A lot of what passes for art today is a kind of minimalism that doesn't stem from the conscientious process of reduction and decantation, resulting in something that would be better described as 'synthetic poverty'.
An approach that's commonly followed to tackle this inherent lack of richness and depth is through automated generation, and stacking layers of texture to simulate richness. This happens not only in the visual arts and interactive arts, it also happens in music.
But after a while, you begin to recognise those patterns of layered texture and those too, become recognisable defects of non-fractal repetition. Non-fractal repetition looks cheap. Lack of chaotic structure is what makes cheap CG look cheap. And it looks cheap because computing fractals to a depth that humans are no longer capable of recognising chaos is more expensive in terms of computing effort.
That being said, for some reason we have developed some tolerance for certain types of uncanny generative and layered art. In decades past when Op Art was a modern thing, it used to be easy to tell it apart from all other types of art because the computational effort needed to create such art wasn't too big. Clever computer programs today without even needing advanced AI, and they could probably create better Op Art than human artists made 50 years ago.
It won't be cheap and it won't be easy, but people who seize the means of computation will succeed. Information doesn't have volition and it can't reflect about itself, but it does disperse in ways similar to a fluid. Information that humans store outside their brains is the cultural backup of our minds. I'm tempted to say it's a law of nature that we'll eventually see a million futures exploding in a million different directions. We're approaching critical mass at an accelerated rate.