AI generated art

[Deleted member 5954819]

Show me any pixel that it has copied.

What exactly should I show you? You just need to study the basics of that AI, which obviously doesn't interest you at all...

Any AI designed to create some kind of art is based on patterns and data that it has already obtained from existing images or photographs. What the AI will do is copy the textures, patterns, or style of the original art and use it for its "own" art. I'll write it again: COPIES—that's not the same as being inspired. In practice, this means that if there is a work such as the painting "Mona Lisa", then somewhere nowadays there is already art that is the same as this painting, only it has a different background and a different person. The fact that you change two things does not mean that you have created something new or original. You just copied it and made it the way you liked it. In short, it creates what someone has already created long before; this is how it works and nothing else.

To understand each other. I'm not saying that AI does a bad job, but if someone calls it artistic and considers it something that an ordinary person can't create, then he's really a complete fool because everything is created only thanks to the fact that some real person has already created something similar before.

 

[XcentY]

It's a shame that it doesn't produce anything; it just copies. These are exactly the facts that you obviously do not understand at all.

What you don't understand is that it does copy as much as an artist copy...
AI is capable of producing images never seen and never produced before.
It does it by activating neurons which stored memories of what it had seen before.

The same goes for an artist.
Artist is capable of producing images never seen and never produced before.
He does it by activating neurons which stored memories of what he had seen before.

That what you don't understand... Human doesn't come with an inner ability to draw an apple...
They have to see an apple before drawing it.
If they have seen a lot of fruits.
They will even be able to draw a fruit which doesn't exist because it will share some characteristics that we identify as coming from a fruit.
The same goes for the generative AI
IA has to see an apple before drawing it.
It you show a lot of fruits to an AI,
AI will even be able to draw a fruit which doesn't exist because it will share some characteristics that we identify as coming from a fruit.

Here, I just prompted an AI to draw me a fruit which doesn't exist on earth

 

[Deleted member 5954819]

The same goes for an artist.

In that case, I wonder what Picasso, Monet, Vermeer, Michelangelo, Van Gogh, Da Vinci, or many other real artists copied when you claim that every artist copies. There is a big difference between copying and inspiration; you obviously confuse these two terms a lot.

While the AI creates something only based on the patterns of real works that it remembers, you can create something new and original that arises in your head, for example, while listening to music. Unfortunately, there is a time when almost everything exists, and there is a lot of everything. But that does not mean that something new cannot be invented and that every artist is limited to copying it.

 

[XcentY]

In that case, I wonder what Picasso, Monet, Vermeer, Michelangelo, Van Gogh, Da Vinci, or many other real artists copied when you claim that every artist copies. There is a big difference between copying and inspiration; you obviously confuse these two terms a lot.

While the AI creates something only based on the patterns of real works that it remembers, you can create something new and original that arises in your head, for example, while listening to music. Unfortunately, there is a time when almost everything exists, and there is a lot of everything. But that does not mean that something new cannot be invented and that every artist is limited to copying it.

Every artist copy. It doesn't mean that what they come up with is not unique.
But takes by example the well known "Creation of Adam"
It's a human with a beard inside a tissue shaped as a human brain pointing his finger to another human younger...
Has the brain of michelangelo not recorded in his memories what a human looks like... or maybe you think the shape of a human is inspired to michelangelo by some higher being ?


Then what inspired the AI to come up with a fruit which doesn't exist if it's not the prompt I wrote ?


So did michelangelo painted "Creation of Adam" inspired by a prompt in his mind ?

You must be registered to see the links

 

[Meaning Less]

Here, I just prompted an AI to draw me a fruit which doesn't exist on earth
View attachment 2565601

Funny how it took your description literally which is why the fruit is "on earth". And to be fair it just looks like a mix of existing fruits like fig, apple, banana.

I'm sure AI art has its place, it is great for personal use, placeholders, backgrounds, or things that aren't in focus and don't involve much thinking.

However I doubt it will replace artists anytime soon, especially since you can't own the copyrights of what you generate with AI, what means everyone else can reuse pictures you churn out from it.

 

[XcentY]

Funny how it took your description literally which is why the fruit is "on earth". And to be fair it just looks like a mix of existing fruits like fig, apple, banana.

I'm sure AI art has its place, it is great for personal use, placeholders, backgrounds, or things that aren't in focus and don't involve much thinking.

However I doubt it will replace artists anytime soon, especially since you can't own the copyrights of what you generate with AI, what means everyone else can reuse pictures you churn out from it.

To be fair, when botanists recorded all the fruits, they've seen that many fruits are sharing common characteristics...
If I ask myself to draw a fruit which doesn't exist on earth... I'm sure you'll say that my drawing looks like a mix of existing fruits...
Because your brain will only try to focus on what he knows... And so you dismiss what AI is producing being of mix of what your brain has recognized based on your own neuron memories... Woaw... Let's dismiss it then !!!
Let's also dismiss the fruits created in Avatar... They too looks like a mix of existing fruits...


But which apple was it ?


Which fig ?



Which banana ?



I don't say it will replace artists... There will be another type of creators that will use AI as a tool to craft new art. Also if we were able to store all the images produced in all the human brains everyday, I'm sure you would find a lot of similar images... Let's just take a little example.
If I say "Michaël Jackson, Thriller"
I'm sure, if you're familiar with the pop culture of the 80's, your brain already produce all kind of informations like images, sound, you may even be able to reproduce the song in your head...

Or If I ask you to hum a song like the American Anthem... Is it by copying, reconstruction or by inspiration that your brain comes up with the melody ?

 

[Meaning Less]

To be fair, when botanists recorded all the fruits, they've seen that many fruits are sharing common characteristics...
If I ask myself to draw a fruit which doesn't exist on earth... I'm sure you'll say that my drawing looks like a mix of existing fruits...

I think you are overethinking about what the AI did there, it just mixed different things it WAS trained with to reach a mixed fruit.

Many fruits share common characteriscs because they are all earth fruits... That's why I said it wasn't the best example of creativity, in fact if you told me that was a real fruit found here on earth I might believe you (except for that random leaf that was poorly placed).

 

[PaullT]

I consider myself a good writer and wanted to attempt to make a visual novel using AI generated backgrounds (MidJourney) with Daz 3d models since I'm no artist. It's not going well so far. My every attempt to make a scene has silly artifacts in it, mostly related to chairs, plant placement, shadows, and room layout. This is the best I could come up with. The problem is that they're all supposed to be scenes from the same college and they just aren't, lol.

 

[XcentY]

I think you are overethinking about what the AI did there, it just mixed different things it WAS trained with to reach a mixed fruit.

Many fruits share common characteriscs because they are all earth fruits... That's why I said it wasn't the best example of creativity, in fact if you told me that was a real fruit found here on earth I might believe you (except for that random leaf that was poorly placed).

Here is a fruit you can find on Pandora...
A fruit created by some artist for the movie Avatar




It's not made by an AI but yet it looks like a mix of earth fruits...

You see, I don't overthing. I think you don't really know how neural network proceed to generate... and how your brain functions...
You think your brain creates art from scratch... Just like from the void came the light... However it's not like that. Your brain is full of tiny bits of informations and when you create, you don't even notice how these neurons will mix informations you already know...
We don't talk about choosing images in libraries and mixing them...
That's not how neural network works....

Now that said... For me, AI is producing art. Yet it's not the AI itself cause it needs human interaction to learn and be prompted... However the way it generates images is not by mixing images but by activating neurons which will return informations to generate an image. I built AI myself from scratch. developping the neural network and feeding it with thousands of samples... But yet... with a simple 2 neurons network, I was able to make the AI learn Payload treshold of different car models but how the AI is capable of generalizing is buffling me. Because nowhere the information has been given to the AI and as we talk about different models. 2 neurons are normally not capable of holding this information but yet it works.

 

[Meaning Less]

It's not made by an AI but yet it looks like a mix of earth fruits...

Sure, the difference is that one was intentional, the other was accidental.

Stumbling upon an image is not exactly the same as creating it.

And the more specific you wanted it to be the harder it will be for the AI to assemble something close to what you want.

 

[XcentY]

Sure, the difference is that one was intentional, the other was accidental.

Stumbling upon an image is not exactly the same as creating it.

And the more specific you wanted it to be the harder it will be for the AI to assemble something close to what you want.

Of course...
It's like if I was asking you to imagine something very specific.
You will certainly have an image forming in your brain but yet it won't be totally what I want.
The same goes for AI.... If I ask the AI something very specific, the image forming in the AI's brain won't be totally what I want.
And that's because it's not my memories but AI's memories...
What you're telling is just that you think of an AI as an extension of your own brain. It isn't

So here if you ask yourself something very specific, you'll come up with something specific to you forming from your own memories. mixing them to even think that this is something new... and it will be intentional.
However with AI it's like you ask another brain something very specific. So you'll have to learn how this other brain comprehend your specific request before you can use AI as a tool to closely match your specific requests that will be less and less accidental...

