i mean, its crazy right? it is completely and totally just a matter of bias and preconceived notions and reactionarism pushing people to arrive at the idea that a person using a generative model soley based on their own work intentionally creating a specific piece of art they have already visualized in their mind is "not art" because... well... um...

well they can't even say! they say "it's not intentional" (it is, he already knows what he wants it to look like!), "it's not work/effort" (god i want you people to see how long it takes me to generate the simplest of anime girls in midjourney), or most bafflingly "the prompt is art but the image output isnt". and nobody can actually explain WHY because there ISNT a good reason! they just have this kneejerk hatred and disgust of this new technology and art form and any attempts at prodding at that feeling are met with aggression or blank stares.

I'M IN CRAZY TOWN

"I would argue the prompt is itself art, but not the output image."

i'll consider agreeing that AI images are art on the condition that we actually start treating them like art. if i ask van gogh "hey what was your thinking behind giving that woman 7 fingers" and he goes "idk the program just does that sometimes" then it's bad art and he's a hack artist, at least in that medium, no different from someone going "idk the paint is just like that sometimes" when asked about the shoddy brushwork and texture of a painting theyve made. dont know OP but a lot of AI image makers seem to want the prestige of Art(tm) when its convenient and hide behind "no its just a funni computer picture" when it's not. can't have it both ways, if this is your medium then take responsibility for it

its very strange you think i do not want ai art to be taken seriously as a medium. why are you speaking so condescendingly to me about this? "prestige of art" girl what? i just want people to recognize one of the mediums i work in is legitimate lol. like "take responsibility" lol? its not that deep. id just like people to concede that just bc you dont Like something doesnt mean its Not Art.

anyways heres a picture i made cus u obviously have a very limited idea of what ai art might be :)

Okay, I mean ... I'm a writer, and have fed LLMs my own writing, and have pushed a button to get written material out of them. Due to the way that these models work, you can't train on only a single artist/writer, because it's just not enough data to get anything coherent, even if they're prolific, but that aside ...

I mean, fine, it's a definitional debate, those are always pretty dumb.

But to me, art is about expression and emotional power much more than it is the mode of production or even really intent. By this metric, there are lots of things that other people would consider to be "art" but I wouldn't: there are plenty of paintings that take technical skill but express nothing and hold no emotion. I think given the nature of painting, it's relatively unusual for a painter to do a full painting without trying to get something across in some way, but I do think it happens.

And the thing is, I don't think that this disqualifies all AI art! I think some of it has expression and emotional power even though it's just a prompt and a model! But I think if we call everything art, there's no point in having the word, it becomes meaningless, and the category of things that I mean when I say art should discount a lot of stuff that has no expression or emotion to it.

So hypothetically, if I had a machine trained only on my stories, and had it write a new story, would it be art?

It would depend on whether it spoke to me or not, whether it had some expression to it, and whether there was a connection to be made. How does it make me feel?

Which means that sometimes when I push the button, I get art, and sometimes (much more often) I do not.

I do not expect this stance to take the world by storm, but I think it helps to explain the distinction in my mind between the AI slop and the stuff that actually makes an impression on me.

Have yall heard of "bad art".

Like. Art that sucks. Art that completely fails at the intended purpose. A joke that's not funny. A drawing that you can't figure out what it was supposed to be. Writing that actively lulls you to sleep. Singing that hurts your ears.

We don't need to invent a separate special category for "whatever they tried, it's not doing anything for me". That's normal for art!!! A child excitingly smearing paint over a piece of paper is very unlikely to get across anything coherent except that they really like drawing. A first-time AI artist is very unlikely to get across anything except that they wanted to see what would happen if a machine drew an image to their specification.

And, hear me out: bad commercial art is a category of art too. Ads are art; they're just ad that generally incites little but hatred, with extremely rare hits. The same is true of AI images that are nothing but an attempt to replace hiring a real human artist for simple illustrations. The person who made them (not the AI. The person who made the prompt) still had to engage with the topic to come up with even the simplest, most boring, basic and failed prompt, and they still make us, the audience, think about the subject matter. (Even if what we're thinking is "I wish I never heard of this thing ever again")

Consider also the concept of "fraud", which is also pre-existing. Remember the story of two little girls who took photographs with fairy cutouts and ended up believed that they met real fairies, and they just kept lying about it? Pretending that something is one kind of art when it's actually another - photoshop job pretending to be a real photo, a fictional story pretending to be something that actually happened. That's not new either. And it doesn't make the fraudulent output not art. It's art with intent to deceive, it's art we might well disapprove of ever being made.

Still art.

The current image generating "AI" is not categorically different from, like, photoshop, with all its filters. And pressing "generate clouds" in photoshop with the default black and white colors, then saving the output, is technically "art you made" - if you want to make a custom screensaver or whatever, this is a perfectly valid way to MAKE one, one that will be your own work that you don't need to credit to anyone else. (You should, of course, not lie about the medium / tool you used to produce it. Unless it would be funny)

I am not willing to accept that every time someone presses the button, the thing that comes out is definitionally art.

But I'm not singling out AI with this!

I don't think that every time someone takes a photograph the product is art! Sometimes people are just taking a photograph to have a reference for later, sometimes it's for verification, sometimes it's to show someone else in an exchange of information, or in the hopes that the camera captures something that their eyes can't, or because they were clumsy and took a picture by accident.

And even then, I accept that sometimes something which was captured by camera on accident, or without intent, can still evoke in me something, and if succeeds in doing that, then I'm still willing to call it art, even if the person who "produced" it doesn't think so, or never knows what they captured.

The thing with the AIs are ... they're just processes, right? And just like we can take human paintings to extremes and ask whether a single brushstroke can be art, or even no brush strokes at all, we can scale back the machines, train them on less data, use fewer nodes. And I don't think it's fruitful to go in a direction where we say that at some certain level then yes, it's art. I think it makes sense to go off whether it makes people feel something, if it does the evocation thing, if it engages.

I think there's a distinction that I personally would draw between "bad art" and "not art", and I think that the vast majority of AI image output is "not art" rather than "bad art", but that's because the vast majority of the output is just people clicking buttons and the machine just spitting things out, and the result is meaningless.

Wait, just press a single button? Since when does machine image generation work like that? I thought you have to input a prompt?

(This is not a rhetorical question. These are meaningfully different situations)

This is a question of definitions.

Can you get an image back with no prompt? Yes, though your interface might not let you, this is not technically difficult to do at all, just something that companies might prohibit. But is a null prompt in fact a prompt? Depends on who you ask.

This is Stable Diffusion from two years ago, and not my own example, but here's one possible result, which depends on the random noise:

I would say that this is squarely not art, and would only change my answer if a sufficient number of people disagreed with me and found some emotional connection, meaning, etc. in it.

Modern Stable Diffusion would give back a better result, in that it would look more like something that existed.

However, I don't necessarily think that "no prompt" is substantially different from having a prompt. The act of choosing a prompt is not what makes an AI image art, in my opinion. In fact, it's trivial to automate prompt generation, and I've done that in the past. You can get GPT to do it.

I think any definition of art where a programmatically random prompt of "cat" is not considered art, but my son's prompt of "cat" is considered art, is probably a bad definition that won't work for actually communicating with people.

