Surveillance as Social Sorting

Privacy, Risk and Automated Discrimination

David Lyon, David Lyon

• Publisher: Routledge
• Year: 2005
• ISBN:
  9781134469031
• Edition: 1
• Topic: Medicine
• Subtopic: Medical Theory, Practice & Reference

Surveillance happens to all of us, everyday, as we walk beneath street cameras, swipe cards, surf the net. Agencies are using increasingly sophisticated computer systems - especially searchable databases - to keep tabs on us at home, work and play. Once the word surveillance was reserved for police activities and intelligence gathering, now it is an unavoidable feature of everyday life. