Star Trek AI art thread

[Michael]

Nothing in here is art, AI cannot produce art, art requires a human creating it and no, typing prompts does not count as creating.

Personally I’m less hung up about whether AI generated imagery can be called art or not, because to me art in the most general of terms is mostly what we humans point to and say it’s art. The intent behind a piece trumps the notion of actually crafting a piece, IMHO. And there’s generative art out there where clever people coded super interesting, abstract experiences that practically no-one would second guess as being art.

The aspect that I can’t get over regarding AI images is that of theft and plagiarism, because none of the images you see here would exist if they didn't just went ahead and stole all the art they could find to feed their algorithm. In my view it’s super unethical and also devalues the work of countless, real artists. And nothing AI art can ever do will let me forget that fact.

 

[BillJ]

In my view it’s super unethical and also devalues the work of countless, real artists.

Aren't human artists also influenced by all the things they see and are exposed to?

 

[Michael]

Aren't human artists also influenced by all the things they see and are exposed to?

That’s true. But from my understanding the way machine learning algorithms are incorporating the “inspiration” that they are fed is different from the way human artists see something they then try to emulate in their own art. The difference is that you will see elements in AI generated images that are clearly copied verbatim. If a human artist would go ahead and trace or copy existing art and present it as their own, most people would rightly call that plagiarism. Problem is the AI just can’t help it.

But yeah, it’s a complicated matter and I’ll admit I don’t get all the technical aspects. I just wish they would have asked artists for their consent or opt-ins before just going ahead and use everything there is to feed the AI. It’s just unethical in my view.

 

[BillJ]

But yeah, it’s a complicated matter and I’ll admit I don’t get all the technical aspects. I just wish they would have asked artists for their consent or opt-ins before just going ahead and use everything there is to feed the AI. It’s just unethical in my view.

It will be super weird when we have AI buzzing around making its own choices on all kinds of subjects. That is when it will really blur, when AI is making its own choices.

 

[Michael]

It will be super weird when we have AI buzzing around making its own choices on all kinds of subjects. That is when it will really blur, when AI is making its own choices.

Agreed. Although, from my understanding, we are still a long way away from something approaching actual consciousness. Stuff like ChatGPT and Midjourney might sometimes seem like it’s almost alive, but from what I’ve been reading, that’s more or less an illusion. And there seem to be experts who’re saying calling what we have now “artificial intelligence” was a mistake, because that’s apparently not what it is right now or tomorrow.

 

[Michael]

Another aspect that’s hugely problematic about machine learning and that’s not talked about enough is the way people in third world countries are exploited to make it all work in the first place. The reason why Midjourney might understand what a helmet is when you prompt it to generate an image of one, is because there are hundreds of thousand of workers – mostly people of color living on the poverty line – sitting in giant call center like offices looking at photos all day and marking up the helmet in the picture. We like to think AI is pretty clever with all it’s already capable of doing right now, but it certainly comes at a cost. And as per usual it’s not people from wealthy nations having to pay that price.

 

[F. King Daniel]

Nothing in here is art, AI cannot produce art, art requires a human creating it and no, typing prompts does not count as creating.

But it makes for amazing, low effort memes and shitposts.

 

[137th Gebirg]

Nothing in here is art, AI cannot produce art, art requires a human creating it and no, typing prompts does not count as creating.

This sounds like it was written by one of my insufferably smug art professors back in college 30+ years ago, who always authoritatively pontificated about what they thought "good art" was supposed to be.

A coder produced the algorithm which is, in-and-of-itself, an art form that happens to build other forms of art. That's "art" enough for me.

 

[Captain pl1ngpl0ng]

none of the images you see here would exist if they didn't just went ahead and stole all the art they could find to feed their algorithm. In my view it’s super unethical and also devalues the work of countless, real artists. And nothing AI art can ever do will let me forget that fact.

this picture here has never ever existed before the poster dragged it out from the abyss that is the interweb...
that is true..
it is something totally new... ..i dont see anything unethical here.
People (like me) who investigate the possibilities AI opens up, we don't see us self's as "artists" or "thief's"
....we are explorers....... we are exploring a tool that will accelerate production of games and movies,
i think it can inspire "real" art...

