AI flooding on Deviant Art

[Mr. X]

I notice there is a lot more AI flooding on DA. All these $5 galleries with no thumbs that flood into the search engine. Seems some people are just casting a very wide net. I would think DA would say No to this but given one of their top earners was somebody with dozens and dozens of these $5 packs of AI pics they probably see this as money.

[Danorian]

I've blocked several artists that just flood my feeds with AI drabble.
I do AI art, but keep it in my Discord page, rather than polluting DA!

[brdiy]

I'd have to agree. Though I'd have to admit, a couple of artists have been successful in using AI to create the perfect likeness of Melissa or Helen Slater's Supergirl, for example, a greater majority are just generic looking and rather repetitive.

[Bladebur]

It's bad. Very few people edit or iterate the resulting images. It's all about prompting, and then producing 30 images that look the same. Nobody even bothers to inpaint the most glaring mistakes, people with a third arm are common (not to mention horror hands). There's a ton of noise, more than the platform can handle.

It was obvious that the monetization stuff would go very wrong with the AI stuff, but the platform is handling it poorly. My watch page is often 75% grey, which is a waste of everybody's time.

[xxxwarriorxxx]

Agree with you all. It’s awful. I did what Danorian did and blocked a few of them. It definitely cleans up the filter and makes what you want to look for easier.

[shevek]

Agreed on all counts. About 50% of my feed is these same bland AI superheroines with not even any creativity in the prompts. I mean, I'm not really seeing any sexy Hindu goddesses or Celtic warrior queens. You give boring prompts, you get boring results. I don't care if the results do sometimes look like Melissa or Helen, Lynda or Gal, because they're obviously not. The people who use AI to make a quick buck are boring people. I bet a lot of them are from Asia, and they're just very utilitarian about it: if it yields positive financial results, then do it.

One of the worst violators that I see are the transformation channels. People used to post great live-action transformation sequences on both SuperheroineTF (Deviant Art) and SuperheroineTFs (Reddit). I even got a series of three transformation customs from the legendary Brad328 thanks to my consistent postings on those sites, and I'm pretty sure the latest custom we're shooting this Tuesday (Custom 15) is also derived from that source. But now it's being overwhelmed by AI. And none of these AI things show any kind of sexy drawn-out transformation process, or special effects or interesting arousal, etc. Not the way you would see a clip from Giga or TBFE, or any number of sexy fetish producers on C4S, or a comic sequence from Manic. It's almost all just simple Supergirl shirt rips, where the Supergirl analogue has suddenly grown huge breasts and massive six-pack abs the moment she pops her shirt open, but you don't know how she got them or how she feels about having them because the art is utterly without soul.

Last year, one of our two interior artists for our comic wanted to start using AI backgrounds to save time on his art. A bunch of our comic creators had to literally stage an 'intervention' where we sat him down at a local pub and told him not to do it. He relented, but as a result it took more than a year to complete eight pages by hand. The results look like real art, however, and there's no fakery in it.

Ugh on AI. Nothing but a plague. I don't allow AI to take over my Microsoft Browser, or to guide my actions on Social Media, or anything. Just say no!

[sneakly]

The renders are pretty damn soulless. Even when the the images are an interesting context, the characters aren't really engaging their surroundings or other characters. I think they do have usage if you are looking for creating novel character or setting, like steampunk or sci-fi pirates. I made a few, but they get boring pretty quick. Just like chatGTP written fiction, it can write, but it can't write well.

Hopefully, DA will add a feature to filter out some of the crap. You can label your post as non-AI and you can toggle them to not be used in by AI engines as well. Unfortunately, like any technological leap forward, we are going to have to wait for someone to really figure out how to apply it well before it becomes worthwhile.

[Femina]

Yeah it's pretty sad. I think AI art can be fun as like... a thing you do for fun on your free time, but I dun much appreciate them flooding up the canals and making all the real art harder to find. It's not yet anywhere near at critical mass but it very well might at some point. The biggest issue is all the AI bros who think they're actually Great Artists because they typed some words in a specific order inside of a highly tailored algorithm and feel like it puts them on par with someone spending hours putting ink to paper.

I'd appreciate it if art sites remained places SOLELY for hand drawn art, but alas there's not many ways to enforce this. It's getting REAL bad on porn sites. Half of Rule 34 is just samey AI crap now.

[ThatOnePervert]

Places like Pixiv are at least bearable, because you can find multiple pictures behind a single search result. Having 30+ search results on DA fill your feed with half assed generic art is miserable. There are a couple of AI artist that I get, like yeah, you've got a niche fetish inside this existing niche fetish that no one was addressing. If this is the only way to scratch that itch fine. But at the very least keep them to yourselves or for God sakes write a caption or something original to go along with your slop.

[argento]

It's bad. Very few people edit or iterate the resulting images. It's all about prompting, and then producing 30 images that look the same. Nobody even bothers to inpaint the most glaring mistakes, people with a third arm are common (not to mention horror hands). There's a ton of noise, more than the platform can handle.

It was obvious that the monetization stuff would go very wrong with the AI stuff, but the platform is handling it poorly. My watch page is often 75% grey, which is a waste of everybody's time.

I agree with you, and the looks are also unseeing or seem to be looking I don't know where. It is a great tool to improve the image. Having the AI do everything is boring.

