battle" and I would say that there was a time when war was only partially about winning. It was also about the glorious forms that were employed and the heroism of its participants. That sort of thinking became ridiculously outmoded a long time ago. Today, war can be prosecuted from an office by office workers who operate drones and other robots that deliver deadly strikes with pinpoint accuracy thousands of miles away, and in the modern way of thinking, nobody cares about bravery or heroism or the techniques that are used as long as your side wins. I think AI will be like that in all the fields of human endeavor. On the subject of the arts, every art form is already dominated by "pastiche" which I think is another word for "sampling." This has become the standard approach and the vast majority of consumers of movies and comics and novels do not seem particularly bothered by the fact that we have musicians who can't play an instrument and visual artists who rely on borrowed images and motifs, and movies that are stylized remakes of older movies and novelists who, I assume, are merely apps. I think the arts are becoming ever cheaper, more disposable and phonier, most people think of "art" purely as entertainment now, and I'm horrified that AI will accelerate all these trends.
I think you may be premature in writing the obituary of bravery and heroism. I think you will find great appreciation for those virtues in Israel and Ukraine, for instance.
But it is certainly true that we now value sportsmen more than we do soldiers. But then, sport has always existed as training for war. Football and soccer are games of open field battle. Baseball and cricket are games of siege warfare. All of the track and field disciplines are martial skills. And the soldierly ideals of fitness and courage apply equally to the athlete. We didn't lose our admiration for these skills, we simply transferred it from soldiers to athletes.
But you are right that there is a lot of pastiche in the arts these days. We might make a distinction in the content industries between art and what we could call soma. The purpose of art is to change how you see. The purpose of soma is to soothe the addled mind. And in terms of volume, soma vastly outweighs art in the market today. So there are really two questions. Can AI make art, and can AI make soma.
I am confident that AI cannot make art, not because of an inherent limits of its capabilities, though I do expect that its limits are inherent and will remain, but because art by its nature is a connection between humans. It is a matter not merely of doing, but of being.
Can AI make soma? Not well, not yet. Will it be able too in the future? I don't know. Is soma something in which the human element truly does not matter, a matter merely of doing, not of being. There is, after all, nothing that soothes the addled mind better than a hug or another's hand in your own. But maybe soma is what you turn to when a hug or a hand in yours is not available. I don't know.
I wish I was as optimistic. Real art vs. soma? Who's to judge? Certainly, in the visual realm, a lot of AI-created art looks professional. Deep fakes are said to be so good that they even fool the experts. Almost everyone's wondering what's really real on the internet. Your visual for Unruly Magic would have taken a professional illustrator many hours and may not have turned out nearly as well. In journalism AI is making huge inroads they tell me. How can I even know? Is literature too emotionally deep or wonderfully humanistic to be overthrown by bots? I have my doubts. There's another thing about these emerging technologies that worry me. Whatever they create is fast and cheap. Already the internet is overflowing with trash. We have to wade through miles of junk with every new search. AI is going to be a junk creator on steroids, so the artist who spends a year on a novel will be completing with hundreds of thousands of new titles that will be coming out simultaneously and his chance of being noticed will be vanishingly small. Sorry. I hate to take the Debbie Downer position on this stuff but that's what I expect will happen. I continue to write fiction but not with high hopes. It's hard to see a way out of this dilemma, but if there is one I think it will have to involve going small. The only way to beat the machine I think will be through a return to traditional media. I do digital illustrations for my novels and stories and I'm spoiled by the facility of working that way, but I keep telling myself I really ought to go back to pencil and paper and paint and canvas and I should use imagery that's as far from photorealism as possible. As writers, I imagine a possible future when we produce our work on small presses, letterpresses and the like. We'll design our books ourselves or with the help of traditional artisans, make small editions of quality that can be sold for a good price. And again, write as truthfully and humanely as we can about subjects that (we hope) a machine will never be able to mimic or understand.
