The problem on why so in its current state won’t match the human creativity is that the generative models aren’t really generative, more like compilational. They can’t generate anything special, they can’t create new genre, artstyle, etc. It would require a GAI to do that, and we are one of even a few revolutions apart from it.