[Bladebur]

Places like Pixiv are at least bearable, because you can find multiple pictures behind a single search result. Having 30+ search results on DA fill your feed with half assed generic art is miserable. There are a couple of AI artist that I get, like yeah, you've got a niche fetish inside this existing niche fetish that no one was addressing. If this is the only way to scratch that itch fine. But at the very least keep them to yourselves or for God sakes write a caption or something original to go along with your slop.

One of the best things of the AI is that it can cover those cases (niche inside a niche inside a niche). But I'd still wish there were a way to filter out or give less weight to very similar images from the same artist. DA has become much harder to search.

[lionbadger]

meh

I kinda feel like this is what people were saying about poser 20 years ago

[spandex4fun]

I stopped using this site years ago cuz of their ridiculous censorship & fucked up site updates. Hear this makes me glad that I left it.

[Dogfish]

I've blocked several artists that just flood my feeds with AI drabble.
I do AI art, but keep it in my Discord page, rather than polluting DA!

It's a shame that so many people try AI art spoil it for everybody by not being able to resist the urge to spam it everywhere.

I mean, the first time I prompted an image and it looked cool, I liked it, but I still had the self awareness to know I hadn't done anything worth sharing.

[Mr. X]

DanO at SHC does actual work with AI cause it does take effort to get expressions and faces to look like someone. But the blast on 11 setting is getting out of hand.

I also don't agree AI is like poser. I agree poser is not hand drawing but it does take levels of skill to get pics that can push someone's buttons.

[Dogfish]

DanO at SHC does actual work with AI cause it does take effort to get expressions and faces to look like someone. But the blast on 11 setting is getting out of hand.

I also don't agree AI is like poser. I agree poser is not hand drawing but it does take levels of skill to get pics that can push someone's buttons.

Yeah AI art is a tool. You can use it well, and with skill, or not.

Has to be acknowledged too though that the ceiling for pure AI art, even by skilled creators, is low.

[Ernie]

Well...maybe there are people with no artistic skill who enjoy actually being able to generate nice images. And maybe some of those people also like to tell stories, and doing it with AI art combined with text is an exciting new avenue to explore. And maybe those people generate 20 images for every one they actually use in a story, and slave over fixing errors or covering them with dialogue bubbles. And maybe they put a lot of work into creating a 100 plus image story and then sell it on D.A. for $5, generating a little income to buy videos with. But I guess that would be evil or something.

[sneakly]

The thing about most art on line is that most of it is the dregs. Most people cannot write a coherent story or develop a character. Most DA stories are pretty unreadable. For every decent author posting fetish content on the DA there are a dozen or a hundred that are putting out pretty awful stuff. I would never tell any of them they suck because I know how hard it is to write. The same is true with visual arts.

This is a new technology that we don't know what the limits of it's potential will be. I have seen one person that had an absolutely amazing Yvonne Craig Batgirl model. He obviously spent a lot of time tweaking it very carefully. I came across a cross post on DA that I thought was a publicity still from the show that I had never seen. The background had the right feel for the show and she looked perfect. I asked the poster if it was AI and he said he had just ripped it from Instagram. Eventually someone is going to figure out what to do with this tech that is actually interesting.

In the mean time, I hope DA puts a cap on uploads. A two month old account shouldn't have 5,000 posts.

[jlocke]

AI seems to have just made another significant progression and some of it has filtered down to the free sites. Face swapping used to be either really bad or really expensive. It's now pretty good (not perfect) and inexpensive/free. I played with AI a year ago and made some cool images that I loved. Hundreds that I didn't, which I promptly deleted. When a bunch of gray pay boxes show up in my DA feed, I unwatch the artist. If there are too many low effort AI post, unwatch.

AI can be a really cool tool to help people with good imaginations and poor artistic skills. I've drawn my entire life, and I fucking suck, like embarrassing. It's awful. But... using an AI engine and carefully refining an image though words, I'm not awful. And I made some things I've wanted to see my entire life.

Is it so hard to capture Wonder Woman, use her magic lasso to obtain all her secrets, blackmail her, and force her to dance at a strip club. I mean, this seems pretty basic to me, but then again, I've never forked over a thousand dollars to make that custom. Too frugal I suppose. However, for zero dollars, I was able to make these:

ezgif-5-9ab4f01dce.jpg (230.64 KiB) Viewed 59754 times

ezgif-5-c7d52c0405.jpg (205.34 KiB) Viewed 59754 times

ezgif-6-f077f4fb3e.jpg (230.59 KiB) Viewed 59754 times

They aren't perfect, but they are way closer to what I've always wanted. And they are light years ahead of some of the AI monstrosities I made last year. AI is only going to improve. I predict within 5 years it will be possible to use AI prompts to create fairly realistic movies. We'll be able to refine scenes, insert dialog. It'll be very useful for creating entertaining, low cost content, and the appetite for content only ever increases.

AI is a tool, if it works, use it. If it freaks you out, probably best to leave it alone.

[hagarb]

And now we have video too.

[personnal]

What do you think of my ai pictures ?

[Danorian]

Some of my efforts in Bing.







Enough for my amateur needs.

[DonShip]

I don't hate AI pictures. They do give very realistic faces/hair, etc (we won't talk about their insanity with fingers). But I am frustrated with the FLOOD of AI images on DA. Regular artists take time, so can't flood us like AI can. The pay-for-view shit is another problem. I'm poor, so cannot afford to pay so many artists a monthly fee. So they are gone to me.

[sneakly]

And now we have video too.

That is surprisingly good. Now just string 60 of them together and you have a five minute short. I saw one yesterday where the girl had an extra joint below her knee.

They will get better. The hands have gotten much better in just a few months. Eventually you'll be able to use a an animated model to control the action and tweak it to perfection, script the dialog, sound effects and props.... and still 95% will still be utter rubbish.

Comparing game graphics from 10 15 years ago with today and what they looked like 30 years ago.... It is going to get better, but there will always be rubbish.

[hagarb]

That is surprisingly good. Now just string 60 of them together and you have a five minute short. I saw one yesterday where the girl had an extra joint below her knee.

They will get better. The hands have gotten much better in just a few months. Eventually you'll be able to use a an animated model to control the action and tweak it to perfection, script the dialog, sound effects and props.... and still 95% will still be utter rubbish.

Comparing game graphics from 10 15 years ago with today and what they looked like 30 years ago.... It is going to get better, but there will always be rubbish.

Thanks. Yeah, it’s a new tool and we’re early adopters, we have to figure out how to make this work to get the best results. Even today we find people calling digital art with layers and brush as "cheating".

[Femina]

Now there's AI flooding the thread about AI flooding deviantart.

xD It's spreading!

[argento]

DanO at SHC does actual work with AI cause it does take effort to get expressions and faces to look like someone. But the blast on 11 setting is getting out of hand.

I also don't agree AI is like poser. I agree poser is not hand drawing but it does take levels of skill to get pics that can push someone's buttons.

“Intelligence is the ability to adapt to change." Stephen Hawking

[superheroinelinks]

some AI is good and some garbage. It depends on which app or website you use

[Femina]

DanO at SHC does actual work with AI cause it does take effort to get expressions and faces to look like someone. But the blast on 11 setting is getting out of hand.

I also don't agree AI is like poser. I agree poser is not hand drawing but it does take levels of skill to get pics that can push someone's buttons.

“Intelligence is the ability to adapt to change." Stephen Hawking

This doesn't actually hold up in all spheres equally and Dr. Hawking would agree. He wasn't talking about individuals for example. Most adaptations are mutation... and most mutations lead to death. We see evolution as adaptation for survival because mutation over a large enough time period always leads toward beneficial adaptive survivability for the species... but the vast majority of individual 'adaptations' lead to the individual creature having a fundamental flaw that impairs their ability to survive in the wild.

This changes in the practical applications of technology of course... but even in this realm not all 'mutation' is to our benefit. Take like, self checkout machines at a grocery stores for example. To 'adapt' to capitalistic pressures companies invest in more and more self checkout machines, firing more employees, fewer people have jobs, fewer people make money, fewer people shop, the economy takes a hit. They adapted... but was it an intelligent adaption within the confines of a Capitalist system? Did the adaptation take into account that in order to function, the system requires that people make and spend money? I don't know that it does... and if every technological industry automated in this way together at the same time, we'd practically have an economic collapse as violent and harmful as the great depression with the amount of fundless people that the change suddenly spit out, most probably resulting in a lot of these corporations that rely heavily on purchasing power having to close up shop. An adaptation needs to be foreword thinking to be intelligent. It needs to take into account the ecosystem it exists in as well... or else have plans to somehow to GTFO of that system and utilize the adaptation within a system that the adaptation benefits them in.

How much forward thinking have ANY of these AI companies ACTUALLY done? Considering how many of them crop up and collapse every week It's basically the new cryptocurrency and NFT thing, the only thought gone into it for these companies is 'how do I adapt to make money NOW!' and it often comes at the expense of impairing their capabilities to adapt in the future. (NFT bros are basically all destitute nowadays)

There's obviously a LOT of advantages that could be gleaned from all this AI stuff, I'm simply worried that none of the people in control of these adaptations are capable or even interested in the direction of intelligent adaptation that Dr. Hawking was ACTUALLY talking about.

[Mr. X]

Why did you have to bring in politics?

[five_red]

The first problem is that you get someone with a particular, very focused, kink who will just spam out hundreds of nearly identical images. Their wonts are very simple, so there's no attempt to do anything other than serve its basic requirements. So, for example, if you like Wonder Woman with ridiculous oversized muscles then you just keep generating images of Wonder Woman with oversized muscles. Nothing else. No further context, no variety, because really your kink doesn't demand it.

The second problem is people who do want to create images with a bit of variety -- for example Wonder Woman with ridiculously oversized muscles, but in different scenarios and vignettes -- but they lack the ingenuity to push the AI in different directions to make the images interesting.

