I gave a talk at the Moral Machines symposium in Helsinki last year, and just heard that a revised version of the talk will be published in an anthology tentatively titled The Ethos of Digital Environments: Technology, Literary Theory and Philosophy. The anthology is edited by Hanna-Riikka Roine and Susanna Lindberg and will be published by Routledge, presumably in 2021 or 2022. Here is an excerpt from my draft of the chapter, where I explore the idea that there might not be that much difference between a neural network that can predict when a human would cry and that involuntary tightness we humans sometimes feel in our chests when we watch a sad movie.
Emotions are often conceived as the determining difference between humans and machines, and indeed, between groups of humans and whatever or whoever they wish to define as non-human. “They don’t have the same feelings we do,” the narrator imagines the wives thinking of the handmaids in Margaret Atwood’s novel (1986, 215); “they don’t seem to feel anything, no pleasure, no pain”, the Terrans remark of the indigenous people they rape and beat in Ursula le Guin’s The Word for World is Forest (1972, 18).Continue Reading →
My latest paper, “Situated data analysis: a new method for analysing encoded power relationships in social media“, started out as an analysis of the data visualisations in Strava, but ended up as something more ambitious: a method that I think can be used to analyse all kinds of digital platform using personal data in different contexts. Here is a video teaser explaining the key points of situated data analysis:
This paper has been a long time in the works, and started off as part of the INDVIL project on data visualisations, where I was tasked with thinking about the epistemology of data visualisations. Working through revision after revision of my analyses of data visualisations in Strava I found that what really interested me about Strava was the many different ways that the personal data it collects from runners and cyclists are presented—or, more precisely, how the data are situated. Once I’d analysed the different ways the Strava data was situated, I realised that the same method could be applied to any social media platform, app or digital infrastructure that uses personal data. So I decided to change the focus of the paper so it was about the method, not just about Strava.
Donna Haraway coined the term situated knowledges in 1988 to demonstrate that knowledge can never be objective, that it is impossible to see the world (or anything) from a neutral, external standpoint. Haraway calls this fiction of objectivity “the god trick,” a fantasy that omniscience is possible.
With Facebook and Google Earth and smart homes and smartphones vastly more data is collected about us and our behaviour than when Haraway wrote about situated knowledge. The god trick as it occurs when big data are involved has been given many names by researchers of digital media: Anthony McCosker and Rowan Wilken write about the fantasy of “total knowledge” in data visualisations, José van Dijck warns against an uncritical, almost religious “dataism“: a belief that human behaviour can be quantified, and Lisa Gitelman points out that “Raw Data” is an Oxymoron in her anthology on the digital humanities. There is also an extensive body of work on algorithmic bias analysing how machine learning using immense datasets is not objective but reinforces biases in the data sets and inherent in the code itself (there are heaps of references to this in the paper itself if you’re curious!).
Situated data analysis provides us with a method for analysing how data is always situated, always partial and biased. In my paper I use Strava as an example, but let’s look at a different kind of data: how about selfies?Continue Reading →
Look, this is the oldest known mirror, reflecting the face of a woman holding it. It is 8000 years old and made from polished obsidian.
I’m working on a book on machine vision, and I want to edit it all enough before summer that I can send it off for feedback. It is so hard to keep just editing though when I keep discovering these new fascinating facts! I had no idea that mirrors have been around for 8000 years. Or that crystal rock was used 4500 years ago to create lenses for eyes for Egyptian statues that are remarkably anatomically correct, at least given possible knowledge of anatomy at the time.Continue Reading →
For my book on machine vision I’m writing a little about Vertov’s wonderful 1924 manifesto written half in the voice of the camera – “I am kino-eye … I, a machine, show you the world as only I can see it.”
I started wondering how involved the other Kinoks were in writing. It’s hard to tell, but Elisaveta Svilova, shown here, was definitely very involved in Vertov’s movies, as the editor who actually made all those fast transitions possible and as co-director later on. Here is an amazing meta-moment near the end of a documentary Kino-Pravda reel where she is shown selecting negatives for the very newsreel she is editing.
