AI is a tool. It lowers the bar of entry tremendously so that you get gazillions of pictures of people whose creativity goes all the way to "blonde woman in skimpy clothing" and "anime girl with big jugs". IT should come at no surprise. These people might not have a good grasp on e.g. anatomy, what good image composition is, hell they might even have poor taste. "one anime girl, big boobs, smiling" takes not a lot of effort to type. If these people could draw, that's what they would draw and it wouldn't be anymore interesting to look at. Hell, there's a ton of artists that actually can draw and that's basically all they draw. With AI it's faster and easier, that's about it. Non-AI art is more often than not garbage too. There's a ton of art lying in some artists drawer, basement or on a harddrive that nobody cares to look at because it's just not that special or interesting. There's also no such thing as objectively good art anyways. For every famous piece of art you'll find a horde of art critics hating it.
It's kinda hard to argue with people about the merits how useful AI art is if they simply don't know what is possible with the tools that exist. As first step, at the current stage of technology, I would move away from seeing it as a thing the AI does. The user is in control. The user makes the art. AI art implies it's art made by some semi-intelligent automaton, which especially for more involved workflows is simply not the case. The text prompt you know from Dall-E and such by itself is a very basic and limited way to interact with image generation models. You can do a lot more with them and you can control their outputs actually pretty well. It just takes skill, knowledge and time. For example you can actually completely control what a scene consists of. You can pose people in generations, down to their fingers. You can insert and remove objects in specific places. You can even control where the light is coming from. I am sure some people reading this now didn't even know that. There is a bar of entry there the "anime girl with big jugs" enjoyer will never cross. It won't stop him from posting 500 pictures like that a day, though. Every time you see fucked up eyes, muscle definitons that are confusing, skin textures that looks alarming, shadows that don't make sense, fingers that are fucked up, style that changes from picture to picture, pixel art that has weird half pixels and artifacts etc... it's because the people generating that picture didn't care about it and just puked it on the internet. It is perfectly possible to get these things right. That needs technique, a process and work. I feel people are in general not very willing to share these workflows, because they see an advantage in "owning" them and not giving others a leg up. Then all of this is very fast moving and hasn't been around for long. So not many people even know how to use all this technology in a sensible way.
I'm old enough to remember when "digital art" slowly started to become a thing, and it was quite similar then. A lot of overnight sensations and dead ends then too. Also a lot of people looking down on it.
Thing is also, AI art won't end here for any appreciable amount of time. It's just gonna get crazier.
