In his Reith lectures on Art, Grayson Perry defines Art and seperates it from craft is that Art is "tat for rich people". He states that what is defined as Art within society, at least since the 19th Century is about a small elite network, previously the Academy, now Art dealers. Mostly Art is about defining status. Art is a useless object that a rich person owns to flaunt the fact they are so rich they can buy something that has no utility at all other than to signify wealth and status. Hence Hirst's Shark, Emin's unmade bed, Mutt's urinal, NFT's.
What the layman considers art - exceptional skill in craft hasn't been a thing in Art since the invention of the Camera and mechanical reproduction (See Walter Benjamin's "The work of Art in the age of mechanical reproduction"). Labour has become divorced from production across the board.
If you get the chance, watch this incredible documentary "Ways of Seeing", produced in the 70's but still absolutely relevant today. You won't regret it. It's brilliant.
For the last seventy odd years the ability to faithfully reproduce what the eye sees is no longer important when you can just take a picture. The whole skill of using a camera is about choice in representing reality - what to shoot, how to use lenses and light, what film stock to use, how to post process it.
The Art world took a turn towards the conceptual. What you represented was more important than the technique used to represent it. Recently the art world has turned to experiential art - happenings - as a reaction against Art as a commodity.
AI art is entirely analogous to photography in that respect. It's a fusion of labour free image production and the conceptual. The skill of AI art is in prompt crafting and curation - what you choose to ask the machine to make, what language you use to make it, and of the images the machine makes, which you choose to show to the world. You can even post process and relight images. With programs like Image 2 text you can compose a composition, choose a subject - do everything a camera does.
The skill in AI art is now linguistic, rather than in learning the rudiments of colour theory etc.
I think it was Marshall Mcluhan (or it might be David Lynch) that said that the real possibilities of a medium are only revealed in the areas where the art breaks down - the distorted guitar, the digital glitch, overexposed film, etc. Arguably we've seen this first - the things that AI is bad at - hands, eyes, form, depth, are the things that make it attractive to a (capital A) Artist.
Movements in Art tend to go hand in hand with our general conceptions of the world. It represents the things we have discovered about the universe and our place in it. The terrifying (to some) thing about AI art is what it says about us as a species. The faculties we regard as exceptional - our ability to speak, produce art, reason - can be reproduced by something that has no sense of consciousness, as far as we can tell, to a level that matches the best of our species (bear in mind how embyronic these technologies are, too). Beyond the idea that AI might replace us - is the suspicion that maybe the things we think, the words we speak, might be derived in a similar statistical manner as Large Language Models. In this we're seeing Freud's Third Existential Insult to humanity writ large.
