The Battle Between Art & The Algorithm
28th April 09
Posted in creativity, data, design, interactive
Technology … is a queer thing.
It brings you great gifts with one hand,
and it stabs you in the back with the other.
–C.P. Snow, New York Times, March 15, 1971–
There’s a battle raging, yet it’s almost Truman Show-like in its subtlety. It’s the battle between art and the algorithm. Between emotion and rationality. Between indescribable magic and perfect information.
Jim Carrey, in 'The Truman Show'
As the granular world of relevance, measurability and accountability tightens its grip on the increasingly emaciated flesh of businesses struggling to re-tool quickly enough to survive, many are rushing too quickly away from striving for the magic that has characterized the work we all admire, no matter what the decade or canvas.
As far as I know, no one is trying to kill me. Yet, I sometimes feel a little like the unfortunate hero at the center of the dystopian sci-fi thriller “Minority Report,” John Anderton. The famous mall scene in which Anderton (Tom Cruise), is assaulted by dozens of individually targeted ads — some of which, much to his horror, even loudly broadcast his name as he passes — represents a world a lot closer to ours than the fictional date of 2054.
It’s a world of perfect targeting. Optimization. Zero wastage. Absolute utility. Total accountability.
More and more of what I see, hear, read and even taste seems exceptionally cunningly targeted at me. My RSS feeds me handpicked news streams. I get perfect movie recommendations via Netflix, books I’ll enjoy via Amazon, uncannily relevant advertising when using Gmail, weirdly familiar music from Last fm. Satnav keeps me resolutely on the data-derived optimum track. And so on.
All remarkable stuff. It seems Arthur C. Clarke wasn’t far off when he noted how “any sufficiently advanced technology is indistinguishable from magic.”
What could be possibly be wrong with all this? In this new world where relevance — of information, of entertainment, of advertising, even of new social contacts — is increasing by the atomically measured second, all powered by the extraordinary power of the Almighty Algorithm, what are we losing?
Well, these shifts are triggering a smoothing out in our experiences, prompting a reduction in serendipity and introducing a spooky predictability to many facets of our lives. It’s becoming clear that ultra relevance comes with a hidden price. Because if everything’s relevant, then nothing’s unexpected, and if nothing’s unexpected, then nothing surprises you, and if nothing surprises you, then that’s a strange, neutralized, vanilla kind of life to lead. Think John Anderton meets Truman Burbank.
We’re talking about the end of surprise.
John Stuart Mill, writing in 1836, coined the term “economic man” in painting a picture of someone who was an ultra-rational being. The day Google was born in 1998 could be said to be the birthday of “algorithmic man” or the ultra-relevant being. Some might argue this is an inevitable cultural outcome of the fusion of technology and economics, the creeping onset of what Mill called “perfect information.” Yet, just as economic man didn’t really exist, nor does algorithmic man.
And right there is the opportunity for marketing: to deliver not just relevance, but revelation.
Surprise is the “killer” form of impact, driving engagement, and powering word of mouth. As the world slides towards increasing reverence of relevance, the opportunity is to re-commit to touching people in powerful ways that genuinely surprise them, whether with products, experiences or communications. We must re-find our ability to craft magic, to move people, to deliver the unexpected, never-seen-before experience, to blow minds and touch hearts. That doesn’t mean fighting against the algorithm; on the contrary, it could mean working with it.
How? A tiny amount of work currently does this.
AKQA’s work for Halo 3 is a brilliant example of how great interactive can genuinely move people by reframing something that they’ve seen before (the honoring of heroes) in a surprising way — in this case through alloying personalization and interactivity with emotion.
Just about anything by Jonathan Harris hits this spot on. His “We Feel Fine” from 2005 is an exploration of human emotion on a global scale. It’s a brilliant coming together of art and mathematics, a fusion of art and the algorithm, resulting in a compelling, immersive experience that touches users (still) because it is profound, simple, beautiful and occasionally funny all at the same time.
