Ok but see why my example of photography in particular applies here.

A photographer can use a specific camera with a specific lens, specific camera settings, shutter speed, film stock etc. They can photograph a specific city skyline from a specific angle and distance and elevation at a certain time of day/night with specific weather and visibility etc. They could then take a unique picture.

The photographer has made all of those creative choices and hardware selection to compose the shot, which is a primary argument as to why a photograph is copyrightable and art.

The exact same photographer could also use a popular imaging-creation AI, and almost perfectly re-create the real photograph they took via AI by carefully using all of those same exact creative choices down to the lens type, in the software prompt.

So the same person would have created a nearly identical image from two different methods. Yet one is currently copyrightable and one is not - despite the same creator putting the same effort and knowledge into composing both versions of the final image.