Return to site

The novel market form of Surveillance,

discussion with Shoshana Zuboff


This article presents the viewpoint of Shoshana Zuboff, author and academic at the Harvard Business School. Her research interest includes social transformation, history of work, technology and capitalism, and also information civilization. This year, Shoshana publishes a new work exploring a novel market form and a specific logic of capitalist accumulation that she named Surveillance Capitalism. She talks about the model of Google and Facebook, with a new, different and very interesting perspective. In what follows, I've tried to script the author's discussion with Leo Laporte and the promotion of her book "The Age of Surveillance Capitalism : The Fight for a Human Future at the New Frontier of Power". The full interview is available here on Triangulation 380.

Leo : Shoshana you are talking to technologists in this book, to people immersed in the internet, technologies, in all of this. But I want to step back and first understand, Shoshana, what Surveillance Capitalism is ?

SZ In some ways, Surveillance Capitalism radically breaks with the centuries-long traditions of capitalism but in other ways there are things that we can understand connected to the history of capitalism. For example, it's long been understood that capitalism evolves by taking things that live outside of the market, that do not have a price, there are not commodities, and finding ways to drag them into the market sphere, so that they are abble to be commodified and sold. So the great historian of the rise of the self-regulating market economy Karl Polanyi in the middle of the 20th century wrote a book called "The great transformation" and he described what he called the three key fictional commodities that allowed industrial capitalism to root and flourish : the first was the idea that you could take human activity that lived outside the market and bring it into the market and call it labor and you could price it and that was the beginning of the ability to have a new structure of wage and salaries labor and so forth. That was the first big invention. The second was the idea of nature that you could take the land and the trees and the lake I'm looking out at, and you could bring into the market and call it real estate or land and you can buy and sell it. The third is the whole idea of exchange which subordinated to the market now became money.
So fast forward to the 21th century Surveillance Capitalism here was the big discovery that allowed Surveillance Capitalism to take human experience, private human experience, you could drag it into the marketplace and call it behavioral data and then with some applications of value added here, and computations there, you buy it, sell it, you create whole new markets based on not just selling behavioral data but selling the computational inferences and predictive analyses from behavioral data that allow interested business customers to lay bets on what we will do, now, soon and later. That's the essence of surveillance capitalism.

Leo : We think of Henry Ford as the inventor of Industrial Capitalism. You've just described a new capitalism based on our private human experience but who came up with Surveillance Capitalism, where did it comes from ?

SZ Great question. I write about the invention of Surveillance Capitalism (SC) at Google in 2001 and your are right: the correct way to think about that is SC is something that was fine line between being stumbled into and being invented which is pretty much how the engineers at the Ford Motor Company felt about the "invention of mass production" where at the end of the day that engineers looked up from moving assembly line and they looked at each other and there was a sense of recognition that they had just broken through to a new kind of economic logic of high volume low unit cost. This logic had not existed before. Similarly at Google we see a situation where in the teeth of financial crisis in the dot-com bust, Google had what most people agreed as the best search engine, they had the snazziest venture capitalists, glamorous and highest status venture investing in their dream but here was the dot-com bust and the reason to believe that Google was gonna go down along with a lot of other small companies in the Silicon Valley and up until that time Larry Page and Sergei Brin had been very vocal and public about their abhorrent of advertising. They regarded online advertizing as something that would deform the online environment and separate online businesses from their users which at that time still conceivably something that you could think about as customers.

Leo : But, they were collecting data about how people used Google to improve the engine. At that time, the deal was, we collect data to improve the product.