P.S. A programmagically random prompt of "cat" is art by whoever wrote the program to randomly choose a prompt.

I've described trying to steer Stable Diffusion as "trying to drive a car by talking to it." It's really great at things like, "hey car, get me to starbucks." Because it knows how to get to starbucks and that's not really conceptually complicated. But if I'm like, hey car, drive towards starbucks for five minutes, but then do a hard u-turn and then pop a wheelie and spin around for a little bit, and then drop by Mike's place, and also get me to McDonalds, it gets really confused. Because individually, all of those things are things it can do, but together, you're basically trying to wrestle this machine into submission and control. The reason AI image generation is handy for a lot of Slop use-cases for a lot of people is because they just want an Image. they don't really care what. And it's really good at that. But the more granular you get, the more it gets confused.

Like. Ok let's take a real example. I'm going to try to generate some images of Artoria Pendragon from Fate. If I ask it something really general, like "Give me artoria pendragon in an epic action battle scene, with full armor," it does that mostly no problem. Like:

This was my second attempt, with very little refinement, and like, it's kind of Nothing, but that's kind of the point of using this stuff the easy way. There are harder ways to use it, though. Like, let's try what I was ACTUALLY trying to do, with immense difficulty, which was to get the AI to make me an image of Artoria Pendragon working at a Subway Sandwich Shop.

As I get more specific, it gets a little more wrong, or wrong in different ways, because it's getting confused. I'm asking it something somewhat unexpected, especially for this model, which was mostly trained on like, anime fanart. So at each step, you can kind of see where it got confused. All of these are a little wrong -- none of them are "Artoria Pendragon standing behind the counter at a Subway restaurant, asking to take my order, making a hoagie." Because that's kind of hard to wrestle the machine into doing.

So, I made this one, with some refinement:

Closest, by far, but still, no cigar. And if you get way too specific, you just get a literal garbled mess of static. You kind of have to be careful with how much you hit the gas. I think of it less like I'm making something and more like I'm negotiating with a very confused preschooler who has relatively little understanding of what I'm actually asking, because they've learned everything they know from like, idk, deviantart or whatever. It's kind of like trying to trick a genie into doing what you want?

Here are the next several iterations I made:

You'll notice again that, we got relatively close, and then I started being too specific, and the computer got extremely confused and -- seemingly but her in an amazon warehouse? wearing overalls? for some reason?

Anyways. AI image generation is, I think, a kind of medium, like any other, but maybe more akin to like, imo, baking. Wherein, many people who engage with it are using ready-made recipes, which are certainly "creative" in the general sense, and certainly takes a decent amount technical awareness and know-how, on some level. but on the other hand, people who make and design baking recipes require a significantly higher level of both technical knowledge and creativity than following a recipe exactly. Both take skill, but one is far more accessible to the average person, at the cost of far less expressive control and capacity for dynamic and creative input. Imagine if I decided I wanted to make the Artoria At Subway image but like, in an art style that I was trying to design myself. The farther you get from the training set, the General Recipe, the more expertise and skill you need to Pilot it. Notably, I don't think this is Unique to AI mediums. For example, during my Slop Project I learned a lot about Garbage YouTube content, and much of it has a similar dynamic (i.e. it's more akin to Baking than it is like, idk, carving a sculpture?) Anyways, I'm going to go back to making Subway Sandwich Shop images until I win against the computer. I'll let you know if I do. Ciao

LET'S FUCKING GOOOOOO VICTORY LAP

THAT IS FUCKING ARTORIA PENDRAGON WORKING AT A HOAGIE SHOP

Bonus Round

Can we talk about this Standwich though? This fuckin Burger King Phone?

when i saw that in the original generation i literally shrieked. its so fucking funny. hello?????

5G Burger

Hear me out- what if you stopped using the plagiarism machine and just drew a fucking picture of Artoria Pendragon working at Subway. As a human, you know what Subway looks like and what she looks like, why waste these resources on this garbage

They're reopening nuclear power plants to power this kind of shut because it's such a drain and look at this bullshit

i mean, its crazy right? it is completely and totally just a matter of bias and preconceived notions and reactionarism pushing people to arrive at the idea that a person using a generative model soley based on their own work intentionally creating a specific piece of art they have already visualized in their mind is "not art" because... well... um...

well they can't even say! they say "it's not intentional" (it is, he already knows what he wants it to look like!), "it's not work/effort" (god i want you people to see how long it takes me to generate the simplest of anime girls in midjourney), or most bafflingly "the prompt is art but the image output isnt". and nobody can actually explain WHY because there ISNT a good reason! they just have this kneejerk hatred and disgust of this new technology and art form and any attempts at prodding at that feeling are met with aggression or blank stares.

I'M IN CRAZY TOWN

"I would argue the prompt is itself art, but not the output image."

i'll consider agreeing that AI images are art on the condition that we actually start treating them like art. if i ask van gogh "hey what was your thinking behind giving that woman 7 fingers" and he goes "idk the program just does that sometimes" then it's bad art and he's a hack artist, at least in that medium, no different from someone going "idk the paint is just like that sometimes" when asked about the shoddy brushwork and texture of a painting theyve made. dont know OP but a lot of AI image makers seem to want the prestige of Art(tm) when its convenient and hide behind "no its just a funni computer picture" when it's not. can't have it both ways, if this is your medium then take responsibility for it

its very strange you think i do not want ai art to be taken seriously as a medium. why are you speaking so condescendingly to me about this? "prestige of art" girl what? i just want people to recognize one of the mediums i work in is legitimate lol. like "take responsibility" lol? its not that deep. id just like people to concede that just bc you dont Like something doesnt mean its Not Art.

anyways heres a picture i made cus u obviously have a very limited idea of what ai art might be :)

Okay, I mean ... I'm a writer, and have fed LLMs my own writing, and have pushed a button to get written material out of them. Due to the way that these models work, you can't train on only a single artist/writer, because it's just not enough data to get anything coherent, even if they're prolific, but that aside ...

I mean, fine, it's a definitional debate, those are always pretty dumb.

But to me, art is about expression and emotional power much more than it is the mode of production or even really intent. By this metric, there are lots of things that other people would consider to be "art" but I wouldn't: there are plenty of paintings that take technical skill but express nothing and hold no emotion. I think given the nature of painting, it's relatively unusual for a painter to do a full painting without trying to get something across in some way, but I do think it happens.

And the thing is, I don't think that this disqualifies all AI art! I think some of it has expression and emotional power even though it's just a prompt and a model! But I think if we call everything art, there's no point in having the word, it becomes meaningless, and the category of things that I mean when I say art should discount a lot of stuff that has no expression or emotion to it.

So hypothetically, if I had a machine trained only on my stories, and had it write a new story, would it be art?

It would depend on whether it spoke to me or not, whether it had some expression to it, and whether there was a connection to be made. How does it make me feel?

Which means that sometimes when I push the button, I get art, and sometimes (much more often) I do not.

I do not expect this stance to take the world by storm, but I think it helps to explain the distinction in my mind between the AI slop and the stuff that actually makes an impression on me.

Have yall heard of "bad art".

