notes from the past, present, and post-future home

Perverse Affordances - Interview with Alejandro Ball

December 31, 2019

Alejandro Ball, curator of the Extreme Views series of online exhibitions at Art Gene’s The Digital U, invited me to exhibit Perverse Affordances on their platform in fall 2018. The exhibition page is still online. Ale also interviewed me as part of the project, and for various reasons, mostly how long I took to answer the questions, it never got posted on the Art Gene site. Sharing them here now, in the spirit of shamelessly dumping old writing into the public sphere.


Can you please talk about your thought process and influences while developing Perverse Affordances?

Definitely. There’s a lot of interest in the art community around GANs, or Generative Adversarial Networks, likely because artists are also image makers. As artists, GANs are maybe our best collaborator and biggest competitor - the biggest existential threat to image-makers since cameras themselves. And of course the images have a surreal, uncanny beauty. But the majority of the GAN-generated images I’ve seen use photos and standard datasets. I would say the piece really started with the desire to see a fundamentally different kind of GAN image.

And secondly, I had the strong feeling that just making images with a GAN wasn’t enough. To do that would be to fetishize tools - and I think making work about tools is important and interesting, but it has to peer a step deeper and investigate what the tools actually are. Really, what they afford. I think one of the most striking qualities of a GAN-generated image is its inhumanity. It sees with a profoundly alien eye. GAN-generated images confront the strangeness of perception itself. So I turned one on social media networks, these presumed-friendly interfaces that mediate most of our relationships, in an effort to make strange the familiar, and make inhuman the social.

This was a many-stage piece and I consider it still ongoing. The first step, collecting the dataset, was actually the most difficult. I used a browser-automation framework called Selenium to crawl popular social media sites and randomly take screenshots.

But, more than just generating the dataset, I wanted to be somewhat scientific about my process. So I worked from a list of the most popular social media sites in the world and tried to include as many of them as possible. Of course, it turns out that a number of them are not english speaking, and in fact a solid six of them are from China. A few can only be signed up for with a phone that has a chinese country code. As part of my research, I ended up hiring a Chinese translator to help make the accounts, and she even let me scrape her weichat account in the browser! Eventually, though this was a tedious process, I ended up with about 10000 images from 10 different platforms. Ironically, training the neural net was easy in comparison.

Once I had the images, I used them like wireframes to create an interactive sketch of what this social network could be. Phase two of this work will be to make it live online, I’d like to build it out so that users can sign up, upload photos, and interact in the uncanny valley of earthly interfaces.

A main concept behind Perverse Affordances is our relationship (people in general) with computer interfaces, e.g. Facebook or Twitter’s interface, how do you feel this components effect, shape, and dictate our actions while navigating the Web?

They are subtle and endless. On a platform, is “friending” unidirectional, or always mutual? On facebook friendship is mutual but on twitter it’s one-sided. I would argue that this makes twitter an inherently more competitive environment. Back when I studied tumblr, I thought its choice of hiding the friend count was particularly important. It allows any tumblr, suddenly discovered, to become an ‘authority’.

How many instagram users have deliberately taken a picture with extra space at the top or bottom so it will crop well to a square? How many people, when making irl physical gestures with their hands are now imitating the ok sign emoji, or the raised hands, or the thumbs up? These platforms effect our actions not only as we navigate the web, but as we navigate our lives in general. Design is never innocent - indeed, it’s the layer at which technology intersects with ethics.

Bill Gate’s recently spoke about the distinction of a platform in contrast to contemporary social media platform, claiming that Facebook and Twitter are less of a platform than Windows OS is, and are more akin to older historical services like America Online. What are your thoughts on this - and how do you believe the interfaces of social media sites contribute to this illusionary effect of obscuring a websites purpose, and or, intention?

If social media networks were spaces, Facebook and Twitter would be suburban gardens. Windows might be city block. Computing is a wilderness. This said, I can’t say I have strong opinions about what is and isn’t a ‘platform’. I am no platform purist.

The ‘purpose’ of all social media platforms (and also most operating systems) is to maintain our attention and collect data about us. This is something that’s becoming increasingly obvious, despite them being in the business of obscuring it. We’re living in the age of totalitarian memory - and beyond surveillance capitalism - there’s something horrible to this kind of permanence, in and of itself. Blockchains and other immutatable distributed datastores make it even more literal. I wonder sometimes if any of us could bear to face the totality of our own data? It’s like imagining a sublime enormity. Confronting your true self in 2019, vipassana with your facebook history.

One of my favourite ironies of Perverse Affordances is that, while engaging with this practice of trying to ‘see’ the invisibleness of interfaces, machine learning algorithms at big social media companies are also ‘seeing’ us all the time (Maybe this is the real ‘abyss that gazes back’). Anyway, the actual process of generating the dataset for Perverse Affordances involved me hiding - by way of generating so many random clicks, far far more random clicks than real ones, that end up being useless noise in someone else’s machine learning dataset. The data about me that Google has is certainly now far worse, after this project. Maybe this kind of thing is like the graffiti of data markets. Productive sabotage. Let’s hope nothing my selenium agent clicked on gets associated with terrorism.

Artificial Intelligence (AI) has captured the imagination of our culture, whether its through art, media, science fiction, or recently current events - all of which picture AI in a negative light. However in my opinion, Perverse Affordances uses AI in a different way; more like a critical analytical tool (or maybe better said, a collaborator). What is your take on the present environment and uses for AI, in contrast to your own use of the technology - and could you also elaborate a little on your thoughts about AI’s cultural persona?

I’m not sure AI is helped by it’s personality, so much as its further mystified. The reality is that right now what we like to call “artificial intelligence” is only “intelligent” when if comes to solving specific types of problems. It’s algebra performed over large datasets, and it is only so good as the data it eats. One could use this to make an argument to be made for data sovereignty, and possibly data markets (if our data allows your ai to be better at some valuable task, our data is valuable). But then, if we start selling our own data, it bears considering that privacy may become a luxury for the wealthy.

But even in start up land, the term “ai” is fuzzy. Besides what we also call machine learning, ai is also the algorithms that guide the npcs in most games. Ai is a marketing gimmick. Ai is a euphemism for mechanical turk. What is “artificial” about this intelligence? Programmers just like me wrote it. It’s made of the data of people just like us both.

As an artist and a computer programmer, your practice is in stark contrast to what some may term as ”art”, could you take us through your practice, both artistic and professional, and tell us how you find synergy between the two?

Definitely. I understand there’s this perception - seldom actually voiced - that art is not technical and computer programming is not creative. Neither of these things are true. Even artists in more traditional mediums like painting and sculpture often end up learning things best categorized as chemistry or materials science. And of course it takes creativity to dig into really crunchy computer science problems. There’s beauty there too, beauty is inescapable (sadly) - an algorithm can be as beautiful as you like.

To me these things don’t feel disconnected - why would they? Software is a medium like any other.

To finish off it would be really nice to hear about your other activities, developing projects, and where you feel your practice will go from here?

That’s a great question. At the moment, I’m contributing to CirclesUBI, which is a democratically-operated blockchain universal basic income. I don’t think it’s enough to make art, or to hold up a mirror and criticize the world. You also have to try and build things that will solve the problems you see. I’ll also be in residence at Trust, “a space for platform design and utopian conspiracy”.

Besides that, I’m going to spend as much of the summer as possible camping and letting the universe blow me around like a dandelion.