SZ Yes and things like improving search and also figuring out that great translation machine because their spelling errors and that would feed into the learning the computer learning for translation and so forth. So at that point what I think of as balance of power between its users I feel like i'm stuck with that. I've often said you should know what data they're collecting and rational decision about whether you're getting enough value back you're putting into the machine but that turns to be an antiquated concept. Unfortunalety, we blew past that a long time right the dot-com bust that's right the Vp's went saying "How are you gonna make money exactly ?" There was a lot of terror and a lot of anxiety. people who documented that period are very eloquent about it. What I like to say is that while these events have been poorly understood and barely, we can make some progress about this balance of power. That's why guys never wanted users because at that point, the company needed its users for this data, to improve its services just about as much users needed the company, so there was this nice balance of power and there's nothing wrong with collecting data if it's very clear and there is a clear social and legal contract that those dta are being used for the benefit of the users. In this crisis, the concept of state exceptiion a little bit going on in Washington today the state of exeception iw what is called upon to say : "all right, we've got such an emergency going on here that we have to suspend our normal values procedures and we've got to do something really dire". Why ? Because it's emergency so essentially the founders state of exception and in the state of xeception they decided to go back to this whole realm of advertizing and see if they could'nt make a kind of breakthrough there and parallel to this as people searching and all of that thare was collateral data off by those activities that Google wasn't using and in half hazrd its stored according to the documentation called data log and the were in servers but they weren't organized. However Google who had been playing around with those data logs understanding that there was tremendous predictive value in those discarded data. When the state of exception went back they began to take more and more data, data which were exclusive to Google and then you add that Google's computational ability which by the may even back then they like to talk about as "AI". When you put that all together they discovered that they had the way to make very very very accurate predictions and kinds of ads that a user is likely to clik on. The CTR (click-through rate) of course became this experience the perceived relevance of an ad so what Google did was at that point you still had a system where advertisers were choosing keywords and deciding where they would place their ads. Google sais "look, we are gonna completely out of that choice and let our new computational abilities and our new data that we're integrating to choice for you based on these new skills and going to tell you where to pur the ad based on whate these analyses tell us." By telling where to put an ad, we can ow target person and people who are mostly likely to click on. The rest is history as everybody knows but what is it happened and what we've got to figure out is : we took data that was more than was required to make a better engine or a great tranlation product and so forth these extra surplus, digital exhaust and things like waste material gonna be upset it and if we take this waste material and apply our unique computational capabilities, what we are axctually producing is prediction products. But what are these predictions ? Let's just zoom out a little bit and decouple idea of CTR. Wat we're really saying is these are predictions of future behaviors you are going to do now, but also soon and later so we have behavioral surplus combined with new machine intelligence to create prediction products that predict the future of human behavior. SO we're getting targeted advertizing on our behavior that makes Google and Facebook in particular the two chief practicers of this very rich and very powerful market. But it's just advertising what's the harm in that you're talking aboutb an inflection point where it becomes you use the term instrument arian power womething a lot less benign what is happeing now is it about to happen it's all around us so you're abseolutely right and I forget word you use but you could ask "What is wrong with that ?" We began with targeted advertising habituating the way that you suggest and of course what companies were learning was the more behavioral surplus I can bring to this operation the better my predictions. So now we are intpo economies of scale so it's not just what Leo is searching for browsing, for spelling, now we've got to invent the methods that are gonna follow Leo all over the web and we're gonna be taking data behavioural from Leo's actions, data for which we canc reate these predictive inferences, how to find that data, how to take those data even if Leo has said he wants those data to be private and even if Leo has refused to share those data. We are moving beyond the simple targeting into a deeper set of methods that are actually saying no matter what you have decided about this, and how data to be sued these methods are now designed to bypass to those kinds of barriers, to literally rob you. Maybe we can habituate to that maybe the next thing that happens this of course begins to spread through silicon valley, to Microsoft but more importntly it becomes default about every startup and app. So there's more competition for behavioral surplus. So now it turns out well th surplus isn't enough we also need to have a variety or surplus it's not enough to have scale we also need scope. What does scope mean ? We gotta get out of this laptop or the desktop we've to get out into the world, the shift to mobility becomes critical, because now the chief supply chain interface becomes the phone and not the laptop. Now we're out in the world so what is what is behavioral surplus that becomes most valuable it's where you are and what you're doing. We want to knwo your feelings, what your face look like, your affective state and this is how affective computing: nice capability where a computer could learn something about your emotions. So now affective computing is a very big growth industry where people use camera, facial and voice recognition, in order to make assement about your feelings and become very predictive behaviroal surplus.

So now we're way beyond taregted ad, we re delving into every aspect of your experience, behavior and feelings.

But it turns out there is a third big discovery : it's not just scale, it's not just scope, the most predictive behavioral surplus comes from actaully interve,ing in your experience and your activities, doing sbtle things that will shape, nudge and tune, and herd you in directions that lead you toward the kind of commercial outcomes that SC customers are paying for.

Leo : Everbody is participating on that. Everything is now in the database. What are we going to do with that ? Your book explain it with a great academic rigor. No one can deny this is happening. But you've just said something: "they go to predicting to influencing". Is that happening now ?

SZ Once you understand these economies of action, conversation I have in the book about SC producing a new quality of power, new species of power. And if I don't explain these underlying mechanisms, it is just science fiction. But once you understand this is already happening, this is all around us, this is not science fiction. Our audience is immersed in this, and we know this, being immersed, it's really hard to see it. Part of the service that I've tried to provide is to know not be immersed in it and come outside and look at it from a different perspective.

So when it comes to this idea that your behavior and shape, that's really predictive because the whole idea now is for the predictions to approximate observations.

... The end of this discussion is available here on Triangulation 380.

All Posts

Almost done…

We just sent you an email. Please click the link in the email to confirm your subscription!