I think AI art is comparable to photography. Photographers do a lot of work behind the scenes to get everything set up, the equipment, lighting, angles, lenses, etc, But at the end of the day, the only action they're taking to capture the art is they press a button, it's not nearly the same amount of work that a painter or a musician puts into their art. So I think the idea of "capturing" art is still a valid thing. Sometimes a photographer can capture an award-winning masterpiece with a spur-of-the-moment photo on some shitty disposable camera. Maybe it took them 1000 bad photos to get that one photo, but they still just captured it from somewhere else, they didn't create the work.

Similarly with AI, a person may have to work with the AI software to setup and craft the prompt that will eventually generate the art, then there may be dozens of iterations of that and fine-tuning to get the result they're imagining, and even after that there may be some photoshopping involved to get it to where they want it. They're capturing artwork from a source that may not be their own creation, just the same as photographers. I think AI art is just as legitimate as other forms of art, it's just open to a wider range of people that can participate because many of the physical hurdles (equipment, space, time, lighting, etc) are not as much of an issue.