Merzmensch – Visual Criticism (2023)
The creative power of generative AI models is fascinating. Since 2016 (with Google Deep Dream) I have been exploring the aesthetics shown and stories told by the machines in essays and my books on AI Art.
In 2020 one of my main projects started. It consists of two parts: “MERZmory” and “reMERZ“. In “MERZmory”, I trained various GenAI models on my own photographs and let the AI reconstruct my visual memories. In “reMERZ”, I fed a language model with essays, poems and stories – new worlds of texts emerged.
There are gems of computational wisdom hidden among these materials. Reading them I came across upon this interesting sentence that the AI wrote to me:
"I was in love with the craft of visual criticism long before the term poetry became fashionable."
At first glance, this statement sounds rather absurd. But only at first glance. In fact, visual stories are older than textual ones. The cave paintings of Lascaux, dated parietal art (around 17,000 years old), the oldest art in human history, were created before Proto-Indo-European languages (around 5,000 BC).
This statement inspired me to use it in different contexts, including as a prompt for diffusion models. The semantics of this sentence seem to trigger a latent aesthetic in the Gen AI, so that it creates new and unusual stories every time.
One of the images with this prompt was this Dadaist collage-like hallucination:
Max Ernst, talking about his collages, also used the word ‘hallucination’, claiming that in this case different semantics and themes are brought together to create new realms.
AI hallucinations, which for me are the most fascinating computational phenomena, are no different – but in this case the machine itself is trying to interpret what it is creating. It fails, of course, but for me this failure is a brilliant creativity of the machine.
To go one step further and bring the image to life with different AI approaches, I still used the same prompt – for describing the movements and action, and also for composing the music (using an older AI model).
And so, after countless experiments, in the year 2023, the following journey into the underworld of the machine was created.
We are contemplating a monochrome metamorphosis of media, a fusion of aesthetics with man, who is both viewer and creator. This is a story about the endless search for new ways of self-expression, locked in the ontological cage of algorithms.
And that journey is just beginning.
You can find more of my works here.