The Unruly Magic image is a good case in point. At first glance, it looks complex. But take just a moment to look at the swirling figures and you will see that they are just half-formed swirls and blobs. A real artist would have drawn nightmare creatures. AI just drew blobs with an eye here or a hand there, but nothing substantial or connected. Whatever the claims may be, most AI art is pathetically easy to spot.
But again, my main point is not based on AI art being pathetically easy to spot, though I expect it will continue to be so, but on human beings demanding to know that the art they are getting is human-created because their whole reason for wanting art is the human connection it gives. My confidence, therefore, has nothing to do with how good AI art gets.
What we should note about soma, is that it is not going to flood the internet. That is a temporary phenomena. If AI gets good enough at producing soma, then it won't be independent writers and artists using AI to put soma on the web. It will big tech companies selling consumers custom soma generators for home use. If AI can make soma to order, there will be no point in shelving generic soma. You will generate soma perfectly suited to your personal tastes and mood of the moment on the fly.
What happens to Amazon KDP and similar services after that will depend on the demand ratio between soma and art. Art will still require a method of publication and distribution, but how big a business that will be remains to be seen.
Of course, it may well prove that people demand a human connection in their soma as well. Certainly the connection between author and reader is considered a key component of book marketing today.
Be of good cheer, therefore, for even if the machine learns to mimic you perfectly, it will still not be you, and in the things that matter most, it is being, not doing, that counts.
You know, I thought a bit about this over the weekend. I'm a writer who also tries to write with depth while being entertaining, and I agree that some people will definitely fall for AI "art." We're already seeing that with the popularity of the most formulaic movies and music, etc. But the key to remember is that not everyone will. And as long as I can continue to sell my books, I'll write for those readers who want something more. I have no idea how long that will be, but I hope it's a very long time.
Perhaps if AI does take over writing soma, then the distinction between soma and art will become more clear and those of us trying to write art will actually have a better defined market to aim at.
This is so great. I'm not worried about AI taking over for human writers. My concerns surrounding AI are more related to this: "We don't want content, we want communication."
I hope this is the case but, I sense the hunger for content--a simple entertaining presence that scratches the itch of loneliness/boredom--may eclipse the need for sincere communication.
It is an interesting question how it will break down. I compared generative AI to a sex doll -- scratching the itch without the human contact. And, of course, sex dolls do exist. There is a market for them, albeit not a large one. Would the mooted sex robot of the future expand that market with a more satisfactory product? Honestly, I'm afraid to ask.
But there is a difference between art and sex. Sex is in limited supply. That is, the number of sexually desirable people who are both willing and available is significantly less than the number of people desirous of sex. All the human-written novels in the world, by contrast, are available for your immediate gratification at the push of a button, and for a very low price. Sex has a supply problem; art does not. If follows that there is a market for substitutes for sex, but no reason for there to be a market for substitutes for art.
Even that particular form of art that serves as a substitute for sex has the same supply characteristics as other art. Its abundance of supply is a noted feature of the web.
To Eric Hoel's point, therefore, the content industries are not lucrative because supply vastly exceeds demand, and always will because once a commodity enters the supply chain today, it never leaves. What void, therefore, are we asking AI art to fill?
The only application for AI art that makes any sense to me is completely individualized art. And yet I can't really see that appealing to people because if you look at the major cultural products of the day, a huge part of the enjoyment people get from them it becoming a member of the fandom, going to conventions, meeting the actors, lionizing the authors. Individualized art could only deepen the loneliness you speak of. Indeed, I think it is reasonable to say that fandom is the single biggest social outlet for all the socially awkward people in the world (of which, I am one). A fandom is the one place you know that everyone wants to talk about the same things you do. AI art would only make the experience less social.
I agree with all of the above. I suppose my concern is more of a cultural trend as opposed to AI specific. The loneliness epidemic coupled with the internet era expectation of instant gratification is what I'm thinking about. Cures for loneliness (communication, intimacy, etc.) can be met instantly by ChatBots, subscription to an AI Partner, paying for an OnlyFans accounts etc. relative to the real McCoy. Do I think this is an immediate concern? No. I worry this may be more of an issue for younger generations who grew up immersed in technology.
