Honestly, as an artist, who is disabled and takes 20X times longer than anyone else to make a simple image, I can tell you multiple ways AI could be used to help people enhance or render images that are not the kind of "auto rendered from google search like prompts from other artists works" that AI is now.
Like what about an AI tool that sets up vectored lines along the lines of your sketch, where those vectors are still editable that has different settings and can learn from what you correct each time that would make vectoring line art faster? What's that? someone to vector your line art for you, based on the sketch/drawing you already did? or at least 90% of it? A dream! That is what AI COULD be being used for.
Or what about an AI based tool that was a lot more like 3d rendering where you set up planes and models, even models made from pre-made bits, or something like an advanced sims-like character creator, and pick out stock textures, or paint your own, or source them from your own photos, and etc... Like using a 3d studio to render art, which is still art, but is geared more towards 2d images and is more user friendly... lets you design and render your own original ideas, and even paint in details, or manually guide what the line weight is everywhere, or otherwise uses the kinds of calculations AI is already capable of now to help you save time filling in the details and getting the end product of your work closer to what you envisioned without spending as many grueling hours over it. Like at some point of micro managing what the AI is doing it becomes a tool again, more like a fun brush, stabilizer, or rendering tool.
Imagine a 3d map of depth you can paint over a work and then apply your own hand-painted textures to [that wraps properly and has the ai help with the perspective but give you points you can edit, and blends down with different layer styles and settings when you are done], so all the work is still yours, but you could cover a complex shape with textures like scaled and ect way faster and more easily or design lighting points and shading colours based around an actual 3d bump of your image that you paint over it in grey-scale or something... Some art programs already have similar but less "intelligent" tools that could be enhanced by the kind of stuff AI is doing now.
The AI could even generate a starting "bump map" in grey-scale for you to edit until it's actually what you want and save you some work. Something similar to existing "find edge" and "buttonize" type functions in photoshop-like programs.
There are so many ways AI technology could be made into tools, free or otherwise, that WOULD help disabled artists, or even people who aren't that good yet, or are very tired or on a time crunch, achieve their creative vision -with work that is still their own- WAY faster and easier. And NONE of that is an argument that we should have AI art the way it is now, scrapped from thousands of other people's work with all the creative decisions being made by a robot and the "artist" left -to try to manipulate what they are going for- out of something with as much comprehension as a google search engine.
As someone who does digital art and is a game making enthusiast [bent on making my own as a one person production team one day] I would LOVE to design some tools for artists to use to make their lives easier or to enable average people, or even teams to make the kinds of game assets they want without having to necessarily learn how to 3d model from the ground up, especially assets that look way more 3d than they actually are to save on processing. After years of experience, I can tell you what my ideal tools would look like, exactly how they would work, and exactly what user input would be required. I might even start making some of these tools myself for open source art programs if I even get on top of the physical space in my life enough to dive back into programming.
What AI can do now could be useful in enhancing a lot of those tools, but -even after experimenting with it thoroughly myself, I cannot say that 1. any of that justifies what AI currently is OR 2. AI and neural networks and etc are REMOTELY living up to their potential.
Of all the art programs I have used, even all the modelling ones that are useful for setting up pose references and etc.. or heck half the stuff in the sims to make characters or rooms... We ABSOLUTELY have the technology to make the kind of super user friendly, no learning curve required art making program of everyone's DREAMS that would actually turn out art that the artist built, designed, and otherwise made from the ground up.
And so long as you are busy whining to defend the most theft-driven corporate excuse of something capable of spitting out an image, we are never going to get there.
How about an AI enhance art program with tools actually designed and created by digital artists? Both digital painters, and 3d modellers? We could have that.
If you sit there satisfied with AN image that "looks nice" that a program spat out for you in a few minutes with no design work on your part, regardless of who's work was stolen to make that, companies are never going to give us the AI based art tools that we all actually want.
There is a difference between "keep any and all calculation enhanced art tools out of digital art forever" and just trying to point out that what AI art as it is NOW is bullshit. The people against AI AREN'T against you or anyone having tools to make your life easier, they are against programs being used to rip off other artists work without credit to churn out something formulaic -and let's face it- not really the final vision you were trying to achieve anyway.
