Jenna Sutela: Mars in a dish

Contemplating nimiia cétiï for New Scientist, 11 September 2018

The artist Jenna Sutela normally divides her time between London and Helsinki, but she has spent the last four months at London’s Somerset House Studios, thanks to a residency with Google Arts & Culture. Here, she’s been either making a video, learning about computers, teaching artificial intelligences to dream, or mastering Martian. Perhaps all of the above. It depends who you speak to.

The artists Google invites to explore the potentials of its machine learning systems normally wind up in its lab in Paris. This time, however, Google engineer Damien Henry (the co-inventor, incidentally of the Google Cardboard VR headset) has been travelling to London to assist Sutela and her artistic mentor, the Turkish-born data artist Memo Akten, in a project that, the more you learn about it, resembles an alchemical operation more than a work of art.

Here – so far as I understand it – is the recipe.

  1. Take one nineteenth-century French medium, Hélène Smith, who made much of her communications with Martians. (The Surrealists lapped this stuff up: they dubbed her “the muse of automatic writing”.) Make up some phonemes to match her Martian lettering. Speak Martian.
  2. Prepare a dish of Bacillus subtilis, a bacterium that we expect would cope rather well with conditions on Mars. Point a camera at it, and direct the video signal through a machine learning system. (Don’t call this an AI, whatever you do. Atkin has issued a public warning that “every time someone personifies this stuff, every time someone talks about ‘the AI’, a kitten is strangled.”)
  3. Lie to your AI. Tell it that your dish of wiggling bacteria is in fact a musical score. Record the music your AI makes as it tries to read the dish. Hide kittens.
  4. Keep lying. Tell it your dish of wiggling bacteria is a text.
  5. – a language.
  6. – a map.
  7. Write down the text. Speak the language. Read the map. Put the whole enterprise into a single twelve-minute video and hang it up in the foyer of London’s Somerset House Studios.

Titled nimiia cétiï and on view in Somerset House Studios until 15 September, Sutela’s video installation is heavy-going at first, but well worth some close scrutiny. Everything you see and hear came from that petri dish: the landscape, the music, the alien script, even its eerily convincing Martian vocalisation. “There is,” Henry tells me with avuncular pride, “absolutely no scientific goal to this project whatsoever.”

The point being that Sutela is one of the first artists, if not the first, to appropriate the rules of machine learning entirely to her own ends. It’s a milestone of sorts. She’s not illustrating an idea, or demonstrating some technical capability. She’s using machine learning like a brush, to conjure up imaginary worlds.

Which is to take nothing away from nimiia cétiï‘s considerable technical achievement. Sutela is forcing her recurrent neural network to over-interpret its little petri dish-shaped world. We’re a long way from inventing a machine that sees pictures in a fire, but these results are certainly suggestive.

«e tesi leca rizini nirnemea riechee sat ze po mizi» as a Martian might say.

Phantom Menaces

written for the Arcfinity blog:

Google’s swooshy new concept video for augmented reality goggles (or “spex”, if you will) has certainly put the virtual cat among the digital pigeons. An attempt, perhaps, to leapfrog the iPad – if Google can persuade us that what we really want is headwear that will let us see things that aren’t really there.

[youtube http://www.youtube.com/watch?v=9c6W4CCU9M4?rel=0]

I recently spent an entire evening doing just that. Aurasma, a start-up spun out of Autonomy (another search giant, incidentally), aims to bring AR to the masses;
that evening, its glamorous representatives pasted digital magic over a south London gallery’s functional white-and-grey surfaces.

Rice packets came to life in our hands to show us how to cook rice. Books spilled their letters into our laps; they took wing and flocked about our heads like so many starlings. Avatars swung their swords blindly about the gallery.

AR is one of a handful of technologies that are likely to transform our lives in the very near future. And I don’t use the ‘T’ word lightly.

People talk about the great things AR can show you. Every wall becomes a picture! Every picture becomes a movie! Every object becomes something other, something better than itself – or seems to.

Oddly nobody talks about AR’s ability to hide things. And since I’d been invited along in the role of Ancient Mariner (stopping one of three with tales of future horror – I am a writer after all; this is my job) it was this ability to subtract from the visual richness of the real which interested me the most.

Never mind the avatars and the rice-packets: these are distractions, no different in kind to movies, posters, fiercely rung handbells, and all the other manifold calls to our attention. Let’s get back to basics here. What does it mean to look at the world through a screen?

The granularity of the world is always going to be finer than the granularity of the medium through which we perceive it. No matter how photorealistic AR gets, it will always be taking information out of the picture plane.

So AR has the potential to render the world down to a kind of tedious photographic grammar – the kind employed by commercial image libraries, whose job it is to reduce the world to a series of unambiguous stills illustrating stock ideas like ‘busy at work’ or ‘looking after the children’.

This is nothing new. Photography has the ability to do this, obviously. But photography cannot be stuck over (or in) your eyeballs twenty-four hours a day.

AR has this potential, substituting the real with a simplified description for anyone wearing the funny glasses.

Lacanian psychoanalysts have a word for this process of simplification: they call it repression. And if AR becomes truly ubiquitous, then we will no longer be able to trust our eyes, and will probably have to develop a neurotic relationship with this technology.

Is AR a good thing, then, or a bad thing? This is the kind of question we’re trying to avoid in Arc. Not because it is a hard question, but because it is a bad question. It assumes we have no agency, no wit, no common sense. It assumes we’re at the mercy of our own technology.

The problems thrown up by AR will not be new. They will be old. They will be fairytale-like problems. (Is that woman by the bar a fairy, a queen, or a crone? Is the wizard touting for trade in the shopping centre a wizard at all, or a mere trickster? Is our Prime Minister really wearing any clothes?) It may be that, in order to navigate this fairytale visual space, AR will give birth to an entirely new set of visual behaviours.

It wouldn’t be the first time. Look at reading. There is nothing ‘natural’ – certainly nothing evolved – about it. But we welcome its effects, and we build upon them, and we celebrate them. In some ways reading makes us less -it’s been shooting human memory in the foot repeatedly since Plato’s day. In other ways it makes us more: it
allows us to share the knowledge and experiences of people we will never meet, of people who have ceased to exist, of people who never existed. AR will do the same.