I also see the flip side. I work in a crisis call center. When I speak to people they express relief that they are finally speaking to a flesh and blood person. The ubiquity of phone trees and automation have made them crave the human touch. Generally (using only my personal experience here), the folks excited about speaking with a human tend to be older (30+). The younger folks I speak with (born after 2000) seem inconvenienced by the fact that I am a human and request I just email them the resources they want and hang up.
With these generational differences, I always wonder whether the younger people will be just like the older ones once the reach the same age, or whether there is a real fundamental difference between the generations. Most of the time it is the former, of course. But sometimes it is the latter.
I think the cult of the individual, and of independence, that has been with us for a couple of generations now, has been enormously destructive. We are living in the first age and place in history in which it is possible for most people to live alone, without a family to protect them or provide for them. We have made that independence the cardinal virtue of our age. The result has been an epidemic of loneliness, a housing crisis, and a decline in lifestyles, because it is still true that two can live as cheaply as one, but everyone insists on living as one.
Is it that this kind of independence is supportable in your 20s but by the time you hit 30 and discover that you have no permanent relationships and lack the skills to create them you begin to hunger for any human contact? Is this why we have so many people in their mid 30s, both men and women, suddenly desperate to get married and start a family and unable to find anyone to marry, having become to set in their independent ways, despite the misery it brings them. Will robot wives and robot husbands fill the void? I doubt it.
Certainly the literature of our age is deeply infected with the cult of individualism and independence. Could literature help to turn it around? Maybe, but not if it is generated by AI. Or maybe the economic stupidity of the younger generation will reduce our living standards to the point where we need each other again and we can be less lonely and die of disease and starvation again like we used to do.
In any case, will AI cure the epidemic of loneliness. Clearly no. Will we figure out where we got things so wrong in our pursuit of independence and go back to living in families and communities again. We can only hope.
Agreed. I don't think AI is really a "cure" for anything. However, its allure and widespread use could create a permission structure for individuals to be alone. As you mentioned: is this something the younger generation grows out of? Haven't a clue.
I remain hopeful when I encounter someone pleased to speak to a human. I think that impulse to speak from one flesh to another will always be there.
I suppose the danger is that, like drugs, it could lead someone so far astray in their search for solace that they no longer have the capacity to make their way back into society. If so, it could certainly contribute to and perhaps prolong the loneliness epidemic. Perhaps that is the great threat of AI, though it is not the one that gets talked about.
But yes, we will lose individuals, as we always have, but society, and its demand for and capacity to provide human contact and human solace will go on, as it always has.
aw man. I think AI is neat, but you're right that people oversell what it actually is capable of. I first got into AI as a way to facilitate more randomness in roleplay groups--so basically, as a glorified generator--and I still think that's what it's best for. It helps me with brainstorming when I'm stuck and I know illustrators who use it the same way, to quickly come up with thumbnails that they'll springboard off of.
I also know someone who creates AI art because she has health problems that prevent her from spending the amount of time she used to on painting and drawing. She's up front about this fact but she gets soooooo much hate and suicide baiting for it.
Which, of course, shows that you're right that people want art that's created by real humans. But still, if people want to be certain whether content is created by an AI or a human, they should probably not suicide bait the people who admit they use AI... that's not going to encourage more people to disclose, you know?
And your book looks right up my alley, oh no. Guess I know what I'm asking for for mother's day. hey, kids....
Yes, bullying people for using AI is petty at best, and vicious at worst. There's nothing immoral about using AI to create content. It's a commercial question. Let people try whatever they want and let the market decide what succeeds. If is actually works, and if its work actually satisfies, then it will triumph in the market. If not, it won't. No amount of bullying with change that one way or the other.
