Rhea Myers

Complaint in the Age of its Operationalization

OpenAI’s social media announcement of its new image model led to a wave of copycat “ghibliizing” of images, named for their imitation of the house style of animator Hayao Miyazaki’s animation studio. Elon Musk’s Whitehouse took only 24 hours to use that model to re-render a photographic record of its rapidly escalating human rights abuses in cute cel animation style. It’s as if someone hopped down on Ketamine had instructed an LLM to precis Walter Benjamin’s “The Work of Art in the Age of its Technological Reproducibility” then treat its warnings as a manifesto for producing kitsch fascist propagada. The results are aesthetically, ethically, and politically revolting.

They are also cringe. They are cringe because everything Musk does is cringe. They are yet another product of a needy, try-hard, middle-aged manchild nerd’s desparation for love and attention, to be one of the popular kids, even and especially as a de facto dictator. If the gap between power and the expression of desire defines cringe, Musk is the Black-Scholes-being-hit-by-negative-prices of cringe.

The rise of a priori dismissal of all AI output as “slop” is the IUD of though-terminating cliches. It stops thinking about AI before it even starts. The evaluation is “slop”, now what was the subject? We can invoke Benjamni’s essay within this dismissal, too. AI slop very obviously lacks Benjamin’s aura. It isn’t even mechanically reproduced, it is mechanically produced. Digital pink slime.

This doesn’t get us very far, though. Digital art is always(-already, if we must) mechanically produced however the cursor is moved. Decrying AI prompting as unskilled in comparison to selecting a brush in Procreate says nothing about AI prompting other than that the person decrying it doesn’t know what they are talking about. Prompting is an art – it is a craft with a theory. You can get bad results in any medium if you don’t know what you’re doing. But seeking to misrepresent those results to people for clicks isn’t informing them, it is exploiting and cementing their ignorance.

Ghibliized images are encountered in this way as a Schrodingerian superimposition of art that is both so good that it can displace a master artist while also being so bereft of that artist’s authentic aura that they are simply cringe. They are slop. Beautiful slop, but slop. And therefore ugly. The torsion on the concept of “slop” is more than a newly confected dismissal can bear. And yet it persists.

Nelson Goodman’s “Languages of Art” is a dry 1970s work of Analytic Philosophy. Among other ideas he works through in the book, Goodman contrasts art that is produced with and without a written score in a way that usefully complements Benjamin’s consideration of mechanical reproducibility. Productive though that is, it’s another of Goodman’s ideas from “Languages of Art” that concerns us here.

When considering the problem of fakes in art, Goodman uses examples of forged paintings being revealed in order to argue that we cannot know which features of an artwork will affect its authenticity in the future. These aren’t a matter of chemical or radiological analysis of images, although these developments have certainly revealed an increasing number of fakes in recent years. Rather it is a matter of looking at the artwork and considering it in a different light.

The features of fakes that later reveal them as such are not noticed when they are first mistakenly accepted as authentic. But when someone points them out, they become obvious. It takes time and effort to establish how to differentiate a fake from an authentic artwork attributed to the same artist. But once the difference can be seen, it cannot be unseen. This is a trapdoor function. It can destroy fortunes and reputations. All based on a shift in the consensus understanding of which sets of perceptual properties underwrite the authenticity of a particular ouvre.

Does ghibliized slop really resemble cels from Miyazaki’s animations? It does to the degree that an image model can detect in early 2025. Reducing Miyazaki’s work to the aesthetics of a pre-CGI line-art style excludes much of the value of his art, though, even if it could be emulated in a way such that there was no remainder for future connouseurs to discover and exploit in order to exclude it from the category of the authentic based on its structure rather than an a priori dismissal.

We are not entitled to live off the epistemic passive income from our investment in a romantic understanding of human uniqueness. It is a failure of critical imagination to simply object to a product’s fulfilment of the limited terms chosen for its initial promotion. It is like a cat chasing a red dot on the floor and feeling pleased with themself when they catch it. While screaming at anyone who points out that the dot is coming from somewhere, and that lasers have other more interesting uses.

It is cringe.