Okay, but if theyre the one writing the prompt, changing parameters and pressing the button to generate it how are they not the one creating it?

Again, the same can be said about hiring a person on fiverr with revisions. You write out what you want, adjust parameters with revisions, and click the "send message" button, but someone else is actually making the art. Just because, in this instance, the "other person" is a computer, doesn't change the fact that the requester isn't making art.

As for the camera analogy, sure, it is similar, but again I think it's missing an important part that makes the photographer the artist. For one, the photographer HAS to have the thing they are taking a picture of directly in front of them. The have to pick an angle and framing for the shot. The picture taken consists of their perspective in that moment, without outside influence. When someone writes a prompt for image generation, they don't really have a direct effect on the output. Sure, you can say that you want a blue flower with six petals, but the chances that you get that out aren't guarenteed. You can say you want it in the style of Van Gogh, or to have the look of a specific camera, or to have specific elements blurred, etc, but ultimately what's returned isn't up to you. Every time I've ever generated AI art, it is never exactly what I went into it wanting. Sure, it gets close, but there will be details off. I've had to accept compromise with the AI. AI art will always only work without exact expectations. People interacting with the AI will only have the option of "close enough." That means fundamentally the prompter isn't in control of the actual artwork, and thus isn't actually the one creating art.