A shot from one of Aaron Koblin's awesome visualizations (http://sandbox.aaronkoblin.com/)
But too little work has played in this area. Only a handful of businesses of any shape, size or persuasion seem to have succeeded in marrying the two — art and the algorithm, magic and interactivity, surprise and efficiency — in how they operate and in what they produce.
How might this be achieved?
One, creative businesses need to cherish and empower the people who understand technology, consumers’ relationships with it, and brands potential uses for it. Technology must be the catalyst of a new creativity, not just a set of new delivery channels or production options.
Two, creative businesses need to create arranged marriages between these people (the algorithmicists) and the magicians or artists. Only a fusion of these two strands of creativity at the earliest conceivable opportunity in a process will lead to the most adventurous outcomes.
Three, in an era where everything that was once solid does appear to be melting into air, creative businesses need to recommit to surprise as a potent strand of engagement. To paraphrase Ming Yeow Ng, one of the founders of Discover.io, if discovery is the new cocaine, then we’d all be wise to get a whole lot better at dealing in not just relevance, but revelation.
We believe this is the opportunity for the marketing and communications industry. Exploiting the awesome power of the new world of relevance whilst creating more surprising and engaging experiences. In short, then, how can we encourage a collision between art and the algorithm?
Far from having all the answers, we have so much to learn. So more immediately, we’d really appreciate your feedback, ideas, and viewpoints.
Go on, surprise us.
(An earlier version of this post appeared on MediaPost, April 27 2009 - http://bit.ly/aWcwq. One correction has been made since the original post was published: I had incorrectly suggested the Halo 3 interactive work was produced by McCann when it was actually produce by AKQA; this has been amended in this version of the post)
35 comments on “The Battle Between Art & The Algorithm”
or sign in using Facebook Connect
Enter your personal information to the left, or sign in with your Facebook account by clicking on the button below:
Ben Malbon
On Behalf Of Kluster, I offer the following food for thought
The reason that you experience ‘perfect targeting’ is because you have made information about you publicly available to the algorithm through your actions online. If you want the algorithm to surprise you, then you must surprise it with behavior outside the ‘norms’ so that it has more obscure data points from which to extrapolate. It’s not that we are experiencing the end of surprise but rather that we are not as surprising as we used to be. With so many ‘recommended’ paths spelled out for us, we often choose the well-trodden instead of those less travelled. We are social creatures, and following the herd is much less cognitively taxing then setting out into unknown territory.
“What could be possibly be wrong with all this? In this new world where relevance — of information, of entertainment, of advertising, even of new social contacts — is increasing by the atomically measured second, all powered by the extraordinary power of the Almighty Algorithm, what are we losing?”
What we are losing is the strength of the signal to noise ratio. With ‘relevant’ information surrounding us and constantly competing for our attention, it is increasingly difficult to find that which is unique and moving when most of what bombards our senses is repetitive and watered-down.
The problem arises from the control of the algorithm. Nearly all sites that generate recommendations use a single algorithm that figures out what content to serve based on user similarities and mass appeal. Once the bandwagon gets rolling, it is difficult to avoid.
What if, however, we could tweak the algorithm to show us what people opposite of us prefer? What if i could get all recommended movies from people in my demographic who rated Space Chimps a 10. Each filter on the algorithm produces a dimension on the dataset being examined. Each filter presents its own unique view of the content. As you increase the number of filters that can be tweaked in the algorithm, an exponential increase occurs in the number of possible ‘content streams’ that can be served. As you begin to compare and contrast multiple perspectives, triangulation of relevant and unique content becomes easier.
By turning control of the algorithm over to the user, people are then empowered to find content that suits them best. The opportunity for the marketing world lies in empowering people to surprise themselves. This creates a positive feedback loop of engagement which only gets stronger and more diverse as you add more people to the mix.
Creativity is essentially the discovery of hidden connections. The marriage of art and the algorithm occurs when people are empowered to explore the many dimensions of the content they are both creating and consuming.
Bravo.
this is great
[...] The Battle Between Art & The Algorithm Apparently design lost the battle with art and now art has set its sights on the algorithm. Is deciding how something is weighted in a myriad of ways based on rational—or is it art. I think it’s art if design myself if you know what to look for. [...]
