The things I quote are the respective context of the answers. They are usually counterarguments of your reasoning. Or differing opinions. That's why I bring up things.
The photographer documents a moment in time.
You seem to try reducing the work of a photographer to that of a random observer with no creative input.
In your eyes, it's inferior art. That's your opinion. But if you want to use that opinion as a premise for further argument, it is pointless. For one, that opinion is not the majority. But mostly because even bad or inferior art is art. And trying to quantify art so you can compare worth of different art is problematic. You do not compare your sunday meal to a Mona Lisa. You compare it to other meals. You compare the painting to other paintings. You compare art photography to your holiday pictures.
is worse than generative AI
As I said. Please do not go down that road of fallacies.
I sometimes argue against AI on a copyright basis
It's the same line as the plagiarism angle. To clarify the word. Plagiarism is copying parts or a whole work of someone else and publishing it under your own name. It is at core a copyright issue. With added complications that often are not legally covered. Copying an idea is not forbidden, but claiming it was your own original idea usually is. Oh, and trivially, the whole not your original idea but claiming it was aspect of plagiarism is completely void, as soon as you tell that gen ai was used.
On a technical level the copyright issue is factual wrong. You would have to accept constructed and ludicrus chains of arguments, that would include things like accepting that reading things into a memory would constitute a copyright violation. People alredy tried that years ago long before ai, for people visiting web sites. They argued that the browser would "copy" things into the memory and argued with copyright.
The output still contains elements of the Mona Lisa, but in a way that nobody can see.
Not really and that's why the copyright and plagiarism arguments fail. It contains knowledge how to construct a thing that looks like a painting. If you give the right prompts, it will generate a thing that from far away might look like that famous painting. For that particular painting though, it might contain a little bit more knowlege, because it is so ubiquitously famous.
It is not unlike a facial composite and describing features to a composite artist. The artist did not see the culprit. But prompts by the witness and a little feedback are enough to create a likeness. Because the artist does know how faces look in general and can connect that look to descriptions.
But if somebody did create a game using an AI trained entirely on their own artistic output, then I would have no objections to playing it.
I appreciate the consistency. But I do not believe that many of the vocal ai opponents would do likewise. It's why I think ethcial type arguments about the training data against ai are shortsighted.
I do not like most ai things in general. Because they look bland and boring. And because I want to spend my recreational time with human made games. And this would not change, just because there would be an ai model, that is 100% free of whatever other ai models are accused of.