Like. Art that sucks. Art that completely fails at the intended purpose. A joke that's not funny. A drawing that you can't figure out what it was supposed to be. Writing that actively lulls you to sleep. Singing that hurts your ears.

We don't need to invent a separate special category for "whatever they tried, it's not doing anything for me". That's normal for art!!! A child excitingly smearing paint over a piece of paper is very unlikely to get across anything coherent except that they really like drawing. A first-time AI artist is very unlikely to get across anything except that they wanted to see what would happen if a machine drew an image to their specification.

And, hear me out: bad commercial art is a category of art too. Ads are art; they're just ad that generally incites little but hatred, with extremely rare hits. The same is true of AI images that are nothing but an attempt to replace hiring a real human artist for simple illustrations. The person who made them (not the AI. The person who made the prompt) still had to engage with the topic to come up with even the simplest, most boring, basic and failed prompt, and they still make us, the audience, think about the subject matter. (Even if what we're thinking is "I wish I never heard of this thing ever again")

Consider also the concept of "fraud", which is also pre-existing. Remember the story of two little girls who took photographs with fairy cutouts and ended up believed that they met real fairies, and they just kept lying about it? Pretending that something is one kind of art when it's actually another - photoshop job pretending to be a real photo, a fictional story pretending to be something that actually happened. That's not new either. And it doesn't make the fraudulent output not art. It's art with intent to deceive, it's art we might well disapprove of ever being made.

Still art.

The current image generating "AI" is not categorically different from, like, photoshop, with all its filters. And pressing "generate clouds" in photoshop with the default black and white colors, then saving the output, is technically "art you made" - if you want to make a custom screensaver or whatever, this is a perfectly valid way to MAKE one, one that will be your own work that you don't need to credit to anyone else. (You should, of course, not lie about the medium / tool you used to produce it. Unless it would be funny)

I am not willing to accept that every time someone presses the button, the thing that comes out is definitionally art.

But I'm not singling out AI with this!

I don't think that every time someone takes a photograph the product is art! Sometimes people are just taking a photograph to have a reference for later, sometimes it's for verification, sometimes it's to show someone else in an exchange of information, or in the hopes that the camera captures something that their eyes can't, or because they were clumsy and took a picture by accident.

And even then, I accept that sometimes something which was captured by camera on accident, or without intent, can still evoke in me something, and if succeeds in doing that, then I'm still willing to call it art, even if the person who "produced" it doesn't think so, or never knows what they captured.

The thing with the AIs are ... they're just processes, right? And just like we can take human paintings to extremes and ask whether a single brushstroke can be art, or even no brush strokes at all, we can scale back the machines, train them on less data, use fewer nodes. And I don't think it's fruitful to go in a direction where we say that at some certain level then yes, it's art. I think it makes sense to go off whether it makes people feel something, if it does the evocation thing, if it engages.

I think there's a distinction that I personally would draw between "bad art" and "not art", and I think that the vast majority of AI image output is "not art" rather than "bad art", but that's because the vast majority of the output is just people clicking buttons and the machine just spitting things out, and the result is meaningless.

Wait, just press a single button? Since when does machine image generation work like that? I thought you have to input a prompt?

(This is not a rhetorical question. These are meaningfully different situations)

This is a question of definitions.

Can you get an image back with no prompt? Yes, though your interface might not let you, this is not technically difficult to do at all, just something that companies might prohibit. But is a null prompt in fact a prompt? Depends on who you ask.

This is Stable Diffusion from two years ago, and not my own example, but here's one possible result, which depends on the random noise:

I would say that this is squarely not art, and would only change my answer if a sufficient number of people disagreed with me and found some emotional connection, meaning, etc. in it.

Modern Stable Diffusion would give back a better result, in that it would look more like something that existed.

However, I don't necessarily think that "no prompt" is substantially different from having a prompt. The act of choosing a prompt is not what makes an AI image art, in my opinion. In fact, it's trivial to automate prompt generation, and I've done that in the past. You can get GPT to do it.

I think any definition of art where a programmatically random prompt of "cat" is not considered art, but my son's prompt of "cat" is considered art, is probably a bad definition that won't work for actually communicating with people.

Honestly, I think the picture created with no prompt is amazing and humanity's cultural heritage would benefit if this button were pressed as many times as possible and the results displayed. The averaged out picture of human culture as observed by an image generator is an AMAZING concept. I want to see more of this. I love how wonky and confusing it is. I love that it clearly has no idea what it's doing. I love that you can still recognize the elements the image is compiled from. I want to see the null prompt result evolve over time. (This is not me trying to make a point. This was my sincere first reaction. I know many people in person who would have the same one. This is fun as hell)

And yes, clicking that button on purpose to see what the machine gives you IS an act of art. Perfomance art, perhaps. Something like humming a song to yourself, maybe.

Like, anyone who's clicking the button with the null prompt is obviously not doing it for a utilitarian purpose (which was your argument for not all photography being art). They're obviously playing with the thing. Doesn't that make it art enough already? They might click the button multiple times to pick the picture that best expresses what they wanted out of the exercise, or they might click the button just once and just take the first thing the algorithm gets them.

So long as it's not an accidental click where they won't ever actually know what the machine made, it's art.

(And if it were that last thing but somenoe later found the abandoned unseen picture and published or displayed it, that'd be art again)

i mean, its crazy right? it is completely and totally just a matter of bias and preconceived notions and reactionarism pushing people to arrive at the idea that a person using a generative model soley based on their own work intentionally creating a specific piece of art they have already visualized in their mind is "not art" because... well... um...

well they can't even say! they say "it's not intentional" (it is, he already knows what he wants it to look like!), "it's not work/effort" (god i want you people to see how long it takes me to generate the simplest of anime girls in midjourney), or most bafflingly "the prompt is art but the image output isnt". and nobody can actually explain WHY because there ISNT a good reason! they just have this kneejerk hatred and disgust of this new technology and art form and any attempts at prodding at that feeling are met with aggression or blank stares.

I'M IN CRAZY TOWN

"I would argue the prompt is itself art, but not the output image."

i'll consider agreeing that AI images are art on the condition that we actually start treating them like art. if i ask van gogh "hey what was your thinking behind giving that woman 7 fingers" and he goes "idk the program just does that sometimes" then it's bad art and he's a hack artist, at least in that medium, no different from someone going "idk the paint is just like that sometimes" when asked about the shoddy brushwork and texture of a painting theyve made. dont know OP but a lot of AI image makers seem to want the prestige of Art(tm) when its convenient and hide behind "no its just a funni computer picture" when it's not. can't have it both ways, if this is your medium then take responsibility for it

its very strange you think i do not want ai art to be taken seriously as a medium. why are you speaking so condescendingly to me about this? "prestige of art" girl what? i just want people to recognize one of the mediums i work in is legitimate lol. like "take responsibility" lol? its not that deep. id just like people to concede that just bc you dont Like something doesnt mean its Not Art.

anyways heres a picture i made cus u obviously have a very limited idea of what ai art might be :)

Okay, I mean ... I'm a writer, and have fed LLMs my own writing, and have pushed a button to get written material out of them. Due to the way that these models work, you can't train on only a single artist/writer, because it's just not enough data to get anything coherent, even if they're prolific, but that aside ...

I mean, fine, it's a definitional debate, those are always pretty dumb.