The most common mistake is to focus too much on the main character when writing the prompt. This is almost certain to give you very generic looking images. I was tempted to make a video sharing all the different tricks I've personally developed for improving the quality of AI images -- including combining different free AI tools to get specific effects -- but there's too much to cover, and I can never decide what to include and what to leave out. But here's a few basic beginner tips that I've found useful:

• If you don't specify facial expressions, the AI will just draw a bland neutral look. Phrases like "she has a shocked expression", "she looks around nervously", "she has a joyful expression", "gritted teeth with anger" nudge the AI towards not only the character's facial expressions but also their body language. Often you have to overstate the emotion, however, to get a comicbook look: "she is very shocked"
• If you don't specify lighting and shadows, the AI will often give fairly mundane looking backgrounds. Phrases like "dimly lit", "hard shadows", "a window with shafts of bright sunlight", can radically change the look and tone of an image. Also, specifying something is in silhouette can also lead to interesting effects; "a monster stands in the open doorway, silhouetted by light from the corridor beyond".
• If you don't specify a camera position, the AI will frame the image as a standard portrait. Phrases like "high vantage, looking down", "dynamic low camera, looking up", and "titled dutch angle camera" (yup, that actually works!), can all help to give the image an interesting look.
• Backgrounds are always hard to get right, but most AIs understand the concept of "in the distance" (for far background), "behind" (for things immediately behind the main character), "foreground" (for things in front of the main character), and so on. For example "a smartly dressman man pointing a gun stands in the foreground, facing the woman"
• Images look better if they tell a story. Rather than focusing on a static pose, add detail into the image that suggests it is part of a narrative. A heroine in an city alley is nice, but it's even better if there's a menacing shadowy figure standing behind her at the end of the alley.

One final point, in closing. At the moment AI seems to be a very Marmite thing: people are either okay with it, or they hate it. If people hate it, they tend to hate all AI, regardless of whether it is just spam or attempts to do something interesting. I think perhaps the spam has tainted a lot of people's opinions, such that they feel negatively towards all AI.

A better approach might be to stop talking about AI in a very binary way -- good or bad. Instead start to praise and recognise those artists (on DA and elsewhere) who use the tools in interesting ways. This might actually encourage other AI users to do better work. Perhaps a more productive discussion might be to post examples of AI work (or links to artist galleries) where you think AI is being used in discerning ways to create interesting results..? AND to explain why you think the work is interesting.


R5

[Ernie]

Great post, five_red.

[tamonicus]

Sadly, I think this is the future. As long as there's a profit motive and next to no expense to generate AI content, creators are going to think "why not?"

The AI content will improve. I hope it never gets as good as human-created content.

[Mr. X]

The first problem is that you get someone with a particular, very focused, kink who will just spam out hundreds of nearly identical images. Their wonts are very simple, so there's no attempt to do anything other than serve its basic requirements. So, for example, if you like Wonder Woman with ridiculous oversized muscles then you just keep generating images of Wonder Woman with oversized muscles. Nothing else. No further context, no variety, because really your kink doesn't demand it.

The second problem is people who do want to create images with a bit of variety -- for example Wonder Woman with ridiculously oversized muscles, but in different scenarios and vignettes -- but they lack the ingenuity to push the AI in different directions to make the images interesting.

The most common mistake is to focus too much on the main character when writing the prompt. This is almost certain to give you very generic looking images. I was tempted to make a video sharing all the different tricks I've personally developed for improving the quality of AI images -- including combining different free AI tools to get specific effects -- but there's too much to cover, and I can never decide what to include and what to leave out. But here's a few basic beginner tips that I've found useful:

• If you don't specify facial expressions, the AI will just draw a bland neutral look. Phrases like "she has a shocked expression", "she looks around nervously", "she has a joyful expression", "gritted teeth with anger" nudge the AI towards not only the character's facial expressions but also their body language. Often you have to overstate the emotion, however, to get a comicbook look: "she is very shocked"
• If you don't specify lighting and shadows, the AI will often give fairly mundane looking backgrounds. Phrases like "dimly lit", "hard shadows", "a window with shafts of bright sunlight", can radically change the look and tone of an image. Also, specifying something is in silhouette can also lead to interesting effects; "a monster stands in the open doorway, silhouetted by light from the corridor beyond".
• If you don't specify a camera position, the AI will frame the image as a standard portrait. Phrases like "high vantage, looking down", "dynamic low camera, looking up", and "titled dutch angle camera" (yup, that actually works!), can all help to give the image an interesting look.
• Backgrounds are always hard to get right, but most AIs understand the concept of "in the distance" (for far background), "behind" (for things immediately behind the main character), "foreground" (for things in front of the main character), and so on. For example "a smartly dressman man pointing a gun stands in the foreground, facing the woman"
• Images look better if they tell a story. Rather than focusing on a static pose, add detail into the image that suggests it is part of a narrative. A heroine in an city alley is nice, but it's even better if there's a menacing shadowy figure standing behind her at the end of the alley.

One final point, in closing. At the moment AI seems to be a very Marmite thing: people are either okay with it, or they hate it. If people hate it, they tend to hate all AI, regardless of whether it is just spam or attempts to do something interesting. I think perhaps the spam has tainted a lot of people's opinions, such that they feel negatively towards all AI.

A better approach might be to stop talking about AI in a very binary way -- good or bad. Instead start to praise and recognise those artists (on DA and elsewhere) who use the tools in interesting ways. This might actually encourage other AI users to do better work. Perhaps a more productive discussion might be to post examples of AI work (or links to artist galleries) where you think AI is being used in discerning ways to create interesting results..? AND to explain why you think the work is interesting.


R5

Well I wasn't arguing the tool so much as the flooding of DA with dozens and dozens of $5 galleries. This bulk buckshotting of products. Been seeing this on other sites as well like DLSite. Same with voice products. So yes the tool is fine, its the cranking out of same old stuff like a candy factory that's an issue.

