Why so much hate toward AI art?

[hysepReC]

Quite a bit. The first two pics were generated by reddit user UnlikelyEmu5 (I manually edited some parts though), and he documents exactly the steps he takes to generate the images. It mostly comes down to the models you use.

But keep in mind, for every good AI image you see, there are about 100 bad gens you have to go through. For images that require no manual editing, the scale is more like 1 to 1000.

See, this is why i don't really understand the "lazy" argument. AI art with edits is somehow lazy compared to using generic HS/Koikatsu models?

At the end of the day, i don't give a shit how the art/renders were made as long as it looks good. If AI art is getting to the point where good, consistent images can be generated then i welcome it.

 

[Meaning Less]

See, this is why i don't really understand the "lazy" argument. AI art with edits is somehow lazy compared to using generic HS/Koikatsu models?

Finding the perfect image you want through a google search is also time consuming you know... Doesn't make it sound any less lazy.

In fact the time wasted to get a image that fits with what you want is another reason as to why AI won't be replacing 3D posing softwares anytime soon.

 

[hysepReC]

Finding the perfect image you want through a google search is also time consuming you know... Doesn't make it sound any less lazy.

And what about devs that just use basic HS/Koikatsu models and put no effort into making them look unique? That's lazy too. If a dev is manually editing generated images to make them look better then i don't really think it's fair to call them lazy.

 

[Meaning Less]

And what about devs that just use basic HS/Koikatsu models and put no effort into making them look unique? That's lazy too. If a dev is manually editing generated images to make them look better then i don't really think it's fair to call them lazy.

I'd say they are just different flavours of lazyness.

One is lazy for not putting the minimal effort into character design before starting their games.
The other is lazy for relying too hard on borrowed/random cgs instead of generating their own.

And even though it is dumb to not create unique characters before starting your game, at the very least they still result in consistent CGs, which is already better in my opinion than good looking but completely random AI art.

 

[ZephyrCloack]

Not to devalue the work that currently is needed to acquire a collection of AI art to suit a project, neither how complex the steps to start creating it.

Though in comparison to photography, where photographers create 10k photos per year on average (

You must be registered to see the links

), discarding the vast majority they throw to the trash, seems to me AI art generation to be the same concept.

Now there is another fact AI art can and will be perfected, sure today you arguably may not consider it lazy, just wait 5 years and the ratio of samples going to the trash will start decreasing and collections of images to be recycle as well.

On another note i don't think people consider AI art lazy, but the projects.

It doesn't matter how a dev does it, but if a dev uses 1 single pic and a wall of text, your efforts will be mostly evaluated by how well the text is written rather than how beautiful the picture is.

 

[Pr0GamerJohnny]

Even when the art won't be shit anymore, it'll still be problematic. AI can't come up with anything by itself, it has to "consume" previous art and therein lies the problem. DeviantArt is in the middle of a shitstorm over AI art because they've let AI companies use their user's content to "feed" their AI.

Basically what this guy said x100.

AI is such an overused misnomer in this day and age, especially by journalists and marketing outlets. There's really nothing "intelligent" about it, it would more accurately be called "huge data processing and interpolation"

It's an excellent tool for creating derivative works with little creative zest. If anyone's worried about AI generated art replacing something, that speaks to just how low the creative quality currently is in that sphere.

My game started out with a lot of AI weirdness that you're talking about, but I'm slowly trying to replace the worst of it and patch things up. Technology is advancing rapidly. Here's a few examples of newer images I generated/seen others generate and included/will include in the game:
View attachment 2351272 View attachment 2351273 View attachment 2351277

Aren't those just essentially remasters of existing images? Or at best a collage of numerous images?

 

[Zuburban]

You guys gotta realise that "AI art" doesn't mean that there's an actual sci-fi sentient robot painting pictures for whoever asks for it, nor is is a "collage" of pre-existing pictures. It's a VERY light piece of software scripts that learns how to depict something by looking at images to guide it's digital paintbrush, just like any real person wanting to create art would. Claiming that Stable Diffusion is "collaging" or using existing images is preposterous, original dataset of 5 BILLION images is 240 TERAbytes heavy! And Stable Diffusion models sizes? In 4-8 GIGAbytes ballpark. And no, it's not "connecting" anywhere to get info, or "search internet real-time", local UI's capable of working completely offline.

Stable Diffusion is but a TOOL, a neural network that you use for synthography. And already, at this very moment, you can use it to produce original art indistinguishable from artworks of highest quality with consistent style and characters - cause it can be TRAINED to create what you need it to create, from perfect hands and specific characters faces to sexual poses and dynamic action. And as you might have guessed - it requires time, skill and technical know-how. All that said - good luck trying to ban or track it, it's gonna be simply unenforceable at this point.

I'm gonna attach a simple explanation image as to what "diffusion" means in regard to creating art, but you more then welcome to doing bare minimum and googling "how stable diffusion works" before you start spouting nonsense when discussing "AI art".

 

[Pr0GamerJohnny]

You guys gotta realise that "AI art" doesn't mean that there's an actual sci-fi sentient robot painting pictures for whoever asks for it, nor is is a "collage" of pre-existing pictures. It's a VERY light piece of software scripts that learns how to depict something by looking at images to guide it's digital paintbrush, just like any real person wanting to create art would. Claiming that Stable Diffusion is "collaging" or using existing images is preposterous, original dataset of 5 BILLION images is 240 TERAbytes heavy! And Stable Diffusion models sizes? In 4-8 GIGAbytes ballpark. And no, it's not "connecting" anywhere to get info, or "search internet real-time", local UI's capable of working completely offline.

