I'm an artist and I find AI generated art fascinating. Is it art? Yes it is. Thd definition of "Art" is anything artificially created (so not spontaniously created by nature or inanimate processes) that doesn't by itself has any real functionality. You could argue that a beautiful vase is a piece of art functional, but in that case the design aspect is the art. The vase doesn't need to be beautiful to be functional. In fact, if the vase was just a plain plastic cylinder it would have the same functionality. So AI "art" is, like it or not, art. But so is a thoughtless scribble on some scrap paper, or a banana taped to a wall. It doesn't mean it's aesthetically any good, or gives it any value. Value is entirely subjective and up to the viewer to decide on. Hence the stupid prices paid for some items -like that banana taped to the wall.
When people say AI art has no soul, that's kind of relative as well. There's a theoretical (though incredibly slim) chance that an AI will generate a perfect, beautiful piece from scratch. But most of the time it has flaws, if not visual, then it's contextual flaws. I like to compare an AI program to a monkey. The monkey is incredibly good at copying aspects of the art creating process, blending colours, brushstrokes, even stylistic choices. The monkey is lightning fast too. However the monkey doesn't understand what it's painting. In time, the monkey might refine it's techniques, make less obvious mistakes, but in the end, it will still be a monkey. Currently it's all about curation, and (extensively) fixing by a skilled artist/designer of the generated outputs. A person who hasn't got the trained eye of a creator, can easily be wowed by most current AI outputs. This is why so many AI art that's being posted looks uncanny, or soulless, because it's full of small flaws that your brain picks up on as "wrong".
I think AI will make for a great tool in the workflow of digital artists though. The dogswork of any piece is right in the middle of the process. The initial work; the sketching, the composition, laying down the base colours and shading is fast and fun. The detailing part and touch ups at the very end is pretty quick and satisfying too. It's the stuff in the middle, the the tedious blending, adjusting and correcting that takes hours if not days and is terribly boring. Now what if an artist sets up an rough image, gives this image to the AI to interpret in a style of their choice, and then take this hybrid output and proceed with the final phase as usual. It would supercharge that artist, making it possible to bring more ideas to life that would otherwise never seen the light of day. Every artist is better off with an apprentice to do the drudgework, even if that apprentice is a highly skilled monkey. 😉