 

[Meaning Less]

And that's because it's not my memories but AI's memories...

Saying that AI has memory is a big stretch. It doesn't know what it is doing, it is code that repeats patterns it was trained to repeat.

However with AI it's like you ask another brain something very specific. So you'll have to learn how this other brain comprehend your specific request before you can use AI as a tool to closely match your specific requests that will be less and less accidental...

Altough the main difference is that AI has no consciousness, and it has no ability to make anything new, it can only mix and match things it was trained with randomly and chaotically because that's what you get by feeding it with tons of images, then it is trained to stop doing it.

An artist can describe you every single artistic choice they made during the drawing process, with an AI, it is just throwing things together accidentally until it doesn't look like pure chaos for other humans.

And unfortunetely the more effective it becomes at not showing you complete chaos is also when you start noticing patterns and similar things that most likely are artifacts of the original artworks it was trained with.

 

[XcentY]

Saying that AI has memory is a big stretch. It doesn't know what it is doing, it is code that repeats patterns it was trained to repeat.

Altough the main difference is that AI has no consciousness, and it has no ability to make anything new, it can only mix and match things it was trained with randomly and chaotically because that's what you get by feeding it with tons of images, then it is trained to stop doing it.

An artist can describe you every single artistic choice they made during the drawing process, with an AI, it is just throwing things together accidentally until it doesn't look like pure chaos for other humans.

And unfortunetely the more effective it becomes at not showing you complete chaos is also when you start noticing patterns and similar things that most likely are artifacts of the original artworks it was trained with.

Because you know what your brain is doing ?

And you think you do something new ?

I think you're overrating your capabilities

Also, the thing is that you totally don't see what's behind and you just judge it from your perspective.

While in fact even the total chaos to your eyes is something that involved neural network memory retrieval.
It's like If you were saying... my dream was so blurry. Total chaos... But yet it was created by your brain.
You judge something by your values.
And of course you'll recognize patterns and artworks of what he learned... Like when you imagine something , if you were honnest with yourself, you could see patterns and artworks you've learned before...

But I will leave you here... We're totally not on the same page.
You don't see how it function and you just await the AI to react one way while it's not.
And sorry but your knowledge on how the brain function is just not going to help you in the matter.
You don't even see the subconsious process of your imagination... You think that when you think of a cat, it only involves conscious thinking but it's not the way your brain works... If you could even deeply think on how when you want to think of a cat, your brain activates a lot of neurons that you're not conscious of... But leave it here... I'm just going to stop debatting...
You don't want to understand how some processes that mimics the mind are just out of your reach...

 

[Kodra_Dev]

The amount of people who don't know how AI works at all but hold a strong opinion over it in this thread is... fascinating.

 

[DuniX]

I consider myself a good writer and wanted to attempt to make a visual novel using AI generated backgrounds (MidJourney) with Daz 3d models since I'm no artist. It's not going well so far. My every attempt to make a scene has silly artifacts in it, mostly related to chairs, plant placement, shadows, and room layout. This is the best I could come up with. The problem is that they're all supposed to be scenes from the same college and they just aren't, lol.

This is the real problem with AIs which is Consistency, Coherency and Intention.
This is because AIs do not have a "Consciousness" so that it can Logically Chain things together and think things through Rationally. Even with Auto whatever that is still just creating Procedural Scripts that are executed and not any kind of Rational Thinking.
It is a essential Filtering and Editing step that is Missing. It's more equivalent to unrestrained daydreaming.

For now that it is left to the humans through the Prompts and establishing Context.
We are going to get a some form of rudimentary memory one way or another so at least the problem of Consistency will be alleviated somewhat.

 

[Deleted member 5954819]

If you want to check some content...
Here a site you might like :

Spoiler: NSFW content

You don't have permission to view the spoiler content. Log in or register now.