This might be my favorite of your posts so far. Thank you for writing it. It showed up in my inbox just when I needed it; I've been struggling to make my own art in light of the concerns you mentioned, among others.
Also, I have my copy of ISABEL already and am looking forward to reading it!
E.T.A.: This is also why the idea of chatbots instead of real therapists is really depressing to me.
I'm not much of a fan of therapists to begin with, but the idea of a robot therapist is utterly chilling. If anything could be guaranteed to drive a sane person to raving madness, it would be a robot therapist. Loneliness is the great curse of our age. What could put the cherry on top of our loneliness more than a robot therapist?
I appreciate your optimism about surviving AI but it seems quaint somehow. The analogy of watching sports and needing to know a real human is involved is a compelling thought, but since sport is form of competition, I found myself thinking about war. Once, to be a warrior required a certain set of abilities and qualities such as courage and valor. War was once a near ritual with armies facing off on "fields of battle" and I would say that there was a time when war was only partially about winning. It was also about the glorious forms that were employed and the heroism of its participants. That sort of thinking became ridiculously outmoded a long time ago. Today, war can be prosecuted from an office by office workers who operate drones and other robots that deliver deadly strikes with pinpoint accuracy thousands of miles away, and in the modern way of thinking, nobody cares about bravery or heroism or the techniques that are used as long as your side wins. I think AI will be like that in all the fields of human endeavor. On the subject of the arts, every art form is already dominated by "pastiche" which I think is another word for "sampling." This has become the standard approach and the vast majority of consumers of movies and comics and novels do not seem particularly bothered by the fact that we have musicians who can't play an instrument and visual artists who rely on borrowed images and motifs, and movies that are stylized remakes of older movies and novelists who, I assume, are merely apps. I think the arts are becoming ever cheaper, more disposable and phonier, most people think of "art" purely as entertainment now, and I'm horrified that AI will accelerate all these trends.
I think you may be premature in writing the obituary of bravery and heroism. I think you will find great appreciation for those virtues in Israel and Ukraine, for instance.
But it is certainly true that we now value sportsmen more than we do soldiers. But then, sport has always existed as training for war. Football and soccer are games of open field battle. Baseball and cricket are games of siege warfare. All of the track and field disciplines are martial skills. And the soldierly ideals of fitness and courage apply equally to the athlete. We didn't lose our admiration for these skills, we simply transferred it from soldiers to athletes.
But you are right that there is a lot of pastiche in the arts these days. We might make a distinction in the content industries between art and what we could call soma. The purpose of art is to change how you see. The purpose of soma is to soothe the addled mind. And in terms of volume, soma vastly outweighs art in the market today. So there are really two questions. Can AI make art, and can AI make soma.
I am confident that AI cannot make art, not because of an inherent limits of its capabilities, though I do expect that its limits are inherent and will remain, but because art by its nature is a connection between humans. It is a matter not merely of doing, but of being.
Can AI make soma? Not well, not yet. Will it be able too in the future? I don't know. Is soma something in which the human element truly does not matter, a matter merely of doing, not of being. There is, after all, nothing that soothes the addled mind better than a hug or another's hand in your own. But maybe soma is what you turn to when a hug or a hand in yours is not available. I don't know.