But to me, art is about expression and emotional power much more than it is the mode of production or even really intent. By this metric, there are lots of things that other people would consider to be "art" but I wouldn't: there are plenty of paintings that take technical skill but express nothing and hold no emotion. I think given the nature of painting, it's relatively unusual for a painter to do a full painting without trying to get something across in some way, but I do think it happens.

And the thing is, I don't think that this disqualifies all AI art! I think some of it has expression and emotional power even though it's just a prompt and a model! But I think if we call everything art, there's no point in having the word, it becomes meaningless, and the category of things that I mean when I say art should discount a lot of stuff that has no expression or emotion to it.

So hypothetically, if I had a machine trained only on my stories, and had it write a new story, would it be art?

It would depend on whether it spoke to me or not, whether it had some expression to it, and whether there was a connection to be made. How does it make me feel?

Which means that sometimes when I push the button, I get art, and sometimes (much more often) I do not.

I do not expect this stance to take the world by storm, but I think it helps to explain the distinction in my mind between the AI slop and the stuff that actually makes an impression on me.

Have yall heard of "bad art".

Like. Art that sucks. Art that completely fails at the intended purpose. A joke that's not funny. A drawing that you can't figure out what it was supposed to be. Writing that actively lulls you to sleep. Singing that hurts your ears.

We don't need to invent a separate special category for "whatever they tried, it's not doing anything for me". That's normal for art!!! A child excitingly smearing paint over a piece of paper is very unlikely to get across anything coherent except that they really like drawing. A first-time AI artist is very unlikely to get across anything except that they wanted to see what would happen if a machine drew an image to their specification.

And, hear me out: bad commercial art is a category of art too. Ads are art; they're just ad that generally incites little but hatred, with extremely rare hits. The same is true of AI images that are nothing but an attempt to replace hiring a real human artist for simple illustrations. The person who made them (not the AI. The person who made the prompt) still had to engage with the topic to come up with even the simplest, most boring, basic and failed prompt, and they still make us, the audience, think about the subject matter. (Even if what we're thinking is "I wish I never heard of this thing ever again")

Consider also the concept of "fraud", which is also pre-existing. Remember the story of two little girls who took photographs with fairy cutouts and ended up believed that they met real fairies, and they just kept lying about it? Pretending that something is one kind of art when it's actually another - photoshop job pretending to be a real photo, a fictional story pretending to be something that actually happened. That's not new either. And it doesn't make the fraudulent output not art. It's art with intent to deceive, it's art we might well disapprove of ever being made.

Still art.

The current image generating "AI" is not categorically different from, like, photoshop, with all its filters. And pressing "generate clouds" in photoshop with the default black and white colors, then saving the output, is technically "art you made" - if you want to make a custom screensaver or whatever, this is a perfectly valid way to MAKE one, one that will be your own work that you don't need to credit to anyone else. (You should, of course, not lie about the medium / tool you used to produce it. Unless it would be funny)

I am not willing to accept that every time someone presses the button, the thing that comes out is definitionally art.

But I'm not singling out AI with this!

I don't think that every time someone takes a photograph the product is art! Sometimes people are just taking a photograph to have a reference for later, sometimes it's for verification, sometimes it's to show someone else in an exchange of information, or in the hopes that the camera captures something that their eyes can't, or because they were clumsy and took a picture by accident.

And even then, I accept that sometimes something which was captured by camera on accident, or without intent, can still evoke in me something, and if succeeds in doing that, then I'm still willing to call it art, even if the person who "produced" it doesn't think so, or never knows what they captured.

The thing with the AIs are ... they're just processes, right? And just like we can take human paintings to extremes and ask whether a single brushstroke can be art, or even no brush strokes at all, we can scale back the machines, train them on less data, use fewer nodes. And I don't think it's fruitful to go in a direction where we say that at some certain level then yes, it's art. I think it makes sense to go off whether it makes people feel something, if it does the evocation thing, if it engages.

I think there's a distinction that I personally would draw between "bad art" and "not art", and I think that the vast majority of AI image output is "not art" rather than "bad art", but that's because the vast majority of the output is just people clicking buttons and the machine just spitting things out, and the result is meaningless.

Wait, just press a single button? Since when does machine image generation work like that? I thought you have to input a prompt?

(This is not a rhetorical question. These are meaningfully different situations)

i mean, its crazy right? it is completely and totally just a matter of bias and preconceived notions and reactionarism pushing people to arrive at the idea that a person using a generative model soley based on their own work intentionally creating a specific piece of art they have already visualized in their mind is "not art" because... well... um...

well they can't even say! they say "it's not intentional" (it is, he already knows what he wants it to look like!), "it's not work/effort" (god i want you people to see how long it takes me to generate the simplest of anime girls in midjourney), or most bafflingly "the prompt is art but the image output isnt". and nobody can actually explain WHY because there ISNT a good reason! they just have this kneejerk hatred and disgust of this new technology and art form and any attempts at prodding at that feeling are met with aggression or blank stares.

I'M IN CRAZY TOWN

"I would argue the prompt is itself art, but not the output image."

i'll consider agreeing that AI images are art on the condition that we actually start treating them like art. if i ask van gogh "hey what was your thinking behind giving that woman 7 fingers" and he goes "idk the program just does that sometimes" then it's bad art and he's a hack artist, at least in that medium, no different from someone going "idk the paint is just like that sometimes" when asked about the shoddy brushwork and texture of a painting theyve made. dont know OP but a lot of AI image makers seem to want the prestige of Art(tm) when its convenient and hide behind "no its just a funni computer picture" when it's not. can't have it both ways, if this is your medium then take responsibility for it

its very strange you think i do not want ai art to be taken seriously as a medium. why are you speaking so condescendingly to me about this? "prestige of art" girl what? i just want people to recognize one of the mediums i work in is legitimate lol. like "take responsibility" lol? its not that deep. id just like people to concede that just bc you dont Like something doesnt mean its Not Art.

anyways heres a picture i made cus u obviously have a very limited idea of what ai art might be :)

Okay, I mean ... I'm a writer, and have fed LLMs my own writing, and have pushed a button to get written material out of them. Due to the way that these models work, you can't train on only a single artist/writer, because it's just not enough data to get anything coherent, even if they're prolific, but that aside ...

I mean, fine, it's a definitional debate, those are always pretty dumb.

But to me, art is about expression and emotional power much more than it is the mode of production or even really intent. By this metric, there are lots of things that other people would consider to be "art" but I wouldn't: there are plenty of paintings that take technical skill but express nothing and hold no emotion. I think given the nature of painting, it's relatively unusual for a painter to do a full painting without trying to get something across in some way, but I do think it happens.

And the thing is, I don't think that this disqualifies all AI art! I think some of it has expression and emotional power even though it's just a prompt and a model! But I think if we call everything art, there's no point in having the word, it becomes meaningless, and the category of things that I mean when I say art should discount a lot of stuff that has no expression or emotion to it.

So hypothetically, if I had a machine trained only on my stories, and had it write a new story, would it be art?

It would depend on whether it spoke to me or not, whether it had some expression to it, and whether there was a connection to be made. How does it make me feel?

Which means that sometimes when I push the button, I get art, and sometimes (much more often) I do not.

