I have gotten fascinated by AI art, which has taken major strides forward in 2022. These images were created by NightCafé using the Stable algorithm, in response to the prompt "typewriter" combined with the names of some artists. (Can you guess which images invoke Hieronymus Bosch, Norman Rockwell, and Gustave Doré?)
This kind of technology has some civilization-destroying potential which I explored in my story "You'll Be Needing This," in Paradigm Shifts. But as we await the robo-apocalypse, we can enjoy these images and reflect on an aesthetic revolution in the making.
Are these images "art"? Not if you define art as something fully under the control of a human artist—but that definition has been questioned in the art world for over a century. These pictures, spat out mindlessly by a computer in just a few seconds, undeniably stimulate feelings, suggest possibilities, and engage the imagination.
For the human user, "making" AI art is completely different from making a conventional artwork. You feed the computer a text prompt and experiment to see what it will generate, which is often surprising. It is both empowering and humbling.
Nobody really understands exactly what happens when the computer creates a particular image, but the general principle is simple: it works with a huge database of pictures that have been labeled by human beings, calculates probabilities, and generates a new picture using those probabilities. Our brains must do something similar when we draw on our cultural familiarity and our vast store of experiences to create something new, yet never completely unprecedented.
Exploring AI art is not primarily a matter of researching computers; it's a dive into human collective memory. The database of images that the computers have at their disposal is the result of billions of human observations and concepts, which can now be explored in a new way.
Will AI art (and text and video) really destroy civilization? Maybe. But what I predict is that AI is going to play a major part in a transformation of human culture that has fascinating potential. I'm glad to be alive to see the beginning of it. I hope to be alive to watch the first complete motion picture that has been generated, from start to finish, by computers. It is sure to be an eerie and thought-provoking experience.
I think we are losing sight of reality.
ReplyDeleteThat's one of the first things I tried to do with Dall-E! See this link for a few of them. https://photos.app.goo.gl/epH6W19aXuTaVVFk7
ReplyDeleteAt this early point in development, I'm curious to know: What if you keep repeatedly feeding in the exact same prompt? How much variation will the AI come up with? If there is variation and you continue long enough, will a cycle turn up?
ReplyDeleteFunny, these remind me of the weird and wonderful typewriters in Cronenberg's 'Naked Lunch' movie...
ReplyDelete