Raw LLM Responses

I agree with you if ai art was copying a very specific set of artwork and reproducing something similar and passing it of as its own, like a knockoff trying to claim to be the real thing. But the issue with this is that in most cases ai art is simply learning from the multiple styles of art and techniques, how an image looks good, just like a human would. In fact the only thing separating ai learning from seeing different art styles and a human is that ai is far more powerful. It is mixing together multiple training weights and biases to create its own version of muscle memory to create something else, which is what humans have done for millennia. It’s not taking multiple artworks, ripping off pieces and joining it together. It’s using it as a basis for adjustment. An artist owns their artwork but they cannot own a style. Hypothetically if you had a permutation and combination matrix powerful enough to create every set of weights and biases for every known style possible, then it eliminates the need for using the Internet as a dataset. I mean, I get that people who use ai art are not artists, but at the same time I am under no illusion that art is not a very selective field. Anyone can get good, but in most cases only people born with certain gifts are recognised or capable of putting down their full creative vision. At least ai art will help people with good ideas put it down on paper without having to pay an artist they probably can’t afford. It’s a support tool not a replacement