Your post reminded me of a quote I once read by JP Morgan:
“There are two reasons why a man buys something: The good reason, and the real reason”
I think we’ve always been hungry for the logical, sane, and quantifiable reason for wanting something, but have always been just as motivated and intrigued by the irrational. It helps explain why Economics is the imperfect science and Psychology has limitless fields of study and interpretation – how we internally rationalize things is often inexplicable.
To me, the internet puts a twist on this age-old tension. I firmly believe that because of this quest to reconcile the “good reason” with the “real reason”, information is addictive, and the internet provides it in spades to anyone with access to a computer. Like any habit, it is hard to break, and we develop a new set of expectations based upon its use and availability.
Your point about taking away the element of surprise (Minority Report) speaks to the removal of “the real reason” we are (perhaps unconsciously) looking for. It is strange for us to have everything functioning, and we may find something off-putting by being targeted as such.
I think the opportunity lies in the fact that we’re constantly searching, and when we find, we are happy. Maybe it is the role of marketers to create that “random find” in a sea of regulation and hyper targeting by way of really cool creative, poignant copy, and innovative use of new media. Rather than tell people exactly what they want, they can offer hints and clues that lead people on their own paths to discovery, using algorithms and tracking to get a sense of what those specific paths are.
Bravo and well said Ben. And Ben your counter argument was great.
Hi Ben.
Very much enjoyed the post.
A quick thought:
Is it not the case that some highly targeted recommendations algorithm serve us content, experiences or products that are not only highly relevant but also have the capacity to surprise and delight us.
In my mind, the spheres of relevance and surprise are not currently mutually exclusive…Sure an Amazon recommendation is not the same serendipitous experience as finding a book yourself through a non-linear journey through the inter-web.
That said, whilst some recommendations are indeed somewhat predictable, I still find that I am frequently surprised by the products recommended to me.
Relevance, in my mind, does still lead to unexpected experiences…
Ha, maybe that means you’re less predictable than me, and no doubt 100 times more interesting. I agree that frequently I feel very well served by amazon, netflix, fresh direct and so on. My point was more about the opportunity for marketing to offset this absolute relevance and provide the surprise, even danger, that seems missing from more and more of our lives, and which I believe we need & in certain cases crave. Cheers for the comments, B.
Great post Ben this is such a hot topic. And Ben K, also love the idea of customisable algorithms with varying degrees of relevance.
A couple of things occur….
As on-line retail gets more and more efficiently relevant, does the role for the off-line retail space become not to compete on this front but to deliver extraordinarily well curated, diverse and serendipitous experiences? Revelation, as you put it.
Does the role of the editor or the curator become increasingly valuable?
I also wonder if social search will help bridge the tension between serendipity and the algorithm. One the one hand, the promise is of ever more relevant and targeted recommendations, on the other those recommendations are drawn from a huge pool of people like me-not just me-so should in theory offer greater freshness and surprise. And i can choose the recommendation of the crowd as a whole or the people most like me-so already I can start to customise the algorithm.
Finally, I always find it hugely comforting to know that a tiny group of films account for a huge proprtion of the difficulty in improving Netfix’ ability to improve their recommendations by that crucial 10%, most famously Napolean Dynamite. It’s nigh on impossible to predict, on past behaviour, whether people will love or hate it. Now that’s magic….
Excellent insight, Ben. But I think you’re touching on only one half of the issue.
This may be passed around marketing departments as a fluff-lined phrase, but it’s also a cultural truth: This is the conversation economy. We’re wired to have a dialogue with our environment – from biomimicry in the natural world to the street art of the urban one. So when one side of the dialog – the input stream – becomes overwhelmingly predictable, the danger isn’t simply that the information we encounter is less likely to surprise us, but it’s also that we lose our own capacity to surprise.
This goes back to Eliot’s point a little bit – the set of expectations we develop inevitably becomes a filter for how we frame ourselves and our understanding of the world, how we relate to our environment.
