When IBM’s Deep Blue chess computer defeated the world champion Garry Kasparov in 1997, humanity let out a collective sigh, recognising the loss of an essential human territory to the onslaught of thinking machines. Chess, that inscrutably challenging game, with more possible game states than there are atoms in the Universe, was no longer a canvas for individual human achievement. Newsweek called it ‘The Brain’s Last Stand’.
Why was the loss so upsetting to so many? Not because chess is complicated, per se – calculating differential equations is complicated, and we are happy to cede the work to computers – but because chess is creative. We talk about the personality, the aesthetics of chess greats such as Kasparov and Bobby Fischer, seeing a ‘style of play’ in the manipulation of pieces on a grid. Chess was a foil, a plane of endeavour, for storytellers as diverse as Vladimir Nabokov and Satyajit Ray, and we celebrate its grandmasters as remarkable synthesisers of logic and creativity. It was particularly galling, then, for Kasparov to lose to a machine based not on its creativity but its efficiency at analysing billions of possible moves. Deep Blue wasn’t really intelligent at all, but it was very good at avoiding mistakes in chess. One might argue that its victory not only knocked humanity down a peg but demonstrated that chess itself is not, or does not have to be, the aesthetic space we imagined it.
And yet Kasparov, after having lost to what he later called ‘a $10 million alarm clock’, continued to play against machines, and to reflect on the consequences of computation for the game of kings. And not just against them: for the past two decades, Kasparov has been exploring an idea he calls ‘Advanced Chess’, where humans collaborate with computer chess programs against other hybrid teams, sometimes called ‘Centaurs’. The humans maintain strategic control of the game while automating the memorisation and basic calculation on which great chess depends. As Kasparov described an early such match:
Having a computer partner also meant never having to worry about making a tactical blunder. The computer could project the consequences of each move we considered, pointing out possible outcomes and countermoves we might otherwise have missed. With that taken care of for us, we could concentrate on strategic planning instead of spending so much time on calculations. Human creativity was even more paramount under these conditions.
Kasparov argues that the introduction of machine intelligence to chess did not diminish but enhanced the aesthetics of the game, creating a new space for creativity at the game’s highest levels. Today, players of ‘freestyle’ chess work with high-end chess systems, databases of millions of games and moves, and often other human collaborators too. Freestyle teams can easily defeat both top grandmasters and chess programs, and some of the best centaur teams are made up of amateur players who have created better processes for combining human and machine intelligence.
These centaur games are beautiful. The quality of play is higher, the noise of simple human errors reduced, making space for the kind of pure contest that the platonic solids and geometries of chess idealise.
We are all centaurs now, our aesthetics continuously enhanced by computation. Every photograph I take on my smartphone is silently improved by algorithms the second after I take it. Every document autocorrected, every digital file optimised. Musicians complain about the death of competence in the wake of Auto-Tune , just as they did in the wake of the synthesiser in the 1970s. It is difficult to think of a medium where creative practice has not been thoroughly transformed by computation and an attendant series of optimisations. The most profound changes have occurred in fields such as photography, where the technical knowledge required to produce competent photographs has been almost entirely eclipsed by creative automation. Even the immediacy of live performance gets bracketed by code through social media and the screens we watch while recording events that transpire right before our eyes.
In fact, the shift might be much more profound for the audience than the artist. Being a critic or consumer of art now relies on a deep web of computational filters and guides, from the Google and Wikipedia searches we use to learn about the world to the recommendation systems queueing up books, songs and movies for us. We rely on computational systems for our essential aesthetic vocabulary, learning what is good and beautiful through a prism of five-star rating systems and social-media endorsements, all closely watched over by algorithmic critics of loving grace. We shape our aesthetic expectations around these feedback loops, finding channels and lists that seem to match our interests and then following them. Google has already introduced a system that proposes responses to your emails based on millions of prior conversations, and the company Narrative Science has been creating algorithmically generated journalism for years.
Today, we experience art in collaboration with these algorithms. How can we disentangle the book critic, say, from the highly personalised algorithms managing her notes, communications, browsing history and filtered feeds on Facebook and Instagram? She exemplifies what philosophers call the extended mind, meaning that her memories, thoughts and perceptions extend beyond her body to algorithmically mediated objects, databases and networks. Without this externalised thinking apparatus, she is not the same critic she would be otherwise. This is true not just in pragmatic terms, in that she might not be nearly as good or efficient at her work, but in biophysical terms as well. Our brains adapt to the tools that we use, like London taxicab drivers with their enlarged hippocampi for mapping the city’s convoluted streets (or the GPS-dependent smartphone users whose shrivelled hippocampi can’t even navigate them around their own neighbourhoods).