I do not expect this stance to take the world by storm, but I think it helps to explain the distinction in my mind between the AI slop and the stuff that actually makes an impression on me.

Have yall heard of "bad art".

Like. Art that sucks. Art that completely fails at the intended purpose. A joke that's not funny. A drawing that you can't figure out what it was supposed to be. Writing that actively lulls you to sleep. Singing that hurts your ears.

We don't need to invent a separate special category for "whatever they tried, it's not doing anything for me". That's normal for art!!! A child excitingly smearing paint over a piece of paper is very unlikely to get across anything coherent except that they really like drawing. A first-time AI artist is very unlikely to get across anything except that they wanted to see what would happen if a machine drew an image to their specification.

And, hear me out: bad commercial art is a category of art too. Ads are art; they're just ad that generally incites little but hatred, with extremely rare hits. The same is true of AI images that are nothing but an attempt to replace hiring a real human artist for simple illustrations. The person who made them (not the AI. The person who made the prompt) still had to engage with the topic to come up with even the simplest, most boring, basic and failed prompt, and they still make us, the audience, think about the subject matter. (Even if what we're thinking is "I wish I never heard of this thing ever again")

Consider also the concept of "fraud", which is also pre-existing. Remember the story of two little girls who took photographs with fairy cutouts and ended up believed that they met real fairies, and they just kept lying about it? Pretending that something is one kind of art when it's actually another - photoshop job pretending to be a real photo, a fictional story pretending to be something that actually happened. That's not new either. And it doesn't make the fraudulent output not art. It's art with intent to deceive, it's art we might well disapprove of ever being made.

Still art.

The current image generating "AI" is not categorically different from, like, photoshop, with all its filters. And pressing "generate clouds" in photoshop with the default black and white colors, then saving the output, is technically "art you made" - if you want to make a custom screensaver or whatever, this is a perfectly valid way to MAKE one, one that will be your own work that you don't need to credit to anyone else. (You should, of course, not lie about the medium / tool you used to produce it. Unless it would be funny)

The worst thing about creative AI right now is that it produces bad results. The writing is bad, the images are bad, and the video is bad. It's impressive, sometimes, that the technology works as well as it does, but it's still bad.

I think if you sit down and go through a few hundred generations, then tweak and edit and inpaint and think intently, you can sometimes get something worth putting in front of people, if you have the right eye for it. I could definitely edit up an AI-written short story into something worth reading, especially if I was the one who had fed it the prompt and gone through the work of having my own ideas to insert. I think at least part of the output would be the AI's, and I could carve away everything that was nonsense or just bad, leaving only a few turns of phrase or some general boilerplate structure ... and this would take more time and effort than just writing the thing myself.

Most people who use generative AI do not want to do any work, and in fact, have no conception of what work would be required. Most of them are consumers, not producers, and they're used to the modes of content consumption, where you don't look closely at the details. Generative AI, in its current state, just kind of sucks when you're in a "press button, get results" mindset.

The stuff generated by "press button, get results" is the vast, vast majority of AI art that you will see, even accounting for filtering effects. There are a lot of people who have no love of artistry producing artwork via machines that are not good at making artwork, sometimes just for a lark, sometimes with profit in mind, and it's threatening to drown out other stuff in spite of being bad.

This is my thesis: generative AI produces bad results, and this is possibly the worst thing about it. If it were able to produce good results, I think that a lot of people would be less opposed to it. If you could get a short story that was worth reading, or a picture worth looking at, for no additional effort of manipulation or prompt engineering or whatever else, then we would be flooded with good art instead of bad art.

When it comes to art, I care about how it makes me feel, and what it's trying to say, and where the intent is, and what ideas it has. AI is not there. Possibly it will never get there. But sometimes I see a picture that the AI has made, and I do feel something in the sweep of the lines, or the composition, or just the juxtaposition of elements. It's just really really rare, and the product of either chance or really careful work on the part of some human. It's not something that the AI can do reliably, at least at the moment. You can also quibble about intent, because the AI "has none", but I find beauty in nature too, which is not trying to make a statement with its sunsets, and whose intents, if they can be said to exist, are mostly about things that are orthogonal to my perceptions, like the plumage of a sparrow or the curved leaves of a fern. To me, art is art because of the way that it can be read and the emotions that I feel when I look at it. Contentious, I'm sure, but I don't find other definitions all that useful.

But the art that the AI makes is, unless expertly guided, bad. And there's a ton of it, and it's impacting the ability of real artists to make superior work.

I think the future I see, if the AI doesn't get better, is one where we have a bunch of cheap shit that's replaced a lot of good expensive things. I am in favor of cheap things, but I'm not in favor of shit. I would love for translation to be as simple as pressing a button. I would love to have a good painting to go with every chapter I write. But we're in a world where the results mostly suck unless you're willing to put in quite a bit of effort and have some expertise in a field of creative endeavor, and that means we're in a world where the products are bad.

I'm interested to see how the conversation shifts if the results start getting better, because that seems to me like one of the sticking points.

The best use case for AI art I've seen, one I'm actually willing to defend, is illustrations for web fiction.

It's generally in the anime style, which means it's pretty simplistic but there's also not much room to be bad. It doesn't say much, but it's not supposed to - the text is doing the saying, the AI art is just providing the picture of a key moment that the author wanted to add a visual to. These people are willing to put in work into, if not prompt engineering per se, then at least trying things again and again until they get something that works for them. And it adds real value without being a complete thought-provoking piece - just being able to visualize the characters correctly is a lot!

Recently When Immortal Ascension Fails Time Travel To Try Again released a chapter with a climactic song performing scene, where multiple characters' efforts came together to create a miracle from the power of rock. And look at this:

I strongly doubt the AI wholesale made this. I am pretty sure photoshop skills of the story author were involved. But the author doesn't draw, and without AI they would not have the pictures to compile into this.

So my answer is, "AI art" by itself is pretty bad, yeah. But it can make a perfectly acceptable ingredient for a mixed-medium work. It can make a starting point for a person with passion and a vision to carve something out of.

(But can't they just commission someone for it?) Every chapter, multiple pieces? Nobody is made of this much money. The story is enhanced by being mixed media, but hiring people for it is a bit out of budget.

(But can't they just learn to draw?) They're a bit busy learning to write (as in, being already a pretty great author, but everybody always keeps learning on the go). Learning to draw to a level where you can consistently depict what you want in the mood you want and other people can get the thing you intended out of the result is a multi-year process and takes way longer than figuring out how to get image generation software to give you something that works. And the story is better for it today, already.

(But isn't this theft?) They're not a multimillion dollar corporation figuring out how to avoid paying their employees. The options are not "the art gets made either way, but the artist gets paid or they don't". The options are "the artist has to find other work either way, the art exists or doesn't". We should have UBI so creatives aren't so desperate for every chunk of income they might have had; in the meantime, while I would not necessarily say this use case justifies all the downsides "AI" brings with it, it's certainly a positive.