[lionbadger]

Well I wasn't arguing the tool so much as the flooding of DA with dozens and dozens of $5 galleries. This bulk buckshotting of products. Been seeing this on other sites as well like DLSite. Same with voice products. So yes the tool is fine, its the cranking out of same old stuff like a candy factory that's an issue.

it's an issue if your business is selling Ms Americana comics that the average person couldn't put together but can now make a reasonable stab at themselves.

I dunno, this thread feels a bit gatekeepery, like I said, early poser was like the first doom game and was everywhere until it matured and some people really mastered it. This argument is just an evolution of "we shouldn't teach the scums to read or they won't need us!"

[Ernie]

Amen!

[Mr. X]

Well I wasn't arguing the tool so much as the flooding of DA with dozens and dozens of $5 galleries. This bulk buckshotting of products. Been seeing this on other sites as well like DLSite. Same with voice products. So yes the tool is fine, its the cranking out of same old stuff like a candy factory that's an issue.

it's an issue if your business is selling Ms Americana comics that the average person couldn't put together but can now make a reasonable stab at themselves.

I dunno, this thread feels a bit gatekeepery, like I said, early poser was like the first doom game and was everywhere until it matured and some people really mastered it. This argument is just an evolution of "we shouldn't teach the scums to read or they won't need us!"

I would love it if people made more stuff. I teach people all the time. That isn't the issue here. Please disconnect the tool concept from the flooding of relentless renders into stores and galleries. In fact I'd love to use AI to generate work.

Please, again, disconnect the tool concept from the concept of flooding galleries with dozens of pics generated from a prompt that effectively becomes spam.

[lionbadger]

again, I'm not seeing the issue if people are filling their stores and their galleries with their renders that they can now make themselves?

it all feels very "smash the cotton gins!" not sure how this is an issue

[five_red]

The issue of people selling these works is, I suspect, a red herring. The most frequent complaints about AI images is not that people are generating them, or that they are selling them, but that the generated images don't stay on a person's personal account space but are bleeding out into the public forums and algorithm driven feeds. The superheroinetf DA group shevek mentioned (full disclosure, I'm a mod there) is now almost exclusively populated by AI stuff. The plus side is that the number of images posted there in an average week is multiple times what it previously was. The downside -- depending upon your point of view -- is that a lot of the images might be perceived as rather samey.

So the volume of work being posted has shot up dramatically as the barrier to entry has crashed through the floor, but the lower barrier to entry has meant there's less originality as the AI can spit out literally hundreds of variations of the same scene in the time it might have previously taken a human artist to draw just one. If people tried to be inventive with their prompting, and posted just the one or two AI variations that worked the best, that might be okay. But all it takes is just one single over-enthusiastic user to over-share what the AI is spitting out, coupled with the fact that (unless you are expressive and creative with your prompting) the AI can generate very samey results even for different users, and you can easily end up with a lot of very similar images being fed into the algorithms, and ending up in people's feeds and/or posted to public groups.

But, as others have pointed out, any time in the past when a technology has dramatically lowered the barrier of entry in some way, there has always been a gush of low-quality work. But eventually things work out. And I expect that is what will happen with AI. In the meantime we just need to be sensible. The superheroinetfs SubReddit (full disclosure, I believe shevek is a mod there) has kept AI down to a minimum -- and for me this is probably the best solution. One forum is mostly AI, the other mostly isn't.


R5

[Femina]

again, I'm not seeing the issue if people are filling their stores and their galleries with their renders that they can now make themselves?

it all feels very "smash the cotton gins!" not sure how this is an issue

Let's be real. They didn't 'make' these pictures xD. They wrote some words in a textbox and a supercomputer made an image out of them and then they hit 'refresh' an inordinate amount of time until the supercomputer finally spit out something that the 'artist' liked. There's no real WORK going on here. They didn't 'make' anything, they just clicked a prompt until something came out. Saying they've made art out of an AI algorithm is like saying that the crew of the US Enterprise were all chefs because they knew what to tell the replicators to make it give them food.

Is it 'fair' to SELL the food you get out of the replicator? In a capitalist world I suppose yes.... I think anyone who BUYS AI artwork is kind of an idiot though. Just go click the refresh button YOURSELF, eventually it'll make you the art you want for free.

[sneakly]

again, I'm not seeing the issue if people are filling their stores and their galleries with their renders that they can now make themselves?

it all feels very "smash the cotton gins!" not sure how this is an issue

Let's be real. They didn't 'make' these pictures xD. They wrote some words in a textbox and a supercomputer made an image out of them and then they hit 'refresh' an inordinate amount of time until the supercomputer finally spit out something that the 'artist' liked. There's no real WORK going on here. They didn't 'make' anything, they just clicked a prompt until something came out. Saying they've made art out of an AI algorithm is like saying that the crew of the US Enterprise were all chefs because they knew what to tell the replicators to make it give them food.

Is it 'fair' to SELL the food you get out of the replicator? In a capitalist world I suppose yes.... I think anyone who BUYS AI artwork is kind of an idiot though. Just go click the refresh button YOURSELF, eventually it'll make you the art you want for free.