Our entire cultural horizon is shaped by the filtering mechanisms of our news feeds, inboxes and search results
The extended mind is now also becoming a space of collective cognition. The critical network of literary reception made up of that critic, the author she writes about, both their friends and followers, the hashtags, links and cross-references that bind these nodes together, all form a much more inclusive tapestry of cultural discourse than was ever possible before. We depend on our friends and social networks to tell us what to think about new creative works, and that process of assessment and sharing depends on algorithmic filters designed to maximise attention, traffic and profits.
Sociologists such as Pierre Bourdieu painstakingly mapped the deeply social dimensions of cultural judgment in the 20th century, but today the deeply intersubjective nature of taste is not just obvious but almost subliminal. Algorithms are shaping the reception of works at the forefront, but also the periphery. The entire horizon of our cultural perspectives is shaped by the filtering mechanisms that populate our news feeds, prioritise our inboxes and rank our search results. And they are, of course, built out of our own collective responses to prior stimuli, modelling a collective aesthetic project that we (often unknowingly) participate in with every click and purchase.
The immediate creative consequence of this sea change is that we are building more technical competence into our tools. It is getting harder to take a really terrible digital photograph, and in correlation the average quality of photographs is rising. From automated essay critiques to algorithms that advise people on fashion errors and coordinating outfits, computation is changing aesthetics. When every art has its Auto-Tune, how will we distinguish great beauty from an increasingly perfect average?
The visceral reaction is to rebel against these simulations of progress and perfection. Many artists today explore the seams and rough edges of digital platforms, creating art out of the glitches and unintended juxtapositions that they can eke out of increasingly complicated creative systems. A few years ago, the American science-fiction author Bruce Sterling helped to champion the idea of the ‘new aesthetic’, predicated on the ‘eruption of the digital into the physical’. The art form celebrates the gritty, avant-garde edge of pixelated shear, and explores the inherent tensions between digital representations of culture and the consequences those representations have when plucked out of two-dimensional server-farm splendour rendered back into richer cultural forms. Carefully crafted physical representations of 8-bit graphic art, giant physical pins emulating the skeuomorphic place-markers in digital map software: these efforts to see like machines are also clearly rebellions against the aesthetic constraints of computation.
Lurking behind these efforts to disrupt the normal functioning of computational culture is a deeper creative need. What we crave most in art, what we reward more than anything else, is surprise. Marcel Duchamp’s urinal, the introduction of perspective to landscape painting, stream-of-consciousness literature – these creative breakthroughs achieve much of their impact by shocking us into some new perspective on the world. Little wonder that the modernist poets were so fascinated by the metaphors of blasts and explosions, or that art has such a long and complicated history with warfare. We need art to surprise us in order to blow up the world, to create fissures out of which the new can emerge.
Computation is not good at this. Algorithms are wonderful for extrapolating from past information, but they still lag behind human creativity when it comes to radical, interesting leaps. So far, they are much better at identifying and replicating surprising content than they are at producing it themselves. Platforms such as Facebook or Flickr’s ‘interestingness’ quotient ultimately measures a kind of surprise, one that draws on information theory as well as aesthetics. We respond to viral memes on social media because they produce something unexpected, often leveraging the deep relationship between surprise and humour. It is telling that so many memes now hide their linguistic tells (more tractable to algorithmic watchdogs than images) inside GIFs and JPEGs that circulate in a kind of shadow economy of surprise.
More often than not, it is a human creator who bends computational tools to achieve a breakthrough
Surprise will remain a human territory, at least for the short term, because it is so idiosyncratic in the first place. Our sense of the unpredictable is so oddly tuned that true randomness can sometimes seem too regular, too predictable, like a long string of coin tosses where the same side comes up many times. At the same time, we are quite choosy about the kinds of novelty that count, a form of distinction that could, in the end, be precisely what we mean by aesthetics. How many art critiques and book reviews boil down to the judgment ‘this is a predictable extrapolation’? Newness is necessary but not sufficient for human surprise. There is a cadence, a significance that we seek in the aesthetics of surprise that reaches deeper than mere randomness. As pattern-seeking animals, we are looking not just for comprehensible behaviours but for signs and portents – stories about the world that allow us to configure reality according to an aesthetic logic.
And that aesthetic, the deeply ingrained logic that the poet John Keats framed in his ‘Ode on a Grecian Urn’ (1819) as ‘beauty is truth, truth beauty’, provides a space for humanity in the age of culture machines. There is a future for human aesthetics in the modulation, the casting of surprise. We will continue responding most powerfully to those creative stimuli that somehow reconfigure our brains, literally allowing us to see in a new way. Machines can occasionally do that for us, like David Cope’s beguiling, algorithmically composed classical music (which disturbed so many audiences by emulating masters such as Mozart entirely too well). But more often than not, it is a human creator who bends computational tools to achieve a breakthrough that somehow becomes more than a recombination or incremental advancement of prior work.
