The typical "AI Artist" is more like an "explorer" going into uncharted worlds and bringing an unknown item back for our world to see. To really be an "AI artist", mabye it would take I would say 50% or more of the image transformed by direct human touch. I would say most are "AI Explorers" discovering and bring back items of interest though prompting and such. Most will bring back AI slop, some might bring back treasures.
To me it's akin to action photography such as nature or sports photography- you go in with a general idea of what to expect, but you're looking for these special moments that capture an idea you want to convey. It's true that many people will bring back "AI slop" but then aren't most photos people take the same way? When digital cameras showed up, we were inundated with low quality photos that couldn't hold a candle to professional photography. Digital artists are at a huge advantage with AI art on the post-processing end such as use of tools like photoshop, and technical artists like programmers have a greater advantage on the front end with things like ComfyUI and the vast number of dials you can turn (similar to how photographers adjust things like aperture and shutter speed). As for your comment about 50% or more transformed by human touch- 100% of AI images are transformed by human touch. First by the humans that created the initial data that the models were trained on, then by the humans who created and released various tools and models to use this data, and then finally by the person who decided what to prompt. In that way, I see AI art is a huge collaboration between many people rather than something just one human did alone.