In the We Feel Fine example, for example, the information and interactions that Harris used should be taken with a grain of salt – even our style of self-expression online, our online persona if you will, has mutated to comply with these expectations of how we OUGHT TO frame ourselves. In a way, our revelation is no longer authentic but a kind of ad hoc, relevance-based self-presentation based on the sort of personal brand we’re trying to build. Did all those expressions of “feelings” really represent an authentic response to a stimulus, or a sentiment based on said set of expectations? I realize this sort of meta-view can be a confusing loop that leads nowhere, but I also think it’s important to ask ourselves these questions as we think about the value and dangers of falling to algorithms to shape how we relate to our cultural environment.
Here’s a “real-life” example of what I’m talking about: Take StumbleUpon – it’s the epitome of relying on an algorithm for the very processes that surprise come from: Curiosity and discovery. Is this wrong? To an extent, because even “discovery” now comes as a database of content regurgitated by the algorithm’s users.
My point here? The antidote to this algorithm-driven robbery of surprise is human-based curiosity. There’s tremendous value – and, above all, art – in what I call “cultural curation.” We need to explore on our own – go down the rabbit hole of lovelinks and reach that fascinating Flickr set by some photographer in Indonesia, that tiny but brilliant art blog by a design student in Amsterdam. And we also need to find the most talented “cultural curators” and turn to them for truly surprising discoveries, for an exploration of human curiosity that no algorithm can offer.
Your post made me think of why Pandora is so powerful. You start with self-selected interests and it constantly surprises you by tossing newness into the mix.
Look at the way Icosystem has taken its Nymbler baby-naming tool through its Hunch Engine technology.
I think we’ve got to find ways to connect ideas and relevant info but in such a way that newness emerges. So, it’s not just matching everything to what I’ve said I like/want… it’s allowing my hunches to take me to new places based upon the interaction of interests, behaviors, and expressed attitudes.
I’m loving the glut of posts about data I’ve been reading and have included this in a round-up of faves from the last few days.
I’m sure we could have worked out a way to do this programmatically but felt it best to allow at least the exciting possibility that I would leave some vital links off my list out of good old fashioned user error - I’ve set a reminder to myself to make a little mistake if I can, perhaps a random typo that will open a link to a new page I never really wanted…
Great post though. And I particularly liked the idea of “turning control of the algorithm over to the user” - YES: I want some dials and stuff to twiddle. I think Last.fm used to let you do that - you could ‘up’ the degree of randomness.. which felt good as a user. I haven’t used Pandora for a while - but when it launched it suffered - in my opinion - from far too much dependence on the music DNA. I remember entering The Wedding Present as a start point. It worked out that I liked twangy guitar music with a strong rhythm and a male vocalist. This matched me to a load of Country music, good old boys playing the banjo.. At the time (Pandora may have got better - I can no longer use it cos I’m in the UK) the social side of the recommendation service probably didn’t have enough users. By contrast, I always felt that Last.fm’s recommendations were better for being ‘more human’. Of course, the mighty algorithm was involved but its job was essentially to extrapolate from the social recommendations of real people. In the past couple of weeks I’ve found Blip.fm’s model - crude though it is - to be quite fun: I tell it a bit about myself up-front, and then add fans to my social filter as I encounter songs that I like. That’s crude, but I feel more in charge. Unfortunately, my initial love affair with Blip has cooled following the ULTIMATE TEST: I wanted to see if they had the Space album by The KLF. They don’t. Nor do Spotify. I’m willing to bet that Pandora don’t (please let me know)… But Last.fm DO.
Awesome!
This is an interesting reframing of the famous CP Snow’s famous essay “The Two Cultures” in which he described the barrier between science and the humanities in the academy. I applaud both your desire to escape the complete rationalization of your desires (After all, some our tastes actually change, over time) and the hope we can make great art of out of the new technology that speaks in “the old high way” to the human heart. It does seem we are entering a cultural moment in which we are looking to art to help us process and understand the explosion of data that is constantly generated and reported back to us. I have endless idealism about the power of great art to accomplish just this feat. But it’s interesting I thought that many of your beautiful examples use a mass of anonymous nodes or memes to create a composite experience, essentially erasing the “subject,” in academic parlance, or rather creating a composite human subject out of an random play of subjective expressions and measurements. “We feel fine” is beautiful and interesting but not sure it totally satisfies. It’s a bit of a trick in the end, reflecting back the mass of data-points in an evocative framework, but I’ve yet to see a great piece of algorithmic art that helped me truly understand where I fit now (a hopelessly solitary subject) in the new sea of data we are all swimming in. Thanks. Great post
[...] ‘battle between art and the algorithm’ is pretty much that. Fascinating article about how the incredible power of the [...]
