As an artist, I think it is a net negative for us. Disregarding the copyright issue, I think it’s also consolidating power into large corporations, going to kill learning fundamental skills (rip next generation of artists), and turn the profession into a low skill minimum wage job. Artists that spent years learning and perfecting their skills will be worth nothing and I think it’s a pretty depressing future for us. Anways thoughts?
Disclaimer, I have spent 100 hours in the last 7 days working with AI art.
I think AI art is absolutely wonderful for art as a medium. I think art is entirely what’s in your head and the illustration part is something else. It’s a skill and a trade. Trades go away as technology advances, and it sucks for those working that trade in the short term but allows greater human productive output in the long term. But with art, I want everyone to be able to participate. It used to be that you had to master illustrating which took years and years and was practically unattainable for most people. Now anyone can create art. And skilled illustrators can create art faster. Hypothetically. We’re not quite there yet.
Drawing isn’t my strength, so my art medium is Blender. For a typical illustration, I do about 90% in Blender and draw 10% by hand. So let me talk about what I know, in the field of 3D art. My art doesn’t personally suffer from this, but I do see it a lot from other creators: In 3D, creating everything you want in a illustration is unrealistic. Therefore many creators download assets and build their scenes with those building blocks. Which is completely valid, it’s what a lot of non-3D artists do too, but you very often see those assets made by other artists strongly influence how a piece turns out. And that isn’t right. You didn’t translate your art to the screen, it got perverted by the difficulty of it. If purchasing and modifying those assets were easier, more of those 3D artists would create images closer to the ideas in their heads.
What I’ve been experimenting with now is about 30% Stable Diffusion, 40% Blender and 30% drawing on top. I absolutely love it. My workflow isn’t very optimal yet, but the quality of my art has improved. I’m excited for the AI tools to get better.
I just wanna say that I know your type of workflow and it absolutely has my respect, as you would need skills in 2d illustration as well as 3d modelling, and you would definitely need to put in the work to make good art. It’s just that AI algorithms will get better and prompting will also get better, so soon you will just need to type very little to get what you want, bypassing the need for 30% AI, 40% Blender and 30% drawing, and into 99% AI and 1% write a sentence.
Thank you. Fundamentally I think the 99% AI split is a good thing. Not for the livelihood of anyone who currently gets paid to create art, since why would a film or games company pay for a hundred artists when their competitors can get similar results with a fraction of the cost. But it’ll take power away from those companies as well. If someone with a passion for cartoons wants to make a short and are able to do so spending their afternoons for two or three months, then that’s likely to be something they do. As opposed to now where a single person would have to spend two or three years, after developing their skills, and still be left with many compromises. If that short then only gets a few views on YouTube, was it really worth years of your social life? So practically the only cartoon shorts that get made are those with backing from capitalists. And yeah, the vast vast majority of art will be low quality, derivative, and unseen, but the vast majority of art is unseen today too. I want anyone to be able to express themselves. And that’s me included. My trade isn’t actually directly in danger yet because many of the art skills I know are technical and can’t yet be done by an AI. But fuck it, I don’t like those parts of the process, I want those parts to get automated, because those parts to me are engineering and not art.