This sounds appealing, but there are two problems.
The first is that AI tools to date are incredibly limited and rigid. MJ is fixated on some arty poses (e.g. portraits with the face tilted back, eyes closed) so a lot of output defaults to those. This wouldn't be so bad if MJ gave you fine control over poses, colour, and so on. But those elements are so entwined in latent space that if you try the same prompt with a different colour you get a completely different result.
The second is that it may not matter. Human slop had taken over the Internet long before AI happened. (Content farm SEO writing, mediocre self-published genre fiction, mediocre genre art, low-effort formulaic video/movie content from the big studios, and so on.)
What's needed is inspired curation and gatekeeping. That's still happening in art to some extent, but it's a foreign concept to most of the creative industries.
So what you get is a conservative cultural process which selects unoriginal unchallenging work, especially if it's supported by effective marketing.
AI curation would be super useful, as an antidote - not just in the arts, but elsewhere.
You can imagine trained AI agents hunting through the slop and finding the gems/stand-out creators, which would add some interesting evolutionary dynamics.
Let's imagine your world, the one where an AI agent hunts through the slop and finds the stand out creators.
How would you feel if all of the stand-out creators it found were all AI? If your answer is "well, that wouldn't happen" then you may be committed to a view for ideological reasons.
Also consider that the feeds for Instagram, TikTok, YouTube etc. are more or less what you are asking for and they exist right now.
I don't think anyone considers Insta, TT, or YT the pinnacle of what's possible creatively.
As for AI gen - AI is making zero art on its own. It's all AI-assisted, whether that means a one word prompt or hours of editing.
I can imagine AI being better than humans at art of all kinds, as and when it gets a theory of mind. Whether that's a good thing or a bad one depends on how that's used. If it's designed entirely for addiction, that's a bad thing.
But we already have that.
More interestingly, I think AI has the potential to be better by breaking out of that loop.
The first is that AI tools to date are incredibly limited and rigid. MJ is fixated on some arty poses (e.g. portraits with the face tilted back, eyes closed) so a lot of output defaults to those. This wouldn't be so bad if MJ gave you fine control over poses, colour, and so on. But those elements are so entwined in latent space that if you try the same prompt with a different colour you get a completely different result.
The second is that it may not matter. Human slop had taken over the Internet long before AI happened. (Content farm SEO writing, mediocre self-published genre fiction, mediocre genre art, low-effort formulaic video/movie content from the big studios, and so on.)
What's needed is inspired curation and gatekeeping. That's still happening in art to some extent, but it's a foreign concept to most of the creative industries.
So what you get is a conservative cultural process which selects unoriginal unchallenging work, especially if it's supported by effective marketing.
AI curation would be super useful, as an antidote - not just in the arts, but elsewhere.
You can imagine trained AI agents hunting through the slop and finding the gems/stand-out creators, which would add some interesting evolutionary dynamics.