Sad Drunk Spock.

 

[blssdwlf]

this picture here has never ever existed before the poster dragged it out from the abyss that is the interweb...
that is true..
it is something totally new... ..i dont see anything unethical here.
People (like me) who investigate the possibilities AI opens up, we don't see us self's as "artists" or "thief's"
....we are explorers....... we are exploring a tool that will accelerate production of games and movies,
i think it can inspire "real" art...

Following on Michael's comment and yours... would the AI image generator you used be able to generate your "Sad Drunk Spock" image if it was not trained on any images? Just curious.

 

[SolarisOne]

Sad Drunk Spock.
2 years ago

Now

Wait, are you stra-tek on Tumblr?

 

[Captain pl1ngpl0ng]

Following on Michael's comment and yours... would the AI image generator you used be able to generate your "Sad Drunk Spock" image if it was not trained on any images? Just curious.

no.. i cant see how that would be possible,
that's like asking someone to make a collage from magazine cutouts, without any magazines.....

SD-Automatic1111 (my choice of ai) was trained on several huge datasets, and all of them got their images from Common Crawl, basically a copy of internet. it is impossible to know what an image generated by ai is getting it's information from or to what extent it uses that information.

 

[F. King Daniel]

Wait, are you stra-tek on Tumblr?

It's a me!
I used to have a link on my sig.

 

[publiusr]

This might be useful for OTOY
https://techxplore.com/news/2023-11-machine-users-superhuman-ability-tools.amp

Here you can spin around the three footer…even if you don’t actually…have it…

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

 

[publiusr]

Just today www.phys.org had a paper called:
"AI model instantly generates 3D image from 2D sample"
--that details Adobe research at ANU...
(LRM: Large Reconstruction Model for single image to 3D).

more
https://techxplore.com/news/2023-11-peek-future-visual-framework-generative.html
[URL]https://medicalxpress.com/news/2023-11-ai-real-actual-human.html[/URL]
[URL]https://techxplore.com/news/2023-11-ai-human-brain.html[/URL]

 

[Captain pl1ngpl0ng]

To anyone who is interested


Ai does NOT copy parts of images from the interent, mixing them together and creating a new image.
This is NOT how it works...

Ai are only TRAINED on images from the internet and how they look, it does not copy....

How is it trained?
it learned how an image gradually change from an image to an unrecognizable soup of random noise (diffusion of an image),
After training, the ai have a noise predictor capable of estimating the noise added to an image.

quote:
"
1.Pick a training image, like a photo of a cat.
2.Generate a random noise image.then..
3.Corrupt the training image by adding this noisy image up to a certain number of steps.
4.Teach the noise predictor to tell us how much noise was added. This is done by tuning its weights and showing it the correct answer.
"

Now the AI can do a reverse diffusion.
From our text input/prompt it uses the noise predictor to estimate how much noise needed to subtract from each step to make an image.

the image produced does NOT come from any parts of artwork grabbed from the internet,
the image is simply created.
and here is a big point: I can run the ai locally without any internet connection!
That is because ai does not store images, it store the informaition about how to create them.

think of it like this:
when you do math like 6-2=4 and 4+4=8 etc.....
you dont store/remember every answer on every possible equation, you just remember the methode.

EDIT: an in depth link: https://stable-diffusion-art.com/how-stable-diffusion-work/

 

[F. King Daniel]

Even if it did, how is it any different from us copying the style of an artist we like? Artists are saying they should give permission for their stuff to be used, but I sure as hell didn't ask Paramount's when I was drawing USS Enterprises as a kid.

 

[Captain pl1ngpl0ng]

^ sure, agree, when we add a name like Rembrant or Andy Warhol to the prompt, the ai recognise the style from it's training, but again, it does NOT copy or steal anything. I just felt it was important to get that through.

 

[137th Gebirg]

There are offline image generators? I like things not reliant on the internet for constant operations, especially with potentially dodgy persistent connections. Where do these exist? Are they available anywhere for download?

 

[Captain pl1ngpl0ng]

i use stable diffusion automatic1111, its not like a discord ai. it is a stand alone Ai. wait a minute i will dig out a link to a recipe on how to install it...