They have a lot of images generated by AI. Mostly NSFW content.
You'll spot the same character quite often. I think it's their use of the same LORA model (maybe chilloutmix)

If you want to know the purpose of a LORA model :

You must be registered to see the links

And if you want to download all the images.
I give you a little tutorial. Very easy (done with firefox) :
1) Download WFDownloader App. (This app gives the ability to download images with a pattern)
1) Navigate to one of the image you like. (Example :

You must be registered to see the links

)
View attachment 2566923
2) Once you're on the image, right click and select "Inspect"
3) It will show you the dev console with the HTML tags of the DIV layer you right clicked.
View attachment 2566932
4) The important line is the one I highlighted where you can see the url of the image
5) Right click on the line and select "copy link address"
6) Check the amount of photos available in the content View attachment 2566938
7) Open WFDownloader App. Then click on "Add" and select "Batch download links with pattern"
View attachment 2566939
8) Enter a name for your batch
9) Paste the recorded url from point 5 in the pattern address
10) Edit the url to get the whole pic folder by adding limit values between brackets.
Example :

You must be registered to see the links

11) give a save folder on your computer
12) It should then looks similar to this :
View attachment 2566943

13) You're good to go. Confirm and play your batch. It will download all the pictures.

But that's still not something that a person can't create in 10 minutes. Here are people who have said that AI can make art that a human can't and that it's better than what a human can do overall. But I don't see any evidence of that. The only thing I can see is that anyone with access to that AI can do it and doesn't need any modeling, rendering, or drawing skills. Based on this, it will almost never have more success or recognition than the work of a real person, and we don't have to deny that, because I could tell you at least five examples even in the middle of the night when it really showed itself in many things

 

[XcentY]

But that's still not something that a person can't create in 10 minutes. Here are people who have said that AI can make art that a human can't and that it's better than what a human can do overall. But I don't see any evidence of that. The only thing I can see is that anyone with access to that AI can do it and doesn't need any modeling, rendering, or drawing skills. Based on this, it will almost never have more success or recognition than the work of a real person, and we don't have to deny that, because I could tell you at least five examples even in the middle of the night when it really showed itself in many things

Sorry but I didn't say that a person can't create such content without the help of AI.
And sorry to disappoint you but I don't care about someone's success or recognition. If you're drawn to that kind of feedback from me, you hit the wrong door.
But to say it can be done in 10 mins by a person... I'd like to see you perform ....

And If I wanted to say that AI outperform human. It's not in the creation itself cause a human is quite capable to create such content but rather to the limitless creation and the amount of content it generates. Google is forecasting 90% of AI generated content by 2025

 

[Deleted member 5954819]

But to say it can be done in 10 mins by a person... I'd like to see you perform ....

Since my visuals don't resemble the style of what you posted here, it probably doesn't make sense to show something I made. But you will find many creators here who do this style, and considering that they are able to create about 500 images per month and render them for their game, I think that I have not written something unrealistic (You can often find visually processed games like this on itch.io if you would like to see some).

Google is forecasting 90% of AI generated content by 2025

I can also predict that when you release the game, it will be extremely successful and you will earn millions. But the reality can be completely different. So we'll see in 2025

 

[XcentY]

Since my visuals don't resemble the style of what you posted here, it probably doesn't make sense to show something I made. But you will find many creators here who do this style, and considering that they are able to create about 500 images per month and render them for their game, I think that I have not written something unrealistic (You can often find visually processed games like this on itch.io if you would like to see some).


I can also predict that when you release the game, it will be extremely successful and you will earn millions. But the reality can be completely different. So we'll see in 2025

500 images a month without the help of AI ?
I doubt...
Don't say they use tools like photoshop, After Effect, Blender, Unity, Unreal Engine and so on... It's full of AI helping tools...

 

[Deleted member 5954819]

500 images a month without the help of AI ?
I doubt...
Don't say they use tools like photoshop, After Effect, Blender and so on... It's full of AI helping tools...

I didn't mean Photoshop; I meant pure software like Blender or DAZ Studio. And maybe it sounds incredible, but I am writing to you about real numbers.

As a perfect example, I will mention a developer named DrPinkCake, the creator of the games Being a DIK. In the nine months it took to develop his last update, 5000 static renders and animations were created (according to the information provided via his Patreon). If you calculate it, it is approximately 550–560 images per month, and it must be taken into account that it is not only about creating the image but also clothes, light, environment, and similar things before you get to the result. In this time frame, he managed to do post-work (which I don't want to take into account since we don't deal with Photoshop here) and coding. And yes, by all accounts, he should be working alone.

Of course, it's definitely easier for him since he has over 15,000 Patreons, so he doesn't have to deal with the financial situation and problems of an ordinary person. But the fact remains that it is not unrealistic.

Edit: No, he doesn't do visuals in the style shown here by AI, but it's still some art and not exactly low quality, so I thought it was a good example.