Stable Diffusion is but a TOOL, a neural network that you use for synthography. And already, at this very moment, you can use it to produce original art indistinguishable from artworks of highest quality with consistent style and characters - cause it can be TRAINED to create what you need it to create, from perfect hands and specific characters faces to sexual poses and dynamic action. And as you might have guessed - it requires time, skill and technical know-how. All that said - good luck trying to ban or track it, it's gonna be simply unenforceable at this point.

I'm gonna attach a simple explanation image as to what "diffusion" means in regard to creating art, but you more then welcome to doing bare minimum and googling "how stable diffusion works" before you start spouting nonsense when discussing "AI art".

Bluntly, this is just a lackey post to obfuscate the fact that at its core, it's image analysis and recognition over a wide data set. It's no more complex than setting a range of acceptable values and boundaries, be they color variance, or pixel geography. This is what a collage is, we just don't normally think of the term being iterated billions of times, but that's trivial with modern storage and processing.

I have no desire to "ban or track it" - as I said prior, AI art doing well is merely a symptom of poor existing quality in some art ecosystem.

 

[ChaosOpen]

I'd say it's less about AI art and more to do with a lack of effort. AI art can be made pretty well, I've seen it some excellent pieces. But those take an amount of effort to craft. In the same way people will tear a game a new asshole if they make lazy CG renders, the same applies to AI art. If you're just hoping to do it the "easy way" then don't expect people to be gentle.

 

[Zuburban]

Bluntly, this is just a lackey post to obfuscate the fact that at its core, it's image analysis and recognition over a wide data set. It's no more complex than setting a range of acceptable values and boundaries, be they color variance, or pixel geography. This is what a collage is, we just don't normally think of the term being iterated billions of times, but that's trivial with modern storage and processing.

I have no desire to "ban or track it" - as I said prior, AI art doing well is merely a symptom of poor existing quality in some art ecosystem.

Factually incorrect. What do you think collage even means? In your version a guy that takes pencil to draw something on piece of paper is collaging graphitte and wood cause of his ability to recognize color and shapes - rather silly.

Again - original dataset is 240TB big, models 4-8GB big - what exactly does in your opinion Stable Diffusion "collaging"? Even if each single image used in training is compressed into 1x1 resolution, a literal pixel, - it's still would be dozens upon dozens of gigabytes.

And yes it's not "intellegent" in a sense that if you leave it alone and give it nothing it starts create images by itself - it's a tool, not a sentient robot.

 

[ChaosOpen]

Factually incorrect. What do you think collage even means? In your version a guy that takes pencil to draw something on piece of paper is collaging graphitte and wood cause of his ability to recognize color and shapes - rather silly.

Again - original dataset is 240TB big, models 4-8GB big - what exactly does in your opinion Stable Diffusion "collaging"? Even if each single image used in training is compressed into 1x1 resolution, a literal pixel, - it's still would be dozens upon dozens of gigabytes.

And yes it's not "intellegent" in a sense that if you leave it alone and give it nothing it starts create images by itself - it's a tool, not a sentient robot.

He is sort of right in a sense that AI art is more of a vast set of individually colored pixels. As the computer doesn't really see an "image" when it draws, it is more in line with individual pixel colors based on a set of rules it determines. As for the billions of pictures that it has before, I believe he is referring to machine learning, in which you feed an algorithm a large number of samples and it determines rules based on what it sees in the samples.

 

[Bigass49]

AI art > 3D characters

 

[Reaper9988]

Just asking. I saw some games with AI art posted in f95 and most of them receive a lot of hate from bystander.

Why though? Is it because people lost part of their soul when their fap to stuffs generated by AI overlord?

I suppose it depends how important the Art part is to one. Command line Art just isn't artistic to me, no matter how good it gets.
It's basically just algorithms doing your work for you by deriving from the countless images they processed and it usally comes out soulless.(Now if we ever get a real sentient AI with it's own world view that Art could be interesting)

As much as I think it's cancer though I won't hate on it I just would like an easy way to ignore it, just like I try to ignore stuff with "assorted" pictures and stuff.
And well selling it is pretty borderline, but it's not like it didn't happen before.

 

[Gizzy28]

funny how there's complaints about theft in a piracy site.

 

[APoc1]

I have seen some really nice AI generated stuff but after I find out it is AI I am not less impressed or moved by the piece but I do feel like it is less special being no person pored their blood sweat and tears into it.

Sure someone had to give the AI the right prompts to get the piece but that is not the same as spending possibly hours either drawing it or rendering it. Frankly I feel the same about a good deal of the 3D daz stuff too its formulaic and lacks feeling to me.

 

[♍VoidTraveler]

Another matter which people seem to ignore is that AI art is far better than no art at all.
Hence, it will definitely be of use to new devs with limited resources who are just starting out.

 

[Meaning Less]

Another matter which people seem to ignore is that AI art is far better than no art at all.

Imagine watching your favorite anime but with daz art.

I'm not convinced, sometimes bad art can very easily tarnish a great text-based story, instead of allowing readers to fully imagine the scene the way they like you are now making them look at low effort art.

 

[♍VoidTraveler]

Imagine watching your favorite anime but with daz art.

I'm not convinced, sometimes bad art can very easily tarnish a great text-based story, instead of allowing readers to fully imagine the scene the way they like you are now making them look at low effort art.

And sometimes people don't really want to imagine the scene, and want to see some art instead to understand how the developer imagines his scene, for example. Not to mention, it won't necessarily be low-effort.

 

[Petermax]

Ai art is bad. There is something humans have that a computer will never have. That is Passion.

 

[sadservices]

ai art is simply early and too fast in its development for humans to adjust to it atm. its like digital art programs on steroids.