I think it can be agreed that the pictures for sale were drawn by a text prompt and not by the hand of a skilled artist. You bring up the food replicator on Star Trek. The question of having the right to sell the food depends on whether or not the person buying the food has access to the food replicator or even understands how to use it. The same could be said of Poser, Second Life or host of other computer assisted drawing tools.

We also have the right to not buy images. If you can generate the images yourself, why don't you? I have messed around with some of these programs and have had pretty limited success in producing anything beyond a momentary bit of entertainment. I have not paid to get access to programs that would allow me to do more sophisticated BDSM images with nudity. If someone does, I guess they have a product to sell, provided they have a market of people willing to pay.

Ultimately, the forums are going to evolve to accommodate the glut of new material, whether you call it art or not. If the food coming out of the replicator is better than the food you get out of your own replicator, you are likely to pay something for it. Frozen meals are cheaper than restaurant meals and easier to make than what I am willing to do for myself.

I expect artists that are popular are going to leave forums being swamped with AI to form AI free forums and some fans will follow them, some will join both and some won't. Mr X does amazing images, but he isn't going to make photo realistic images of Gal Gadot dangling from chains.

[five_red]

All Alan Moore did was write text prompts for Dave Gibbons to draw. So Moore didn't really 'make' Watchmen...

Of course Moore did more than just write text prompts, he told a story. And that is the skill with AI prompting -- not describing a picture but telling a story with an image. That's why I wrote my post above regarding basic things like lighting, camera position, facial expressions, the positioning of elements in the frame, etc. And then you can use those elements to help sell a story.

I remember seeing an interview with, I think it was, Philip Oakey (of The Human League) about when electronic music first hit the mainstream in the late 70s, and the perception that the musicians pressed a button and the synthesiser wrote the music for them. To be certain, programmable synthesisers did make producing music easier, particularly if you wanted to have a variety of sounds without mastering lots of different instruments, but you still had to be skilled as a composer to make music people actually wanted to listen to.


R5

[DonShip]

AI is a plague on all social media. It's at the point there is no time to figure out if photos are real women/celebrities, Or AI.

Example



Gorgeous, right? Yeah, except her left arm has a right hand. Or maybe both arms merge into one hand. Jeeze.

[Mr. X]

Well the lawsuits have started and so AI may get a neutering.

[Femina]

I think it can be agreed that the pictures for sale were drawn by a text prompt and not by the hand of a skilled artist. You bring up the food replicator on Star Trek. The question of having the right to sell the food depends on whether or not the person buying the food has access to the food replicator or even understands how to use it. The same could be said of Poser, Second Life or host of other computer assisted drawing tools.

We also have the right to not buy images. If you can generate the images yourself, why don't you? I have messed around with some of these programs and have had pretty limited success in producing anything beyond a momentary bit of entertainment. I have not paid to get access to programs that would allow me to do more sophisticated BDSM images with nudity. If someone does, I guess they have a product to sell, provided they have a market of people willing to pay.

Ultimately, the forums are going to evolve to accommodate the glut of new material, whether you call it art or not. If the food coming out of the replicator is better than the food you get out of your own replicator, you are likely to pay something for it. Frozen meals are cheaper than restaurant meals and easier to make than what I am willing to do for myself.

I expect artists that are popular are going to leave forums being swamped with AI to form AI free forums and some fans will follow them, some will join both and some won't. Mr X does amazing images, but he isn't going to make photo realistic images of Gal Gadot dangling from chains.

Bear with me here, some of this may sound more aggressive than it's meant to be.

This doesn't really stop the purchase of AI art as being something that's a really dumb idea from a consumer's standpoint. You CAN go do it yourself... with practically zero effort, all it takes is ten or so minutes of internet browsing to learn how to do. I once generated some 100 or so Captain Marvel images off Bing's AI imager over a half hour or so and wound up with a solid fifteen or so decent pics... but I didn't CREATE them, Bing did. I spent no more effort than I would googling images of Captain Marvel and the results were unilaterally far less expressive and appealing than the 20 or some out of 100 images I would find from a Google search five or so years ago before the great AI flood began burying them. As per likening it to something like Daz or Poser I don't think the average person understands what it takes to actually produce a render in these 3d programs that anybody actually WANTS to see. There's a 'shared' element of most of DAZ/Poser work in that the producer of the image is utilizing a group of models and tools created by other people... but they have to spend time, patience, effort, and INTENT to make an interesting image out of it... oftentimes every bit as much time and effort as somebody who can just draw it. As I've said elsewhere I have Daz 'characters' I've spent HUNDREDS of hours tweaking and perfecting, all of which typically involving as many as a dozen customized 'maps' for surfaces and the like in the effort to make the image something I WANT to see which I absolutely did create myself... and while I still would not declare myself or other Daz/Poser artists on equal footing artistically as the gifted human beings who can make magic with pen and paper (or indeed... MODEL the tools one uses in Daz/Poser which is essentially another thing entirely from the programs themselves), It remains an artistically 'expressive' medium. In order to make something that somebody wants to look at out of your 3d program, you have to set it all up juuuuust so carefully. You have to know how to EXPRESS yourself within it.