When Immortal Ascension Fails is my favorite for how it uses this, but plenty of other works use AI for covers and illustrations. (And sometimes the authors of these same stories scrape together money and get a commission from a real life artist for an actually good rendering of their characters or scenes. And it is good. And the simpler, shittier, more approximate AI illustrations still have a role to play)

TL;DR:

Asking the voice actors on my video game project to consent to having AI models of their voices created specifically and exclusively for the purpose of making the dialogue say the player character's name out loud at appropriate junctures no matter what the player chooses.

the more I understand the role of AI as an assistive tool, the more I realize how much we need to be discussing it openly in those terms. how much we need to treat it like any other form of disability advocacy

and when you recognize AI as an assistive program, so much of this looks very different

#interesting point but i think it misses some important aspects#if someone is using prompts to generate artwork thats less making art and more commissioning an AI to make art for them#and if someone orders a commission it an asshole move to turn around and say 'i made this'

if you use a photoshop filter to make an image of clouds, is it an asshole move to turn around and say "I made this" or did you commission photoshop to make it?

The current image generation technology is a tool, it's not a person. Using it is not like commissioning art at all, because there isn't a person on the other end who puts their creative intent into the process. The only creative intent is yours, and the "AI" just blindly throws stuff together to offer you in response to a prompt.

the more I understand the role of AI as an assistive tool, the more I realize how much we need to be discussing it openly in those terms. how much we need to treat it like any other form of disability advocacy

and when you recognize AI as an assistive program, so much of this looks very different

Look if you have never written a robots.txt I think you might not have a reasonable understanding of how AI training data is collected and what controls it is possible to enact. Everything on your Tumblr is and has been fair game for the entire internet to see forever unless you password protected it.

I get why everyone is yelling about this, because loudly complaining about AI is what's cool, but your data has probably been scraped by OpenAI years ago.

You should always assume that a bot is ignoring all the polite signs you've written and scraping every public page of your website forever. This is always true. If you don't want someone to see something you've made the only sure option is to never put it publicly online or ever give it to someone who might put it publicly online. Everything else involves the risk reward trade-off of making your eight mutuals laugh vs. getting reposted to ifunny.

You guys remember NFTs, right? Remember how we would make fun of them because they would claim unique ownership of an image, but we could access the image anyway, right? . Remember "Right Click + Save As?" Something being Readable on the internet means you can make a copy of it, since the computer has access to the whole thing so that it can send it to the GPU to the monitor and then to your eyeballs - this is just how computers work.

AI works by 'Right Click + Save As'-ing a lot of times very quickly. It's currently not actually illegal to Right Click+Save As, whether you do it a few times slowly or a lot of times quickly.

Your posts are in an AI model

and then Tumblr decided to sell them to AI models.

Now, don't get me wrong, tumblr selling out the users to AI companies is bad, yes, they shouldn't do that. It sucks.

but don't lets get this confused: your posts were already in there. Tumblr selling them is about tumblr making some money and about the AI models having more exhaustive post collections. It's not about your posts being in an AI model, vs not being in one. That battle has already been lost.

Can you find your post on google? Then it's almost certainly in an AI model already. Think about it: These AI sites showed up before all the sites were making deals to sell their users' content, right? How do you think they built them in the first place?

They scraped the posts. Just like google and bing and such do when they build their search indexes.

It's a fundamental part of how the open web works: you want your posts on tumblr to be visible to users, right? You want them to be readable?* Like, look how much stuff broke when twitter changed their whole read-while-not-logged-in policy, ruining a bunch of thread links/NSFW links. And if it's visible, it's scrapable. That's what the AI models were built on.

I've done website scraping before (not for AI models, of course. I was doing search engines and website archival), this is just how it works. You hire a few relatively smart CS graduates and tell them "build me a scraper that'll give us a bunch of tumblr posts" and they go off for a month or two and come back with a database of a few billion posts, and you stuff that into your AI model. That's how they got all the deviantart and flickr and twitter and pinterest and so on posts. They didn't pay for them: they just took them.

They only ever pay for this shit because either:

ANYWAY: my point is, don't treat this "oh no tumblr is selling our posts to AI" like it's a big thing that might happen and it would be bad to happen. Yes, it's bad, tumblr shouldn't do this, this'll let AI models get continual updates of content for far easier than just scraping them would be, tumblr betrayed user trust, and so on...

but realistically, this is not a black and white matter of "if only tumblr didn't do this, then we'd be safe from AI models!"

Nope. We already lost that battle. I'm sorry, and it does suck, but that's just how it is. The avalanche has already started, it's too late for the pebbles to vote. * I'm assuming here that you don't run a private blog that's set to only followers or something. You'd be safer then, of course, but you're not really my target audience for this rant

I guess my point is: Don't freak out about this being a big thing that you're gonna have to fight to stop. Having anxiety over this is not useful. Instead, you need to realize that this battle is already lost, and reconsider your strategy from there.

I'm not saying "you need to give up and let AI companies take everything you make and do whatever they want with it". I'm saying that the specific battle of stopping AI models from getting your posts is a lost one.

So rethink your strategy: What else can be done? Can we limit what the companies do with our data? Can we stop them completely? Can/should we change where we post artwork to keep it from getting acquired by AI models in future? Does this need a political solution (like "ban generative AI"), not an individual solution from artists?

All those are options you could look at and decide on, plus I'm sure there's tons more I haven't thought of. I'm not trying to tell you want to do, just saying that you should choose your battles, and when choosing battles, don't choose one that is already lost.

Some things you can do:

Thoughts on just, like, posting a lot of weird sludge

I just think it would be fun is all

Economic anxiety has a way of bringing out reactionary sentiment in anyone if they're not careful.

It is deeply, deeply frustrating to watch it play out in front of me in leftist spaces such that self-proclaimed leftists are using actual, literal fascist arguments about Real Art vs. Fake Art and Real Labor vs. Lazy Button-Pushing.

These things don't become any less bad when you SAY your enemy is "some rich techbro" while calling broke disabled hobbyists "evil soulless automatons".

The central logic doesn't become true when you SAY you're targeting an inhuman machine while you screech obscenities about a great replacement at its operator.

When you say one minute "there is no unskilled labor, only undervalued skills", it doesn't magically absolve you of saying "nooo, you were supposed to automate away the BAD and DEMEANING jobs with no financial safety net for the workers, not THIS one I consider RESPECTABLE" in the next breath; it only makes you a fucking hypocrite.

"Fair use for me but not for thee" is not a rational position to prevent plagiarism and forgery; it's just a means to codify an ingroup and an outgroup.

"Degenerate art" is always, ALWAYS reactionary and proto-fascist thing to believe in, even if you wrap it up in other fancy words because you know "degenerate" is a Bad Word. "There is Good Art that makes society better and Bad Art, if you can even CALL it Art at all, that will rot our brains and turn us all into mindless drones if it's allowed to survive" cannot be made into anything but a reactionary position! Period! End of!

"Lazy button-pushers" are EXACTLY what corporations want you to think ANY automation operator is, so they can take credit away from those employees and criminally underpay them. They said the same damned thing about digital artists back in the early days of Photoshop. They say the same thing about overworked VFX artists today. You are DIRECTLY helping them make it worse with this argument.

