Archive for December, 2022

Art and Bitterness

December 15, 2022 6 comments

If I were making gallery art I would make a robot that goes around art galleries rating the paintings and an installation of an empty room with 2 monitors in it, one showing machine-generated art produced continuously as quickly as possible, the other looking at the first machine’s product and categorizing it by style, movement, and quality.

That’s all I’ve got to say about the AI art debate, really. There’s some theory behind it, which follows, but if you get it, you can move on to the next blog.


Still here?

  1. there’s some anthropological theory about communication. First, art is communication – that’s clear enough, right? And communication implies at least two communicants – a speaker and a listener. Both of these are active in any act of communication: an utterance (“speech act” in the disciplinary jargon) is not complete until it’s been received. In fact its meaning is not fixed except by the receiver – think of a gift (as Christmas looms). Is it a good gift? Appropriate to the person and occasion? You only know when the receiver smiles. Or a joke – does it serve its purpose? Only the audience can tell you. All speech acts are like this.

    Thus, also, art. Which brings me to a common misunderstanding about AI art: people fixate on the part the machine does, but we humans remain in a partnership with the machine because we select what we think is good out of its products. All the AI art produced so far gets subjected to this process of selection or reception. Without a receiver, it’s not even art, as such. Much like a bicycle is not a mode of transport without someone to ride it.
  2. the craftsmanship has already sailed. When Walter Benjamin philosophized about how industrial production kills the specialness inherent in individual objects, he was observing a process that had already been at work for a century. And he feared that art, as maybe the last bastion of the human trace in the stuff we interact with, was being rendered obsolete by photography (an art based almost entirely on reception, btw). The camera could make images instantly, without even requiring a human operator. And Art (as in gallery art) responded by shifting its value structure, in a way the general public still hasn’t caught up with… about a century ago. Gallery art has spent that century wrestling, more or less angrily or amusingly, with the obsolescence of craftsmanship, until now it’s old hat and there’s simply no agreement about what makes an artwork good. So (as people often state without fully articulating their question) we’re mostly left wondering why the gallery owner/curator chose to show us this rather than that. We’re focused on the receiver, that person being someone other than ourselves.

    Benjamin worried that machines outperform human craftspeople in producing most of the things we use. The one thing human artists had over machines (apart from a quasi-religious value in being human) was the quality of their thought. But AI art promises to replace that in short order – not because the machines are smart but because the receivers of their products either can’t tell the difference or will tend to prefer the machines’ production, while continuing to ignore/discount their own role as selectors. And the more they select, the more the machines will learn what pleases them.
  3. Benjamin’s fixation on the aura of the individual crafted object is hard even to imagine these days: art is already ubiquitous and easy to find, provided you’re happy with its translation into pixels. The internet offers us the world’s biggest library, at our fingertips. If you were thinking of making something, chances are it already exists. You’d just have to search to find it.

    Borges wrote a story – The Library of Babel – about this situation, and about how it might actually be easier to write your own book than to search an almost infinite library for the extant one – even with our modern search tools it’s most likely to be hidden way down in the cheap end of the search results. But AI art promises (perhaps) to short-cut that search. It will make something maybe close enough to what you wanted – close enough that the extra value in making exactly what you want (if, in fact, you are able to achieve that) is a frivolous luxury.

So my modest proposal is, if you’re going to replace the hand, why not also the eye? AI can throw out images at a rate no human viewer can keep up with – the logical next step is to replace the viewer.

But why would you do that? Why even go to the expense and effort of building this machine that does not need a human at any stage? The machine doesn’t need it, either.

That’s why this is a blog post, not a gallery installation. I don’t need to build it in order to question the continuing value of making more art (which has been actively in question for a century).

…so is art – or Art – dead? No, I don’t suppose so. Gallery Art continues to fulfill social and economic functions for those people rich enough to deal in it and it has been demonstrating that it doesn’t even need authors for decades now. As for popular art – the stuff you can retweet from google image search or DeviantArt, well…

I expect it to go through a process like vanillin. Back in the 17th century (when commissioning a portrait told everyone you were important) vanillin commanded a king’s ransom. You could only get it from orchids in Mexico. Louis XIV dazzled his mistresses with it. Then manual pollination allowed it to be transplanted around the world and the middle class could afford it as a treat. Then in the late 19th century someone figured out how to extract it from pine wood and it became a bi-product of the paper industry – one of the cheapest commodities known to man. And now people use the word “vanilla” as a vague insult denoting something with no special qualities, too ubiquitous to be valued. And more vanillin is consumed, in more contexts, than ever before.

Art, like vanilla, is already everywhere. Commercial art is already unwelcome noise in our urban landscape. And there are efforts afoot to make everything a projection surface or, better, an oled screen, capable of surrounding us with movies 24/7. It takes a lot for an image to break through that noise.

Maybe, eventually, we will have less of it. Maybe machines can finally get us there.


I needed to read a few things to write this minor rant. Those things contain much more interesting insights. So I recommend:

Signs of Recognition, by Webb Keane. An anthro study of how people tell their listeners when they really need to listen. Also contains a thorough examination of the speech act process.

Fictions, by Jorge Luis Borges, which includes the dazzling stories The Library of Babel and The Lottery of Babylon, both of which have implications in a thousand astonishing directions.

The Work of Art in the Age of Mechanical Reproduction, by Walter Benjamin. Which seems foolish at first but then he keeps going and it’s… really good in the end after you’ve thought about it for a decade or so.