I wish I was as optimistic. Real art vs. soma? Who's to judge? Certainly, in the visual realm, a lot of AI-created art looks professional. Deep fakes are said to be so good that they even fool the experts. Almost everyone's wondering what's really real on the internet. Your visual for Unruly Magic would have taken a professional illustrator many hours and may not have turned out nearly as well. In journalism AI is making huge inroads they tell me. How can I even know? Is literature too emotionally deep or wonderfully humanistic to be overthrown by bots? I have my doubts. There's another thing about these emerging technologies that worry me. Whatever they create is fast and cheap. Already the internet is overflowing with trash. We have to wade through miles of junk with every new search. AI is going to be a junk creator on steroids, so the artist who spends a year on a novel will be completing with hundreds of thousands of new titles that will be coming out simultaneously and his chance of being noticed will be vanishingly small. Sorry. I hate to take the Debbie Downer position on this stuff but that's what I expect will happen. I continue to write fiction but not with high hopes. It's hard to see a way out of this dilemma, but if there is one I think it will have to involve going small. The only way to beat the machine I think will be through a return to traditional media. I do digital illustrations for my novels and stories and I'm spoiled by the facility of working that way, but I keep telling myself I really ought to go back to pencil and paper and paint and canvas and I should use imagery that's as far from photorealism as possible. As writers, I imagine a possible future when we produce our work on small presses, letterpresses and the like. We'll design our books ourselves or with the help of traditional artisans, make small editions of quality that can be sold for a good price. And again, write as truthfully and humanely as we can about subjects that (we hope) a machine will never be able to mimic or understand.
The Unruly Magic image is a good case in point. At first glance, it looks complex. But take just a moment to look at the swirling figures and you will see that they are just half-formed swirls and blobs. A real artist would have drawn nightmare creatures. AI just drew blobs with an eye here or a hand there, but nothing substantial or connected. Whatever the claims may be, most AI art is pathetically easy to spot.
But again, my main point is not based on AI art being pathetically easy to spot, though I expect it will continue to be so, but on human beings demanding to know that the art they are getting is human-created because their whole reason for wanting art is the human connection it gives. My confidence, therefore, has nothing to do with how good AI art gets.
What we should note about soma, is that it is not going to flood the internet. That is a temporary phenomena. If AI gets good enough at producing soma, then it won't be independent writers and artists using AI to put soma on the web. It will big tech companies selling consumers custom soma generators for home use. If AI can make soma to order, there will be no point in shelving generic soma. You will generate soma perfectly suited to your personal tastes and mood of the moment on the fly.
What happens to Amazon KDP and similar services after that will depend on the demand ratio between soma and art. Art will still require a method of publication and distribution, but how big a business that will be remains to be seen.
Of course, it may well prove that people demand a human connection in their soma as well. Certainly the connection between author and reader is considered a key component of book marketing today.
Be of good cheer, therefore, for even if the machine learns to mimic you perfectly, it will still not be you, and in the things that matter most, it is being, not doing, that counts.
You know, I thought a bit about this over the weekend. I'm a writer who also tries to write with depth while being entertaining, and I agree that some people will definitely fall for AI "art." We're already seeing that with the popularity of the most formulaic movies and music, etc. But the key to remember is that not everyone will. And as long as I can continue to sell my books, I'll write for those readers who want something more. I have no idea how long that will be, but I hope it's a very long time.
Perhaps if AI does take over writing soma, then the distinction between soma and art will become more clear and those of us trying to write art will actually have a better defined market to aim at.
This is so great. I'm not worried about AI taking over for human writers. My concerns surrounding AI are more related to this: "We don't want content, we want communication."
I hope this is the case but, I sense the hunger for content--a simple entertaining presence that scratches the itch of loneliness/boredom--may eclipse the need for sincere communication.
It is an interesting question how it will break down. I compared generative AI to a sex doll -- scratching the itch without the human contact. And, of course, sex dolls do exist. There is a market for them, albeit not a large one. Would the mooted sex robot of the future expand that market with a more satisfactory product? Honestly, I'm afraid to ask.
But there is a difference between art and sex. Sex is in limited supply. That is, the number of sexually desirable people who are both willing and available is significantly less than the number of people desirous of sex. All the human-written novels in the world, by contrast, are available for your immediate gratification at the push of a button, and for a very low price. Sex has a supply problem; art does not. If follows that there is a market for substitutes for sex, but no reason for there to be a market for substitutes for art.
Even that particular form of art that serves as a substitute for sex has the same supply characteristics as other art. Its abundance of supply is a noted feature of the web.