The same old fucking trick of making you uncertain of your financial future so you lash out at other victims of the system because you "can't take the risk" of coming together to fight the actual enemy? Is working a FUCKING treat on way too many people who pride themselves on Not Being Like That - and it's even worse because a lot of the time pointing this out will get nothing but denial because maintaining pride in a leftist, progressive, pro-labor, pro-human Identity is more important to way too many people than ACTUALLY identifying the root of reactionary sentiment and the strategies used to spread it.

It makes me genuinely feel like I've fallen into a Fox News convention, hearing all these blatantly reactionary arguments and actively self-defeating strategies to Protect Labor.

OP please let me know if this is not the place, but I wanted to add a little bit of an alternative for folks to consider while challenging these beliefs within themselves.

AI art has done real harm to real artists, and I don't think it's any more productive to deny that than it is to deny the harm done by the arguments the op laid out.

But ask yourself: is that harm happening because of AI, or is it happening because of the ways in which people create and utilize AI? Are there ways to create AI art that do not or would not cause harm? (What about training an AI on data sets based around consent and fair use? What about AI models that aren't costing or making anyone any money?)

What is the nature of this harm? Oftentimes it's a monetary issue we're talking about; people make AI art that looks like an artists' in order to avoid paying for a commission. Would this harm exist if money didn't have an impact on someone's livelihood? Again: is it the AI, or the context? Could this happen without AI? Does it? (Hint: people have been getting angry about "copying styles" for a while now.)

We can talk about "stealing art" and "stealing credit", and I think there's a discussion there worth having, but is that conversation limited to AI art? ("Original OC do not steal" has been a joke for like a decade.)

In my opinion, at least for the moment, the majority of the harm AI art causes is a direct result of living under capitalism. There's an element here of consent/fair use, but again, I think a lot (but not all) of the conversation there also comes back to who is profiting off of who's work.

That doesn't mean AI art as it is right now is fine & cool & so we don't need to worry about it, but it does mean that it's probably a lot more worthwhile to direct our focus towards regulating AI art, than it is to bicker over what "counts" as "real art" like we're trying to recreate the fucking Hierarchy of the Genres.

IDK, I think saying "actually, this wasn't created by a human being" is a pretty good place to draw the line on what art is for the purposes of law and copyright. That's actually established law since that whole "non-human primate took a photo" lawsuit that PETA bankrolled.

I don't actually care about the philosophical definition of art when it comes to procedurally- generated images. (There is no such thing as "AI" right now, and I refuse to cede that ground to the PR game of billionaires.) Saying "this isn't art by the accepted legal standards because a human being didn't actually create the art" isn't the same thing as saying "this particular style or subject is degenerate," and that's a totally shit line to draw in the sand.

Like, it's just not a defensible position, and it is, at best, an attempt to rile people who don't want to have their livelihoods automated away. I'm not playing along with this speedrunning of Godwin's Law.

Procedurally-generated art is the pet project of billionaires who have openly admitted that they can't do it without violating the copyright and IP of basically every artist who's ever had a picture of their work posted online. It's bankrolled by Sam Altman and Elon Musk. It is not morally neutral to support something that is definitively intended to remove the need for billionaires to pay artists. It's just not.

If all of the theft-based databases were purged and procedurally generated art was based only on a database of public domain art and opt-in CC art, that would be wonderful. But it is not any more morally neutral to say "I want to make art and I feel I'm owed it so I'm going to steal from you" than it is to look at an indie author and say "but i wanted to read your book so I pirated a copy. I know this means less money for you, but like... I wanted it and I didn't have the money for it."

(And if you think people haven't said this directly to artists and authors on this site, I invite you to pay attention. They have and they do.)

As it stands now, PGI/Procedurally-generated images are based on theft, and no amount of "you're a bad leftist" finger-wagging will change that.

Stealing money out of your fellow worker's pockets today under the idea that this will all be fine Someday After The Revolution is bad praxis, no matter how many "you're a bad leftist if you disagree" bows you dress it up in.

And if that's how it is, fine. I'd rather be thought a bad leftist by someone who thinks they are totally cool to steal from other workers at the behest of billionaires - every creation of PGI moves the needle a little further in the direction Musk funded it to go, so. 🤷🏻‍♂️

I reject the argument entirely. It has meaningless foundations.

Uh-huh, so spending hours wrangling a computer INTO getting the result you want isn't labor, or human intent, or human creation.

Is fractal art "not art" then? CGI? Defining this labor as "not real labor" is, in fact, EXACTLY what the billionaires want. Which one of us is ceding ground to them?

Defining procedurally generated images as "theft" because they have looked at images that already exist, when other forms of art do the same, IS, in fact, "fair use for me but not for thee." You can reverse image search nearly any generated image and NEVER find a result similar enough to have grounds for a copyright infringement claim. Copyright infringement is about the end product, NOT about the process. Even if you could claim that edits are INHERENTLY copyright infringement no matter HOW different they are from the source material - which you can't - then fine, go ahead, point out which pixels were "stolen" from your work, if it DOES work the way that PUBLISHING INDUSTRY GIANTS - you know, people with more money than god who are at risk of having maybe a little bit less of that money - say it does, source-dude-trust-me, while the people actively WORKING on the tech are happy to document how no it fucking doesn't work that way.

YOU are, indeed, being a giant fucking reactionary.

You are PROVING my point.

YOU are the one ceding ground to billionaires, because this is a new thing, and you have just eaten up a whole bunch of absolute LIES because they're shouted at a loud enough volume, and you are, in your own words, rejecting the fundamental argument of anyone pointing out how that's dangerous.

Good fucking job! You sure did show those billionaires what's what!

I guess there could be a case made for whether creating and keeping a dataset for training and further model improvements could be counted as copyright infringement, but sure it does not apply to actually using a model, and creating images from it. Not that it matters when "creating a model" and "using a model" are notoriously conflated in these discussions.

And yeah I can definitely assure that discussions on this site on "is ai art real art" are never in the context of law and copyright, but mostly reactionary nonsense like

but this stuff is also something that you need to pay attention to. And yes, I see economic fears as justified, but also we can recognize nuance in that, for example, Google Translate is

here instead we get... this.

@sailor-kaiju if you're that interested in having this argument, fine, let's go.

You dismiss the comic I brought up and its subsequent discussion as meaningless insecurities, but they are nothing of the sort. The artist is absolutely correct - not just about his own art, but about how art works in general. Your art has just as much reason to be considered "stealing" from whatever you take inspiration from, as does my writing. Recognizing that fact is something all artists should do, and I see plenty who do so. Your lack of that self-awareness does not make you any better than those of us who have it.

Meanwhile, I'm surprised you see the idea of UBI as unrealistic Marxism. It's a lofty goal, but trying to put a stop to the use of an effective piece of technology is even less practical. I've been saying for years that it's the real answer to concerns around automation, as have many others - starting long before generative AI got as big as it is now.

If you're not convinced by those arguments, good news - I have plenty more of them! Multiple tags' worth, even. It's a common topic on this blog, one where I've said and shared all sorts of things I don't feel the need to reiterate. I'd expect people who've been following me for a while to be familiar with them by now, unless they're avoiding engaging with the topics entirely.

Whatever is the case, I would recommend against perceiving someone posting things on their own blog as them coming into your house and spitting in your face. Doesn't seem like a healthy perspective.

And I’d expect people who have followed my blog to understand my POV on art more since I talk about it quite regularly and yet here we are!

