12/30/2022
“AI Art” as a descriptive term was an unfortunate choice for many reasons and its likely too late now to meaningfully change this, but I can’t help sharing some thoughts about it.
The key problem with the term is that it immediately throws the question of authorship of the Art into question and leads many to conceive of a an erroneous scenario of how the Art came to be.
The AI is not an artist. It has not been trained to be an artist. It is not training to be an artist. It has made no aesthetic decisions and it has no intent, let alone an intent to actually create art. It’s not a collection of computer code specifically created to be a virtual artist.
A better description would be something like cybernetic collaborative lucid dreaming between humans and artificially intelligent prosthetic digital representations of human memory. Put more simply, the current crop of AI Art tools work the same way that dreaming and hallucination work, only digitally. A gigantic database of images linked to descriptive words and phrases was used to create virtual visual memories. The AI Art generating code uses user imputed words and phrases to digitally remember or hallucinate from the prosthetic memory of images. Like Lucid Dreaming, the human is never in full control of the dream, but can influence it. The images that are produced only become Art, once a human views them and makes the aesthetic decision that they are Art. Most AI generated images are actually discarded as not art, or at least not good or usable art.
I am not a fan of the new term Synthography either, as Image Synthesis is not limited to AI tools and has already been in use for decades. https://en.m.wikipedia.org/wiki/Rendering_(computer_graphics)
I am sure it’s too late, but my preferred term for the technology would be:
Lucid Synthetic Dream imaging
or
LSDi as an acronym.
The resulting Art itself would simply be a:
Synthetic Dream
FWIW, no names of any Surrealist artists nor the term surrealism were used to generate these images.…
