Last night I went along to the Photographers’ Gallery to see an artist who is working with neural network models to create images: the work Mario Klingemann was discussing is available most comprehensively on his Tumblr account.
The talk was extremely relevant to everything I have been thinking about recently, also to many of the themes I touched on in the A4 Critical Essay (especially in relation to flickering signifiers and digital culture) and some of the collaborative ideas I have for A5.
When I first read about it, I must admit I misunderstood – I thought he was taking data extracted from human neural networks and asking a programme to make images with them. What that might achieve I do not know…Not sure it is even possible. Photographer, Chris Friel, was mixing audio data with images a while ago, somehow, so perhaps I thought Klingemann was doing something along those lines. But I very much doubt that data would lead to actual pictures of faces, no matter how grotesque but maybe they would – maybe I should try it!
In fact, he takes the sort of computer models which learns to recognise objects and faces, trains them and then gets them to produce images with the knowledge they have acquired. His powerful computer can generate in the region of 5000 images a night and then he sifts through them. This type of tech is the sort you can find at Open Source – a community of developers who make their software available to anyone can use or better it. Notably, when I searched for more information, the first article I noticed from Open Source pointed out that empathy was one of the key concerns for software developers at the moment, along with how these developers can earn a living (same issue everywhere – things in the information economy cost very little to make and are easy to copy and reproduce.) Concern about the erasure of empathy is everywhere too at the moment. This week’s New Scientist mag has a big article in which they refer to ‘techlash’ – backing away from technology and understanding that something needs to be done to address some of its less positive outcomes.
“IT IS a bit too early to start taking bets on 2018’s word of the year, but “techlash” is surely worth a punt. In recent weeks, growing disgruntlement with giant technology companies has hardened into a full-blown backlash. The prevailing mood is that Amazon, Apple, Facebook and Google have grown too big for their boots (see “How Google and Facebook hooked us – and how to break the habit“).
The complaints against the big four are many and varied, from enabling the dissemination of fake news to monopolistic behaviour and rampant hawking of user-generated data. These are genuine concerns that need to be addressed. But, as yet, they do not add up to much more than an incoherent howl of frustration.” (Lead article, 2018)
Several things emerged for me in terms of what I have been thinking about throughout this module.
This work is not about the aesthetic quality of someone’s work. Yes, he chooses the images that he likes, which express something of his subjective taste, and I will discuss more this further along. He does, however, discuss the ersatz quality in them. But the aesthetics seem like a side to the main course, which is his process – and the questions of collaboration as Klingemann allows his technology to do its thing.
I have been thinking about aesthetics in art a lot lately. To put it bluntly, aesthetics alone can at times be desperately boring. That’s not to say I don’t see things which are sublime, astonishingly beautiful, awe-inspiring. I recall reading about a photographer who said photography is boring – and I’m afraid agreed with him. I wish I could recall who it was. He wasn’t saying all photographic projects were boring. But he was saying photography itself can be. I think perhaps one reason (there are others) for this is aesthetics have been completely and utterly taken over by advertisers and mass production. Roy Ascott eloquently sums up how art has changed and is continuing to change, “The emphasis by the artist moves from content to context, from object to process, from representation of the world as a given, to the construction of worlds in emergence, from certainty to contingency, from composition and resolution to complexity and emergence. In short our focus has shifted from the behaviour of forms to forms of behaviour.” (2002:p2) (my italics) Advertisers and production methods combine to make representation a given – or at any rate, representation is incredibly easy to achieve with a click of a few buttons or a swipe of a thumb over a screen. Aura is perhaps achieved now more through context rather than content. Stan Dickenson’s work for BoW was a process and the striking patterns that emerged from it were almost ‘by the by’ – any computer can generate such patterns and print them in high quantity within moments on an IKEA or Habitat picnic set. It was his extraordinarily clever context which contributed to the work being so effective and successful.
Klingemann’s work and his ambition to create a model which no longer needs his input at all makes us question our relationship with technology. What emerges is important and whether or not we ‘like’ the results matters too – but not in the same way we as it might if we were looking at a highly accomplished or expressive pencil drawing. Just as Bryan Eccleshall’s Digital Rain seems more about quantity, digital distribution and the merging of signifiers from different parts of the digital spectacle; news, social media, history and art – modern and not so modern, copied digitally but originally from different mediums and production methods etc…
The discussion about aesthetics in art has been going for some decades but at a time when it is possible to create an image in moments on our phone and upload it seconds later to a platform drowning in images this conversation seems more important than ever.
What does emerge from Klingemann’s models is strange, and in his words ‘creepy’. It’s not for nothing he calls himself Quasimodo on his Instagram feed. Catherine Banks and I chatted about this creepy aspect of some digital entities and images recently. We humans recognise when something isn’t right. A genuine psychopath can trigger nausea in a relatively healthy person (I have experienced this and it’s a very unpleasant feeling). The video of Snapchat combinations I made as an A3 experiment had the same effect. Quite a lot of Klingemann’s work is on the edge of that feeling. He suggests the models ‘aren’t there yet’ – they don’t quite know how to create something real enough to be thought of as beautiful although some of the artifacts they generate may appeal to certain tastes. And so his work is a clear example of a form of behaviour and the results make us feel uncomfortable in the same way as a psychopath might. Klingemann is working without the primary objective of creating something beautiful or even striking to look at – although he does choose images which must pass a certain subjective taste test. (But it is not usual documentary where the application of ‘beauty and artistic taste’ is at times morally and ethically questionable.) Even so, his images might nevertheless be described as an analogy for the times we live in. Afterall, the internet is described as a monster in the latest edition of New Scientist;
““We’re no longer talking about harmless search algorithms. These companies permeate fundamental aspects of our lives,” says Olaf Groth of Hult International Business School in San Francisco. “These companies often don’t really know what they’ve built,” says internet policy adviser Anri van der Spuy at Research ICT Africa in South Africa. “It’s sort of a Frankenstein problem.”
And the monster looks unstoppable” (Heaven, 2018) So although
All the way through the talk I was wondering, is this work Klingemann’s exteriority or the models’? I came away being fairly certain it was his own and an expression which also represents our society at the moment just as Giacometti’s tall thin twisted sculptures represented the one he lived through. So although the work might be described as primarily concerned with process, the balance between content and context is more tricky to ascertain. Advertising is the place where content seems to dominate, where accepted norms in relation to aesthetics are easily recognised. Advertising extends into most people’s minds and communicates ‘art’, ‘pretty photograph’ ‘beautiful’. And this work is far removed from all of that. I also found it helpful to think about analogy and will discuss this in relation to my own plans for A5 elsewhere, no doubt. I’m incredibly glad I went along.
Ascott, R, 2002. The Grand Convergence: art, technology, consciousness in a planetary perspective, Ecole normale supérieure, Paris https://www.academia.edu/4972226/The_Grand_Convergence_art_technology_consciousness_in_a_planetary_perspective