The Horror of AI Art
--
We asked the machines to learn from us, and it produced cosmic horror. What does that say about us?
As a content creator, I’m delighted by AI. I have access to limitless art without copyright. I can create eye-catching images for my Medium posts, beautiful music to play behind my YouTube videos, evocative covers for my books, even the poems that might proceed each chapter, all without the meddlesome IP that might legally haunt me some day, and without paying one cent to a human being.
Therein lies the irony. Is this the end of the gig economy? What use will we have for artists in our content driven age? How long before AI masters prose, assembles a factual news story (or a fake one), a screenplay, or a novel?
That already happened. Knowing this, or, more precisely, now knowing that you didn’t know this, makes you doubt everything you read.
First we lost the gatekeepers. Now we’ve lost the humans altogether.
“Every story tells a story that has already been told” — Umberto Eco, 1984
Of course there is nothing new under the sun. After 400 generations, anything we feel has been felt before.
And that’s why AI works.
AI art is derivative, which makes it less than original. You feed the computer a few words, select an established art style, and give it an image to build on. The emotion you feel is derivative, too, trained to appreciate or emote based on hundreds of years of artistic endeavor. If you feel serenity in the AI’s Impressionist landscape, you have Monet and Degas to thank, not the computer. The human hand still lies behind that random generator.
That may be the true horror. AI pulls from everything we’ve made, an internet full of images. Why, then, do the results so often horrify us? What does that say about us, and the vision of the world we’ve presented?
For this image on the Dream app, I typed in “World War 2” and overlaid a photo of myself in a cave. The result: a soldier stands forlorn on a field of devastation as a dive bomber appears…