Inevitably, the generation of those surprising works will grow more dependent on computational tools, following in the steps of every other field and industry. And so we come to the second major opportunity for human creativity in the face of increasingly intelligent, competent and aesthetically capable machines. In order to survive, but more importantly to thrive, in the age of algorithms, we need to cultivate a deep respect for algorithmic literacy and the capacity to ‘read’ the impact of computational influences on our work – not necessarily to resist those influences, but to understand them and use them to become better humans.
To follow this argument, we have to contemplate how computation is changing our fundamental cultural grammars of action. Our ubiquitous smartphones, sensors and platforms are more than just new nouns on the stage of cultural practice. They are generating new verbs and grammatical relationships, many of them so obvious that we no longer even pause to contemplate the godlike powers encoded in the phrase ‘to google’ something. As the media theorist Lev Manovich’s work on Instagram suggests, the general dissemination of cellphone cameras goes beyond making photography more accessible; it fundamentally changes what photography means. We are starting to perceive the world through computational filters, perhaps even organising our lives around the perfect selfie or defining our aesthetic worth around the endorsements of computationally mediated ‘friends’.
In other words, digital platforms are imposing exponential shifts in creative practice and possibility. It is possible, sometimes even trivial, to make genetic code sing or translate pitch into colour. Perceived as patches, programs, little nests of scripting, databases and sensors, such things are just demonstrations, methods or practices. But these tools are, first of all, endlessly recombinant, allowing stunning concatenations of creative cause and effect. Second, and more importantly, they depend on the same universal ideology of computation that drives Silicon Valley’s continued expansion into cultural life. According to this logic, it’s all data, from DNA to footstep counters, from high-frequency stock trading to videoblogging. And while that makes more aspects of cultural life accessible to algorithms, it also makes all the same aspects tractable for art.
As creators and appreciators of the arts, we would do well to remember all the things that Google does not know
Artists can now, in a very real way, make art out of the stock market or the emanations of our smartphone radio antennae. Art that is surprising, and new, and computational. Art that reverses the almost gravitational force currently sucking agency, money and meaning out of 20th-century industries and redistributing them to a small technological elite. For that reason alone, the emancipatory power of art is vitally important as we come to terms with the deep consequences of cultural computation. Surprise might get us to the door but it is after our neurons are reconfigured, after we can see what we could not before, that really interesting things begin to happen.
The role of art in creating new forms of legibility, of literacy, will become crucial for humans attempting to swim in the ocean of computation, that vast deep we can only appreciate through metaphor, analogy or creative interpretation. At times, it seems like the only way we can really come to terms with the full consequences of algorithms is when they play in our cultural spaces, like Google’s DeepMind machine-learning algorithm in its path of conquest from Atari videogames to the mythically creative game of Go, or its equally mesmerising and psychedelic image-processing cousin, Deep Dream.
Human creativity has always been a response to the immense strangeness of reality, and now its subject has evolved, as reality becomes increasingly codeterminate, and intermingled, with computation. If that statement seems extreme, consider the extent to which our fundamental perceptions of reality – from research in the physical sciences to finance to the little screens we constantly interject between ourselves in the world – have changed what it means to live, to feel, to know. As creators and appreciators of the arts, we would do well to remember all the things that Google does not know.
The creative response to computation needs to pursue both of these avenues – the human capacity and receptivity for surprise; the need for new creative computational literacies – to contend with the sea change we are living through. It’s a challenge that goes beyond coming to terms with algorithms, and one that might become a question of survival on an intellectual level, if not a physical one.
The remarkable precipice we stand beside now is one where our tools are, in a transformative way, just as plastic as we are. Our algorithmic systems are watching us, learning from us, just as we learn from them, creating the possibility for a complex dance of intention, anticipation, creativity and emergence based on individual people, algorithms, and the social and technical structures that bracket them all. This is terrifying and breathtaking all at once, and it’s artists that we need most of all to make sense of a future in which our collaborators are strange mirror machines of ourselves. Aesthetics has always been the unforgiving terrain where we assess pragmatic reality according to the impossible standards of the world as we wish it would be. Computation is a parallel project, grounded in the impossible beauty of abstract mathematics and symbolic systems. As they come together, we need to remain the creators, and not the creations, of our beautiful machines.
‘What Algorithms Want: Imagination in the Age of Computing’ by Ed Finn is out now via MIT.