Hi all. This is a great discussion. some thoughts:
@Dylan Viner
“Is it not the case that some highly targeted recommendations algorithm serve us content, experiences or products that are not only highly relevant but also have the capacity to surprise and delight us.”
indeed. however, surprise is dependent on change. let’s say you’re looking at your netflix recommendations every day. in this time span, they won’t change and quickly will surprise be lost. sample the recommendations every 6 months or 2 years, and then yes, there will be more change, more data to be sampled, more branches in the search space, and thus more potential for surprise and discovery.
one of the elements of surprise is sampling frequency
@Patricia McDonald
“As on-line retail gets more and more efficiently relevant, does the role for the off-line retail space become not to compete on this front but to deliver extraordinarily well curated, diverse and serendipitous experiences? Revelation, as you put it.
Does the role of the editor or the curator become increasingly valuable?”
yes. the editor/curator has to create an experience, a story for the customer to create/discover themselves and others within.
and along this same line of thought in regard to Ben’s original question of how do we marry art and the algorithm to create surprise. that’s where the role of the curator is essential. a curator frames the experience and then gets out of the way.
“Don’t tell people how to do things. Tell them what to do and let them surprise you with their results.” ~ General Patton
[...] dias atrás li um texto muito bom no BBH Labs. Falava sobre a importância de não se preocupar apenas com a relevância das idéias. Faz todo o [...]
[...] in Pedro Qualquer coisa que eu escreva vai soar idiota perto desse texto (apesar da palavra “art” me incomodar) que encontrei no BBH [...]
Great stuff Ben, and something I touched upon recently as well, looking at it from a PR perspective: http://bit.ly/GVWyA
Same general premise, that while measurement (choose your definition) is important, art (choose your definition) is equally important, and as you’ve stated, may be a key differentiator as more agencies have access to the ‘algorithm’.
I look forward to hearing more on this from BBH Labs.
I love your point about the art being the part that will remain impossible to commoditize. Clearly, mastery of the algorithm and of data is by no means easy (not least because it’s evolving so quickly) but I’m especially interested in how data can either inspire or even become art. Hence the marriage proposal I’m making. B
Posting a message I tweeted a while back:
“Just in Paris, saw amazing piece by Hantai, made me think about the ‘art & algorithm’ (he scribed text on canvas for 1 year) http://is.gd/ztce”
It shows that artists have long been interested in the algorithm. It’s great that algorithm-ists are now getting more interested in art.
Agree. Surely data and art are simply two sides of the same thing: life. The algorithm is just ’smart data’. B
Thank you all for sharing your thoughts.
I come from the business of using algorithms to help people find love (Not the most easy algorithm in the world).
If you are positively surprised by a suggestion of Amazon or any other program – then this is not due to accident – then it is a well working algorithm, based on more than the “I like” dimension with access to several information sources – like online activity.
I have to admit that this discussion is very much like religion Vs. Science discussion, which I would bet we all had over dinner table and some beers.
The question in both discussions is how much do we want to “know” based on data and statistic analyses – and how much do we want to leave to chance based on believe.
I am not saying that Religion = Art – but I have not meet one person who can predict Art - if you even can that.
But I have meet enough persons that can with 95% predict a person’s buying behavior – with enough data.
So from a company (Economic) perspective, I very much understand why where they a placing their money.
I am impressed by companies that take art way – but I can promise you, they will never take the same way if does not show some value on the bottom-line. So we can only support the companies that goes that direction.
So going back to the question: Yes, I believe in targeting and in algorithms. I do not want to date 1.000 women to find the right one. I do not want to receive 1.000 Amazon recommendation – I want to receive 2 good ones.
