The Archival Look
Last week, I came across two images that epitomize the two trajectories of our new AI-generated images culture. The first is a video of an anthropomorphic egg schizo-morphing into a self-cannibalizing demon complete with humanoid teeth. It is, as they say, nightmare fuel.
If anything corroborates Hayao Miyazaki’s position on artificial intelligence—that it is "an insult to life itself"—it’s probably this genre of AI visual: a Slick Horror. For some reason machine intelligence is particularly adept at creating imagery that maintains a hyperreal iPhone sheen while adding unsettling, impossible elements: convoluted six-fingered hands, melting distorted faces, gaping orifices, mutant teeth.
Software is intuiting our collective anxiety about the dentist. See another example below:
Love it or hate it, Slick Horror is probably the most novel AI visual aesthetic. But is it the most compelling? Is the newest kind of image a new technology produces necessarily its most important? I would argue no.
First off, the schizoid-melting-hallucination already feels dated. Diffusion models like DALL-E have improved substantially since their commercial release in 2021. Slick Horror is already an artifact of the recent past.
Secondly, the genre-specific nature of the aesthetic limits its application. Like Surrealism before it, the genre precedes the aesthetic. Surrealist art, fashion, architecture, and film have more in common with each other than their medium-specific brethren. So too with horror.
But let’s play devil’s advocate. It’s popular to presume cultural innovation is based entirely on material (i.e. technological) inputs. Developments like abstract painting are second order effects of inventions like photography. You get Bauhaus architecture and the International Style from steel beams and concrete. New ways of making things, so the thinking goes, is always the subtext of cultural innovation. Ergo Slick Horror. You can only create something like the video above with AI.
This reminded me of a prescient quote by sci-fi author and critic Bruce Sterling from his 2012 Wired “Essay on the New Aesthetic”:
Valorizing machine-generated imagery is like valorizing the unconscious mind. Like Surrealist imagery, it is cool, weird, provocative, suggestive, otherworldly; but it is also impoverished.
The essay is a response by Sterling to the British critic and artist, James Bridle and his (then) recently hyped SXSW panel concerning the New Aesthetic. The panel argued for integrating computer vision and the inherent attributes of screens into art and design. The New Aesthetic in Bridles words was an attempt to “rupture the interface between the digital and the physical.”
Like Clement Greenberg’s approach to the picture plane in his 1960 essay “Modernist Painting”—the New Aesthetic championed flatness as the fundamental attribute of the screen. Remember when Jonny Ive banished skeuomorphism with iOS 7 in 2013? That was the New Aesthetic.
This made sense when Photoshop was the most common digital image-making tool and the ever-improving photographic software on every smartphone fundamentally changed the image culture of the era. From this, we got DTC marketing—color blocked and hi-res. From this, we got corporate Memphis cartoons. From this, we got minimalist logos, with every conceivable brand giving themselves a sans-serif makeover. (The argument for this being sans-serif fonts are more legible on small mobile screens.) When you start to look for the twin trajectories of simplification and flattening, you see it everywhere last decade.
And then it stopped.
THE ARCHIVAL LOOK
AI is the dominant image-making technology of the future. And while the Uncanny Valley effect of non-human intelligence creating unimaginable horrors may excite some—there is a broader shift at play here. This occurred to me when I saw the second image I mentioned. Find it below:
Created by AI artist Stephan Vasement, the visual is interesting because it sits somewhere between an illustration and a photograph. It feels highly processed, despite the fact that there is no original image to be processed. Its aesthetic is identifiable but only vaguely. It sits somewhere between a heavy metal album cover, 80s fashion photography, and a movie poster for a swords-and-sorcery epic.
Flat and slick is being replaced with archival graininess. Photorealism is being modulated with the illustrative. The Archival Look is exceptional artifice, an attempt to rescue aura from techno-sterility. It’s the aesthetic equivalent of digging through your flatscreen’s options menu to turn off ‘auto-smoothing’—that annoying preset that makes everything you watch look like Masterpiece Theatre.
The use of the word archival here is intentional. I’m borrowing it from the fashion world’s vocabulary. Archival in that context is the fulfillment of a certain idea of luxury where brand imbues a quasi-mystical power to a garment. One feels as if one is buying a real thing. The luxury object doesn’t lose its shine, but rather increases in curiosity and reverence over time. In short, it attains an aura.
Unlike Slick Horror, the Archival Look is already creeping into mass culture. Last Sunday’s episode of Industry introduced noise and grain to the show; a departure from the sheen of the first and second seasons. Billie Eilish’s music video “Lunch” recreates the blown-out saturation of Pulp’s “Like a Friend". Perhaps, I’ll make a compendium of examples.
Some may see this and think it’s all style over substance—just another class of faux-vintage filter, Valencia for Zoomers. But what is most interesting about AI visuals is that they are all style and no substance. I highly doubt “Lunch” was created using Stable Diffusion. I’m not arguing that. It may not even have been filmed on a retrotech digital video camera, even though it looks like it was. No matter.
The Archival Look takes “all style no substance” as our baseline understanding of images moving forward. And it’s from style—not the implements used to derive it—that meaning and aura are obtained.
