Discussion: The soft power of biometric surveillance

Excerpted from Jathan Sadowski and Frank Pasquale:

At one end of the spectrum of control are the technologies that enact their power in subtle ways. They are increasingly ubiquitous and become subsumed into the background of everyday life due to their ‘invisibility’ (Star, 1999). That is, they can be functionally invisible because people no longer notice their relationships and interactions with the technologies and/or physically invisible because they are intangible or hidden.

Political philosopher Giorgio Agamben [31] explains this particularly curious mode as an “operation of power that does not immediately affect what humans can do — their potentiality — but rather their ‘impotentiality,’ that is, what they cannot do, or better, can not do.” Through this way of operating, power is not limiting my capacity to do an action — in the conventional way of constraining subjects — but instead is making it very difficult for me to not do an action. For example, when few people had cellphones it was easy to not own or use one, but now that almost everybody does — and increasingly more parts of our life our tied to the constant communication and platform capabilities afforded by the device — it’s nearly impossible not to also conduct your life via a cell/smartphone (Peppet, 2011; Morozov, 2014). The same can be said of automobiles: nothing forces any given person to buy an automobile, but when infrastructure is constructed with private vehicles assumed, and when there are scant other alternatives, it becomes difficult to not make the choice.

Urban surveillance technologies — especially as they are implemented as part of massive, networked systems — anchor the subtle end of the spectrum of control. Consider the closed-circuit television (CCTV) arrays that already blanket the streets and buildings of major cities around the world. CCTV is now emerging as a kind of “fifth utility” within cities (alongside gas, electricity, water, and telecommunications). “Once CCTV systems are installed, their logic is inevitably expansionary. Economies of scale are very marked — once a system is built and monitoring personnel are employed, it makes sense to cover larger and larger areas” [32]. In the ‘smart city,’ such surveillance systems rush down the path towards ubiquity; they become subsumed into the background of everyday life: always present, tirelessly watching, but rarely noticed.

CCTV is a flexible technology, with the potential for added layers of sophisticated software incorporated into the hardware — such as biometrics that are linked to the in-depth, personal information held and managed by data brokers. The proliferation of surveillance systems as part of ‘smart’ initiatives, which are then enhanced by advanced analytics, changes the very political economy of what it means to be a city dweller.

Let’s consider further the example of biometrics, which identify, measure, and collect a biological trait or group of traits [33]. There are a wide variety of types of existing biometrics, with more in development. Some of the most common focus on physical traits: faces, fingerprints, irises, retinas and DNA. Others focus on behavioral traits: voice, signature, gait (how a person walks) and keystrokes (speed and timing between key presses). In practice, biometric technologies employ a standard process across different types. A sample of the biological trait is collected using a sensor of some kind, such as a camera for faces or a telephone for voices. Through the use of an algorithm that extracts information from the biometric sample, the trait is then converted into a digital representation called a “template,” which can be stored in a database. The larger the database, the more templates there are to verify or identify subjects. The key component, though, is the algorithm used to construct the template. This is the feature that distinguishes one biometric recognition system as ‘better’ than others on markers like: can the algorithm quickly extract biometric information? Can it do so in a variety of environmental circumstances? Can it create a template that is accurate?

The potential role of biometrics in the information economy is huge — especially for the massive data-brokerage industry. During a 2013 U.S. Senate hearing, Senator John D. Rockefeller IV, then Chairman of the U.S. Senate Committee on Commerce, Science and Transportation, said: “In 2012, the data broker industry generated $156 billion in revenues. That’s more than twice the size of the entire intelligence budget of the United States Government — all generated by the effort to learn about, and sell, the details about our private lives.” [34] Biometrics present new ways to convert data into profit, a figurative strip-mining bodies (and their actions) so that ever more actionable information can be extracted from them. This analogy captures the degree of intrusiveness that biometrics have when they hone in on particular biological traits and pull them out of the context of the rest of the body, person, and environment.

The finer grain, personalized data provided through biometrics would be like gold in the data brokers’ servers, enabling companies to significantly fine-tune the way they target potential customers and providing government agencies with additional ways to oversee populations. “Despite consumer data broker companies’ clear links with credit rating agencies, revenues numbering in the billions, exemption from state regulations to protect consumers from identity theft, and documented data breaches, the average citizen has likely never heard of these powerful corporations” [35]. Since these brokers collate data and construct profiles through whatever means available, by adding biometric algorithms and databases to the mix, these brokers, and crucially their clients, accumulate troves of data to the point that they may know more intimate details about persons (including income, debt, illness, criminal records, and drugs taken) than their families do. Some high-end stores already use facial recognition software to alert clerks and salespeople that a VIP or a celebrity is in the store (Salinas, 2013). With large enough databases, what’s to prevent stores from identifying even non-VIP customers who walk in the door?

The acceleration of profiling and personalization is a natural consequence of big data business strategies. Firms at the center of the big data economy claim that their data troves reverse the common economic law of diminishing marginal returns. The more data a firm has, the more its existing store is worth, since contextualization of profiles enables ever greater power of sorting, control, price discrimination — and even blackmail.

These implications, among others, are consequences of the ways biometrics allow — and encourage — more intensive commodification of physical bodies. “Biometrics break bodies down into their component parts in ways that allow them to be marketed more easily in the transnational marketplace … The flimsy material body is rendered rugged as biometric technologies make the body replicable, transmittable, and segmentable” [36]. We’ve heard of the data economy, but how about the face economy, or iris economy or gait economy? There are entire corporate sectors eager to mine that data and put it to use in any number of ways: data brokers construct in-depth consumer profiles replete with biometric templates; salespeople and store security use biometric emanations to pull up your reputation from the database; and your identity is pinned to your location, which is better tracked as you move through the streets, public squares and shops. Insurance companies, for instance, are hungry for the somatic data provided by personal health and fitness monitoring devices (Sadowski, 2014a). Imagine what they, and others, could do with the knowledge and power provided by diverse types of biometrics.

Thus, biometrics present a way to not only dividualize people at minute scales, but also provide the means to intensify commodification — via strip-mining the newly available sources of data — and control — via biopolitical management — of people, all while the ‘smart city’ constructs a conducive platform for these activities.
The technological systems installed within cities to make them more connected, efficient, secure, and smart don’t exist in a vacuum. They “absorb and reproduce the dominant cultural values of the contemporary political economy” [37]. At the subtle end of the spectrum of control, the systems act on us in ways that are functionally and/or physically invisible; few even know about data brokers, intrusive surveillance, and the ways we become incorporated into the data flows of capital. And even then, we “consent” by default because the options to not do things that pull us into the logics of these systems — such as not using digital platforms, not using smartphones, not going to stores and streets without a mask, not living in a populated area — can hardly be considered real choices for the vast majority of citizens.”