No, you're barking up the wrong tree here. I love tech, I work in the industry and intend to continue to do so. I hate AI art because it's just bad. It's awful. AI produces images that don't make physical sense, have a weird glossy "digital art" sheen and most of all have zero artistic merit. I only want to see art (and read text) produced by human beings.
Der Kuss is an oil-on-canvas painting with added gold leaf, silver and platinum. It is widely praised and held in high regard.
To me, it looks like the man has a broken neck, and the woman's been decapitated, her head rotated 90° and reattached to her torso by the ear.
This isn't to say that AI art is flawless on the first go — just that quite a lot of famous "good" art has exactly that specific flaw too.
Even though GenAI is getting better, and even though competent users know how to work around this with the right kinds of effort, normal people don't study the thing they want to get an image of before asking an AI to make it for them, and therefore will generally fail to notice when the image they've just made is still very wrong at at least one entire layer of abstraction about how the things it's depicting function or relate to each other.
Have a look at the dreck in the linked article. The software was not asked to produce a picture of an impossible waterfall that couldn't exist in our physical world. Everything about the prompt indicates the user was looking for a physically real scene. The picture of the girl in a meadow even used hashtags to indicate to the algorithm that it was expected to produce an instagram-like photo, but there's a seed pod floating in mid air, her dress is missing a critical strap, the lacework of the dress doesn't make sense and the direction of lighting is totally inconsistent across different parts of her figure.
For example:
> The software was not asked to produce a picture of an impossible waterfall that couldn't exist in our physical world.
And, given my lack of knowledge of waterfall geology, I have no idea why that's impossible.
> her dress is missing a critical strap
Her left shoulder? I've seen real dresses where one of those slips off sideways for whatever reason. It's like the dress version of self-opening flies.
If you meant something else, I missed it.
I do see the seed pod; but (perhaps due to so many photographers using fill lighting or photoshop, I don't know) I see nothing wrong with the lighting.
I see nothing noteworthy about the lacework either, but I wouldn't expect to as a guy whose total knowledge of things like that is "they are often intricate and frilly".
There are oftentimes things I can indeed spot in GenAI images, when I make them I often find it best to make four at a time, pick the one that was closest to what I wanted, then highlight the errors and do img2img for another cycle, for a dozen repetitions, to get what I want.
But most people don't look that closely.
It's like this: https://xkcd.com/1015/
https://x.com/neuralviz/status/1848176393282326595
I enjoy it and I think has tons of artistic merit