And look, no judgments here on anyone getting away with this... if you CAN make money in the system DO it, we all gotta eat... but don't DELUDE yourself into thinking you're an artist while you do it. None of those images have anything of YOU in them. They have more of EVERY OTHER ARTIST in the world inside of them than they do you, because the algorithm is taking from every artists' work that has ever been imported into the algorithm. We don't all have the same skills and abilities as one another, certain artforms are simply beyond the individual without a significant time investment, but MOST of us are fairly naturally attuned with SOMETHING that we can put our artistic energies in. If you ARE one of these people collecting batches of soulless images for sale, and you're making money doing it, I wouldn't tell you to STOP. I wouldn't even ADVISE you stop. Don't stop doing something that's putting food on the table! I WOULD advise that if you are doing this and you want to be an ARTIST... take up a hobby you enjoy OUTSIDE of the profit motivation. Find something you can put YOURSELF into on the side, and then you WILL be an artist... and maybe if you get good enough at that thing you can make money off of THAT too. I wouldn't, but again no judgments... I'm sort of insane in this venue, I can't produce ANYTHING artistically if it's for money xD.

AI is here. It's totally valid to chant 'It's here, it's queer, get used to it' and all that. It's another to say it's the 'new art'. AI isn't 'the hip new music' the old generation doesn't care for... when the old generation's music fell out of style, the new kids weren't just passively clicking buttons until a song they liked materialized out of the void, the kids were still working their assess off to create the NEW music, and still imparting all of that music with their 'soul'. Similarly Digital art didn't END the artistic expression of the artists who adopted it, it didn't make producing the art faster or less intentional, it still required their time, effort and self expression.

There is nothing intentional in AI art, 90% of the time it doesn't even recognize most of the elements of the prompt you input into it. There is no self expression in AI art. There is no SOUL in AI art. There never will be. There never can be. The AI might one day become self aware and IT may become an artist with the output. One might conceivably argue that the CODE in the algorithm is a thing of artistry and thus the developers may be artists of a sort. But nothing the algorithm outputs can ever be something that makes YOU an artist, because nothing of you is inside the art.

[sneakly]

I think it can be agreed that the pictures for sale were drawn by a text prompt and not by the hand of a skilled artist. You bring up the food replicator on Star Trek. The question of having the right to sell the food depends on whether or not the person buying the food has access to the food replicator or even understands how to use it. The same could be said of Poser, Second Life or host of other computer assisted drawing tools.

We also have the right to not buy images. If you can generate the images yourself, why don't you? I have messed around with some of these programs and have had pretty limited success in producing anything beyond a momentary bit of entertainment. I have not paid to get access to programs that would allow me to do more sophisticated BDSM images with nudity. If someone does, I guess they have a product to sell, provided they have a market of people willing to pay.

Ultimately, the forums are going to evolve to accommodate the glut of new material, whether you call it art or not. If the food coming out of the replicator is better than the food you get out of your own replicator, you are likely to pay something for it. Frozen meals are cheaper than restaurant meals and easier to make than what I am willing to do for myself.

I expect artists that are popular are going to leave forums being swamped with AI to form AI free forums and some fans will follow them, some will join both and some won't. Mr X does amazing images, but he isn't going to make photo realistic images of Gal Gadot dangling from chains.

Bear with me here, some of this may sound more aggressive than it's meant to be.

This doesn't really stop the purchase of AI art as being something that's a really dumb idea from a consumer's standpoint. You CAN go do it yourself... with practically zero effort, all it takes is ten or so minutes of internet browsing to learn how to do. I once generated some 100 or so Captain Marvel images off Bing's AI imager over a half hour or so and wound up with a solid fifteen or so decent pics... but I didn't CREATE them, Bing did. I spent no more effort than I would googling images of Captain Marvel and the results were unilaterally far less expressive and appealing than the 20 or some out of 100 images I would find from a Google search five or so years ago before the great AI flood began burying them. As per likening it to something like Daz or Poser I don't think the average person understands what it takes to actually produce a render in these 3d programs that anybody actually WANTS to see. There's a 'shared' element of most of DAZ/Poser work in that the producer of the image is utilizing a group of models and tools created by other people... but they have to spend time, patience, effort, and INTENT to make an interesting image out of it... oftentimes every bit as much time and effort as somebody who can just draw it. As I've said elsewhere I have Daz 'characters' I've spent HUNDREDS of hours tweaking and perfecting, all of which typically involving as many as a dozen customized 'maps' for surfaces and the like in the effort to make the image something I WANT to see which I absolutely did create myself... and while I still would not declare myself or other Daz/Poser artists on equal footing artistically as the gifted human beings who can make magic with pen and paper (or indeed... MODEL the tools one uses in Daz/Poser which is essentially another thing entirely from the programs themselves), It remains an artistically 'expressive' medium. In order to make something that somebody wants to look at out of your 3d program, you have to set it all up juuuuust so carefully. You have to know how to EXPRESS yourself within it.

And look, no judgments here on anyone getting away with this... if you CAN make money in the system DO it, we all gotta eat... but don't DELUDE yourself into thinking you're an artist while you do it. None of those images have anything of YOU in them. They have more of EVERY OTHER ARTIST in the world inside of them than they do you, because the algorithm is taking from every artists' work that has ever been imported into the algorithm. We don't all have the same skills and abilities as one another, certain artforms are simply beyond the individual without a significant time investment, but MOST of us are fairly naturally attuned with SOMETHING that we can put our artistic energies in. If you ARE one of these people collecting batches of soulless images for sale, and you're making money doing it, I wouldn't tell you to STOP. I wouldn't even ADVISE you stop. Don't stop doing something that's putting food on the table! I WOULD advise that if you are doing this and you want to be an ARTIST... take up a hobby you enjoy OUTSIDE of the profit motivation. Find something you can put YOURSELF into on the side, and then you WILL be an artist... and maybe if you get good enough at that thing you can make money off of THAT too. I wouldn't, but again no judgments... I'm sort of insane in this venue, I can't produce ANYTHING artistically if it's for money xD.