The camera shows her, then the negatives, back and forwards, showing the negatives in negative and finishing with an negative image of her face. Lilya Kaganovsky describes this as “a fusion of object and subject” and writes that Svilova looks directly at us. And yet it seems to me that in that last image it is the film itself somehow that is looking at Svilova?Continue Reading →
I have a fabulous research environment right now, and while obviously some of that is due to having had funding to employ brilliant researchers (thank you ERC) we’ve been doing a lot of other things that are working out really well, some formal (weekly research group meetings and a Digital Humanities Network with lunch meetings a few times a semester) and some informal. I’ve realized that developing a research environment that is good for us and makes researchers happy is one of my top priorities, and so I want to think more systematically about how to do it. I’m sure some researchers are quite happy alone in their dens, and that’s certainly been the model in the humanities, but a lot of us like collaboration and actually talking about our research with each other. So I asked people on Twitter what works for them, and here is the list I have compiled, from my own experience and from Twitter. We are doing some of this here at Digital Culture, but can certainly keep working at trying more of these tips!Continue Reading →
I spent some of last week at a wonderful larp (live action roleplaying) camp for kids run by Tidsreiser, and had a wonderful time. I have secretly wanted to try larping since I was a teenager, but there weren’t any local ones, then I didn’t dare try, and then I sort of forgot and just settled into being a boring grownup. Luckily, one of the advantages of having kids is you get to try out new stuff. So after a year of sitting around watching the kids battling and sneaking around the forest with their latex swords, and dropping them off at the Nordic Wizarding Academy (Trolldomsakademiet), I’ve started joining in a bit, and I absolutely love it.
After chatting with the fascinating game masters and larpwriters at last week’s camp, and trying out some more different kinds of larp there, I started thinking about what a great tool larping could be for teaching and research dissemination – perhaps especially in subjects like digital culture, or for our research on the cultural implications of machine vision, because one of our main goals is to think through ethical dilemmas – what kind of technologies do we want? What kinds of consequences could these technologies have? What might they lead to? A well-designed larp could give participants a rich opportunity to act out situations that require them to make choices about or experience various consequences of technology use. This post gathers some of my initial ideas about how to do that, and some links to other larps about technology people have told me about.Continue Reading →
Readings: Understanding Video Games Chapter 5, Woke Gaming Chapter 6 (Kristin Bezio: The Perpetual Crusade: Rise of the Tomb Raider, Religious Extremism, and the Problem of Empire. (p 119-138))
Learning goals: After doing the reading, taking the quiz and attending the class, students can
- Explain how video game aesthetics incorporate game mechanics as well as visuals, sounds, etc.
- Use some of the terms in Understanding Video Games chapter 5 to describe games
- Explain Said’s concept of orientalism and discuss it in relation to video games
One of our goals in MACHINE VISION is to analyse how machine vision is represented in art, stories, games and popular culture. A really common trope is showing machine vision as hostile and as dangerous to humans. Machine vision is used as an effective visual metaphor for something alien that threatens us.
My eight-year-old and I watched Ralph Breaks the Internet last weekend. I found it surprisingly satisfying – I had been expecting something inane like that emoji movie, but the story was quite engaging, with an excellent exploration of the bad effects of neediness in friendships. But my research brain switched on in the computer virus scene , towards the end of the movie, because we see “through the eyes of the virus”. Here is a shot of the virus, depicted as a dark swooshing creature with a single red eye:
And here you see the camera switch to what the virus sees. It is an “insecurity virus”, that scans for “insecurities” (such as Vanellope’s anxious glitching and Ralph’s fear of losing Vanellope) and replicates them.Continue Reading →
Last week I was in London, where I visited Pierre Huyghe’s exhibition Uumwelt at the Serpentine Gallery. You walk in, and there are flies in the air, flies and a large screen showing images flickering past, fast. The images are generated by a neural network and are reconstructions of images humans have looked at, but that the neural network hasn’t had direct access to – they are generated based on brainwave activity in the human subjects.
The images flicker past in bursts, fast fast fast fast fast slow fast fast fast, again and again, never resting. Jason Farago describes the rhythm as the machine’s “endless frantic attempts to render human thoughts into visual form”, and frantic describes it well, but it’s a nonhuman frantic, a mechanical frantic that doesn’t seem harried. It’s systematic, mechanical, but never resting, never quite sure of itself but trying again and again. I think (though I’m not sure) that this is an artefact of the fMRI scanning or the processing of the neural network that Huyghe has chosen to retain, rather than something Huyghe has introduced.