To Eric Hoel's point, therefore, the content industries are not lucrative because supply vastly exceeds demand, and always will because once a commodity enters the supply chain today, it never leaves. What void, therefore, are we asking AI art to fill?
The only application for AI art that makes any sense to me is completely individualized art. And yet I can't really see that appealing to people because if you look at the major cultural products of the day, a huge part of the enjoyment people get from them it becoming a member of the fandom, going to conventions, meeting the actors, lionizing the authors. Individualized art could only deepen the loneliness you speak of. Indeed, I think it is reasonable to say that fandom is the single biggest social outlet for all the socially awkward people in the world (of which, I am one). A fandom is the one place you know that everyone wants to talk about the same things you do. AI art would only make the experience less social.
I agree with all of the above. I suppose my concern is more of a cultural trend as opposed to AI specific. The loneliness epidemic coupled with the internet era expectation of instant gratification is what I'm thinking about. Cures for loneliness (communication, intimacy, etc.) can be met instantly by ChatBots, subscription to an AI Partner, paying for an OnlyFans accounts etc. relative to the real McCoy. Do I think this is an immediate concern? No. I worry this may be more of an issue for younger generations who grew up immersed in technology.
I also see the flip side. I work in a crisis call center. When I speak to people they express relief that they are finally speaking to a flesh and blood person. The ubiquity of phone trees and automation have made them crave the human touch. Generally (using only my personal experience here), the folks excited about speaking with a human tend to be older (30+). The younger folks I speak with (born after 2000) seem inconvenienced by the fact that I am a human and request I just email them the resources they want and hang up.
With these generational differences, I always wonder whether the younger people will be just like the older ones once the reach the same age, or whether there is a real fundamental difference between the generations. Most of the time it is the former, of course. But sometimes it is the latter.
I think the cult of the individual, and of independence, that has been with us for a couple of generations now, has been enormously destructive. We are living in the first age and place in history in which it is possible for most people to live alone, without a family to protect them or provide for them. We have made that independence the cardinal virtue of our age. The result has been an epidemic of loneliness, a housing crisis, and a decline in lifestyles, because it is still true that two can live as cheaply as one, but everyone insists on living as one.
Is it that this kind of independence is supportable in your 20s but by the time you hit 30 and discover that you have no permanent relationships and lack the skills to create them you begin to hunger for any human contact? Is this why we have so many people in their mid 30s, both men and women, suddenly desperate to get married and start a family and unable to find anyone to marry, having become to set in their independent ways, despite the misery it brings them. Will robot wives and robot husbands fill the void? I doubt it.
Certainly the literature of our age is deeply infected with the cult of individualism and independence. Could literature help to turn it around? Maybe, but not if it is generated by AI. Or maybe the economic stupidity of the younger generation will reduce our living standards to the point where we need each other again and we can be less lonely and die of disease and starvation again like we used to do.
In any case, will AI cure the epidemic of loneliness. Clearly no. Will we figure out where we got things so wrong in our pursuit of independence and go back to living in families and communities again. We can only hope.
Agreed. I don't think AI is really a "cure" for anything. However, its allure and widespread use could create a permission structure for individuals to be alone. As you mentioned: is this something the younger generation grows out of? Haven't a clue.
I remain hopeful when I encounter someone pleased to speak to a human. I think that impulse to speak from one flesh to another will always be there.
I suppose the danger is that, like drugs, it could lead someone so far astray in their search for solace that they no longer have the capacity to make their way back into society. If so, it could certainly contribute to and perhaps prolong the loneliness epidemic. Perhaps that is the great threat of AI, though it is not the one that gets talked about.
But yes, we will lose individuals, as we always have, but society, and its demand for and capacity to provide human contact and human solace will go on, as it always has.
Thanks for sharing this! Agree with "Art is the expression of human experience."
I wrote similar thoughts in my post below
https://ranas9.substack.com/p/why-am-i-on-substack
Have been collecting my thoughts on AI, art, language and the human experience and this really helped.