That artist may be correct for themselves. That’s on them. But honestly, I am trying not to be rude to that artist because they didn’t choose to be in this conversation, which is why I keep turning away from that argument…but that mentality is that of a really insecure artist. I have NO issue knowing how I am different from a machine making content and it baffles me how someone couldn’t.

-I- don’t just replicate and vomit it back up. And I do art that is freakin derivative as HELL and can STILL explain all sorts of symbolism and choices made to create it the way it was created. assuming that’s “all” artist do that is not only disrespectful to artist but to art as a whole. If some artist do nothing but replicate…that’s on them, but with respect to the artist of that comic, if all they are bringing to their art is copying maybe they SHOULD be navel gazing about what they bring that an AI doesn’t. Mosy artist stop being replicate machines once they learn the technical skills and start using then to SAY something.

Does your art not SAY something? Are you not DOING something with all those references and skills? Why do you think all art is summed up only by the technical skills used to create it?

UBI is great, AI is effecting my rent now. I’m not living in a perfected future, I’m living here and now. How does UBI fix the issue happening currently? would YOU like to call my mortgage company and explain I won’t be paying my bills because UBI is the most ethical system? i’ll DM you their number, I’m sure they’ll take philosophical possibilities for my rent payment.

I have seen your arguments. And sorry….they do nothing for me. To the point I usually don’t engage cause you’re so clearly in the hole but I didn’t realize you were actually AGREEING with OPs terrible take in that post. The arguments are inherently flawed, often about what art even is, as a concept. Much like that comic.

Also, gimme a break about “taking it personal” when you’re talking about my profession AND passion. This isn’t theoretical, you are talking about ME when you talk about artist and AI art. i’m sorry your uncomfortable that you were talking down about a skill and profession and didn’t expect someone from that profession to react. But that’s what happens when you put out negative opinions publicly. The weird “Oh you’re getting too personal~” tactic isn’t going to work cause yeah ofc this is personal. I am very passionate about art and paying my bills.

Everyone uses or is gonna use AI. It’s inevitable. It sucks. You won’t have much of a choice anyway. Much like fast fashion. But just admit you want the pretty pictures and don’t want to pay to commission them and/or work to learn how to make them yourself.

I don't think I've ever followed your blog. I remember interacting positively with you in the past, but I don't follow many people.

I didn't say anything about artists being replication machines who don't have distinct symbolism and choices they can explain. My point is that beneath those choices, there's so much more you're taking for granted. Can you name every piece of art that might have had the tiniest subconscious influence on your choices of symbolism, or on developing your understanding of art to think to use symbolism the way you do in the first place?

When I write a scene, my idea of what that scene might consist of is shaped by any number of my previous experiences with media, to teach me about elements a scene can have and how they can impact it. When I write a sentence - even one outside of a story, like this one - I come up with a concept for what the sentence is meant to accomplish, and rely on subconscious parts of my brain to fill in gaps as needed, extrapolating from its existing information to come up with words likely to meet the criteria. That is exactly the same as the process of prompting an AI.

In order: of course it does, of course I am, and I don't think any such thing. I'm not sure what prompted any of these questions, but they seem to be a response to a thing I didn't say.

My condolences.

If this topic stresses you out that much, I would not recommend reading my posts on the matter. Have you considered blacklisting relevant tags?

I want to use AI to more conveniently complete projects that I probably could not feasibly bring into reality any other way. I have no problem acknowledging this. But I was advocating for AI art well before I became interested in actually using it myself, not for that reason at all.

so my position on "AI" art is - video game opponents have a better claim of being an "AI" than the image generating machine, much like the text generating machine. at least they are trying to imitate what a real AI would do, internally. at least they are reacting to circumstances based on a model of the (simplified digital) world; - the "art" part is giving the machine the prompt, not what the machine does. any more than photoshop is the one doing art when you add a gradient or a filter to your picture. just because the process is obscured and randomized doesnt make it any less mechanically programmed; - like, art is communication. the image generator isn't communicating anything to you, it doens't have an internality to communicate. the prompter does. and they're not communicating with the AI, they're communicating with the viewer. like using text to speech or something like that. a technological intermediary; - art is NOT in fact universally accessible. "i am disabled and I do art and teach art" is not an argument any more than "well I am disabled and I don't need a wheelchair" is an argument against wheelchairs. there are different disabilities??? i mean like everyone can do SOME form of art but they arent fucking. mutually substitutable. not everyone can draw a picture and i dont mean in a skill way. someone could have been a fantastic artist and then lost the ability bc their hands got fucked up; - speaking of mutual substitution, machine generated art is not the same medium as traditional and digital art any more than those are the same medium as photography. the means of obtaining the result are categorically different, come with categorically different challenges and skill requirements. you wont judge a photo on how photorealistic it is and you wont judge a painting on how difficult it must have been to stage. machine generated art is a third thing and must be clearly labeled in any situation where the difference between a photo and a painting too would be relevant; - aaand the means by which it is obtained are sketchy as hell. in a perfect future utopian UBI reality it maybe wouldnt be a problem but in the current now reality it hella is. and regulation based on existing flawed legal frameworks is a better solution than "waiting for the underlying issue to be solved" much like "homeless shelters" are not a better solution to the problem of homelessness than "housing for everyone" programs but are hell of a lot better than nothing; - also I think it would still be an issue in the perfect future utopian UBI reality because its a deepfake machine. its just a fact of life for photos but people not wanting the machine to read their art style and learn to replicate it are valid regardless of monetary concerns. what the machine does in remixing images is not in fact the same as a human looking at the picture and reverse engineering the technique by which it was done / having thoughts and ideas about how to do a similar thing. it is ethically and practically different

so there you go, youre both wrong and im right. praise me :P

AI has problems re: algorithm training and image ownership, but this image and this image only can stay.

Hot take about hot takes (a hot metatake, if you will)

If your "AI art isn't real art" take would also rule out

-photography -collage -surrealism -dadaism -readymades -generative art

you are not making some Inspired Point about Art™, you are simply being reactionary.

Tags from a happy customer:

You are describing collage.

You are describing L.H.O.O.Q., a piece of art created by Marcel Duchamp.

It’d be so much easier if people were able to detach the question of whether something is Art or not from a value judgement.

Something being art doesn’t mean it’s good! It doesn’t mean you have to like it! It doesn’t mean it’s impressive!

You can just say that it’s very, very unimpressive art that you don’t think warrants clout or awards or admiration. It’s fine.

I've seen bad art, I've seen uninspiring art, I've seen art that I genuinely believe is made for bad reasons with bad intentions (much rarer than people tend to say).

Still art, though.

However, photography passed off as painting (to make the realism look impressive while the composition etc are boring), a painting passed off as a photograph (to make something look like remarkable practical effects / physical ability when in fact it's just drawn), a collage passed off as fully your own creation when other people's work defines key parts of it, are all fraudulent.

The problem with AI art isn't when it exists, it's when it tries to pass itself off as something else.

(Also there's a problem with the economic system we live in. Institute UBI and people will suddenly be a lot less resistant to technological advances making human labor redundant. That's not a problem with AI art per se, but it IS a problem that any and all proposed legislation concerning it has to contend with)