AI is here. It's totally valid to chant 'It's here, it's queer, get used to it' and all that. It's another to say it's the 'new art'. AI isn't 'the hip new music' the old generation doesn't care for... when the old generation's music fell out of style, the new kids weren't just passively clicking buttons until a song they liked materialized out of the void, the kids were still working their assess off to create the NEW music, and still imparting all of that music with their 'soul'. Similarly Digital art didn't END the artistic expression of the artists who adopted it, it didn't make producing the art faster or less intentional, it still required their time, effort and self expression.

There is nothing intentional in AI art, 90% of the time it doesn't even recognize most of the elements of the prompt you input into it. There is no self expression in AI art. There is no SOUL in AI art. There never will be. There never can be. The AI might one day become self aware and IT may become an artist with the output. One might conceivably argue that the CODE in the algorithm is a thing of artistry and thus the developers may be artists of a sort. But nothing the algorithm outputs can ever be something that makes YOU an artist, because nothing of you is inside the art.

I am not saying AI art makes some schmoe on DA an artist. Most of the "art" it creates is the product of scanning through millions of images and language text and tweaks to the algorithm. But ultimately, it is a tool. The static images that it currently creates are usually pretty dull and lifeless. Given some time it will be come a tool to tell stories. There will come a day when AI will be used to create a best picture. It won't be because Gal Godot, covered in mud and ropes looks good. It will be because it helped someone create a completely new view of a world that could never exist.

Unfortunately, we are going to put up with a lot of bad imagery first. The number of people that are actually good writers is smaller than the number of writers. The number of good story tellers that can write is even smaller. This isn't going to be any different.

[Femina]

I am not saying AI art makes some schmoe on DA an artist. Most of the "art" it creates is the product of scanning through millions of images and language text and tweaks to the algorithm. But ultimately, it is a tool. The static images that it currently creates are usually pretty dull and lifeless. Given some time it will be come a tool to tell stories. There will come a day when AI will be used to create a best picture. It won't be because Gal Godot, covered in mud and ropes looks good. It will be because it helped someone create a completely new view of a world that could never exist.

Unfortunately, we are going to put up with a lot of bad imagery first. The number of people that are actually good writers is smaller than the number of writers. The number of good story tellers that can write is even smaller. This isn't going to be any different.

I think you're misunderstanding my perspective? The 'TOOL' of AI is what it is. What it can and can't do is well defined, what it 'may' be in the future is, if not actually measurable, at least now firmly in the realm of the theoretical.

What I'm saying is that, (From my point of view) AI as it currently exists, regardless of how good it may become, regardless of whatever theoretically masterful visions of worlds unknown it may show us... regardless of if it ever creates images that far surpass what any human being is capable of, affecting and emotional... it will never make ANYONE an artist, because nothing of anybody will ever actually be in the 'art'. The same is true-ish of AI writing tools. Even the worst author of the worst script of the worst film imaginable... is more of an artist than someone clipping together paragraphs that an AI algorithm has spit out at them. The only reason using AI to write things isn't PLAGERISM... is because nobody actually wrote it to begin with. The way its currently utilized as a writing 'tool' is the SAME methodology as plagiarism. Again, to be clear, it's NOT Plagiarism... because nobody wrote it, but USING the algorythms sentences and paragraphs in place of one's own is the same METHOD as using another persons prewritten sentences and paragraphs in place of one's own. Be a shit author, with shit ideas, and you'll be more of an artist than you ever will be using AI. That said I think I'll grant that SOME of one's own effort does still go 'interspersed' into AI 'writing' in a way that is simply impossible with AI pictures since it allows for the 'author' to edit and manipulate the outputs freely. I 'guess' that qualifies. I guess, if you're generous with your own input while using an AI writing tool, you might still be an artist... just a lazier artist than most writers... a MUCH lazier artist... and never free of suspicion of being a hack since one probably can't ever PROVE how much of the output was them vs the machine... but I'm REALLY digressing now I think xD!

To be an artist, you have to put something of yourself into the art. I like to think you need to put EFFORT into it as well... but I'm sure there's situations and circumstances where folks have done some decent artistic whatnots with minimal effort in the whole of time and space. Or not... idk I'm not omniscient or anything... I'm not even the art police! I'm going on no campaigns against the usage of AI. These are just my opinions on the matter of their qualifications as real art or not.

[DonShip]

Just playing around with KreaAI last few days (free version, so might be my problem), I find it enhances rendered women better than line drawn, comic book art better. The rendered women get more realistic faces and hair, but I don't see much difference when I enhance a line drawn woman. Example, I enhanced Holli Wood.

[DonShip]

Some of my heroines enhanced. All rendered, but the only things really noticably enhanced are their faces and hair.

[DonShip]

I would love to be able to take the above images and POSE them differently. KreaAI hasn't allowed me to do that so far.