That can happen with giving algorithm access to data – by that I agree with Ben Kaufman and Maria Popava: We have create surprise and changes our self. Take a new path – go down the rabbit hole.
Thanks Martin, an interesting slant on the debate; always good to drag people like us out of our tiny little world of brands and marketing. That said, I guess what’s so intriguing about where we are now in marketing is that the immense influence of the algorithm is forcing us to completely re-appraise the prominence and nature of ‘the art’ through which we have historically delivered messages to people. I think it makes it that ‘art’ more exciting, and more powerful. I think it’s leading to a new creativity.
Great post.
All I hope is that the internet gets used as more than just a canvas. We’ve already got enough surfaces to paint on.
What we create online isn’t fine art, it’s architecture, it’s engineering. What we need is back-end beauty. Look at any great feat of engineering or product design, or achitecture and art is inherent at every level - it’s the form, the experience, the originality.
Great artisitic web content already exists today, but we need to appreciate it for what it is.
It’s not what you hang on the walls, it’s the walls you hang it on.
Thanks for your thoughts Copyblighter. Could not agree more.
What’s clear (even to the sceptics) is that the internet is much more than ‘another canvas’. It’s a completely game-changing paradigm shift in how we communicate, how we interact, how we entertain ourselves, how we create value, how we do business . . . even how we leave traces of ourselves on this planet when we eventually depart.
Which is why it’s so utterly cool.
B
True. The web is cool. It’s an infinite playground.
But it’s also causing an insatiable hunger for immediate gratification, for change, for the new. Which makes enjoyment an increasingly transient experience, practically devoid of anticipation.
Because there’s more to see than can be seen, we have to pass judgment in an instant. And, even when we do stick around long enough to appreciate something, how long until the next big thing lands?
My point is this: when we make this many connections, can any of them be truly meaningful?
In the past, surprise was driven by one content/stuff/places/sites that made us stop and stare. It was singular, something consumed and forgotten. Today, surprise has taken on a different shape, and is being driven by ongoing experiences. It is participative, shared (and sharable) and spread out across platforms. In this new world, surprise is about delivering a journey, not just a destination.
Overall, this conversation reminds me of something my grandfather used to say: “You can please some of the people all of the time, all of the people some of the time, but you can’t please all of the people all of the time.” I guess the best we can hope for is to please enough of the people enough of the time…
Remains a fascinating topic of debate. Am reminded of Joshua Porter’s excellent question, “how do you programme for serendipity?”
Two things I think will be critical in “marrying” art and the algorithm are:
1) the increased portability of the social graph increasing our ability to deliver recommendations anywhere based on friends’ and friends-of-friends’ preferences
2) Data mash-ups: combining our netflix preferences with our Amazon choices with our grocery shop to deliver increasingly sophisticated recommendations-targeted but with nuance and breadth
I’m sure it’s been widely read by now but an excellent article in Wired this month highlighting precisely this tension: Google vs. Facebook=the algorithm vs. the social graph http://bit.ly/F5yLF
[...] connecting it, how does differentiation happen, other than the obvious product possibilities? This very interesting article (via @vijaysankaran) discusses the battle between art and algorithm. Amidst the quest for perfect [...]
This is all something we can cheer for. But almost everywhere I have been managers and companies become terrified when coming face to face with someone with genuine creativity and vision. Most people are risk adverse. Artists take risks.
you might be interested in this: [http://faithinthealgorithm.net]
Especially the first article of the ongoing series: Faith in the Algorithm, Part 1: Beyond the Turing Test [http://arxiv.org/abs/0903.0200]
Alberto, these papers look fascinating. Have downloaded and look forward to reading. Thanks for the recommendation.
Nice introduction but deceiving examples… How could you describe such a beautiful insights with such poor examples…
I agree with the core but not with the edge
Hi Jean. How kind of you to comment so constructively, but sorry to disappoint so comprehensively. Please share your examples so when we revise the article we can include them. They sound as though they must be very interesting. Cheers.
[...] The Battle Between Art & The Algorithm 34 comment(s) | 9574 view(s) [...]