Huyghe uses technology from Yukiyasu Kamitani’s lab at Kyoto University. A gif Kamitani posted to Twitter gives a glimpse into how the system uses existing photographs as starting points for figuring out what the fMRI data might mean – the images that flicker by on the right hand side sometimes have background features like grass or a horizon line that is not present in the left image (the image shown to the human). Here is a YouTube version of the gif he tweeted:
The images and even the flickering rhythms of the Kamitani Lab video are really quite close to Huyghe’s Uumwelt. At the exhibition I thought perhaps the artist had added a lot to the images, used filters or altered colours or something, but I think he actually just left the images pretty much as the neural network generated them. Here’s a short video from one of the other large screens in Uumwelt – there were several rooms in the exhibition, each with a large screen and flies. Sections of paint on the walls of the gallery were sanded down to show layers of old paint, leaving large patterns that at first glance looked like mould.
The neural network Kamitani’s lab uses has a training set of images (photographs of owls and tigers and beaches and so on) which have been viewed by humans who were hooked up to fMRI, so the system knows the patterns of brain activity that are associated with each of the training images. Then a human is shown a new image that the system doesn’t already know, and the system tries to figure out what that image looks like by combining features of the images it knows produce similar brain activity. Or to be more precise, “The reconstruction algorithm starts from a random image and iteratively optimize the pixel values so that the DNN [DNN=deep neural network] features of the input image become similar to those decoded from brain activity across multiple DNN layers” (Shen et.al. 2017) Looking at the lab’s video and at Uumwelt, I suspect the neural network has seen a lot of photos of puppy dogs.
I’ve read a few of the Kamitani Lab’s papers, and as far as I’ve seen, they don’t really discuss how they conceive of vision in their research. I mean, what exactly does the brain activity correspond to? Yes, when we look at an image, our brain reacts in ways that deep neural networks can use as data to reconstruct an image that has some similarities with the image we looked at. But when we look at an image, is our brain really reacting to the pixels? Or are we instead imagining a puppy dog or an owl or whatever? I would imagine that if I look at an image of somebody I love my brain activity will be rather different than if I look at an image of somebody I hate. How would Kamitani’s team deal with that? Is that data even visual?
Kamitani’s lab also tried just asking people to imagine an image they had previously been shown. To help them remember the image, they were “asked to relate words and visual images so that they can remember visual images from word cues” (Shen et.al. 2017). As you can see below, it’s pretty hard to tell the difference between a subject’s remembered swan or aeroplane and their remembered swan or aeroplane. I wonder if they were really remembering the image at all, or just thinking of the concept or thing itself.
Uumwelt means “environment” or “world around us” in German, though Huyghe has given it an extra u at the start, in what Farago calls a “stutter” that matches the rhythms of the videos, though I had thought of it as more of a negator, an “un-environment”. Huyghe is known for his environmental art, where elements of the installation work together in an ecosystem, and of course the introduction of flies to Uumwelt is a way of combining the organic with the machine. Sensors detect the movements of the flies, as well as temperature and other data that relates to the movement of humans and flies through the gallery, and this influences the display of images. The docent I spoke with said she hadn’t noticed any difference in the speed or kinds of images displayed, but that the videos seemed to move from screen to screen, or a new set of videos that hadn’t been shown for a while would pop up from time to time. The exact nature of the interaction wasn’t clear. Perhaps the concept is more important than the actuality of it.
The flies apparently are born and die within the gallery, living their short lives entirely within the artwork. They are fed by the people working at the gallery, and appear as happy as flies usually appear, clearly attracted to the light of the videos.
Dead flies are scattered on the floors. They have no agency in this Uumwelt. At least none that affects the machines.
I’ve given a few more versions of the “algorithms and society” talks from this spring. You can still see the videos of those talks, but here are a few links to new material I’ve woven into them:
Social credit in China – this story by the Australian Broadcasting Company paints a vivid picture of what it might be like to live with this system. It’s hard to know exactly what is currently fact and what is conjecture.
Ray Serrato’s Twitter thread about YouTube recommending fake news about Chemnitz,and the New York Times article detailing the issue.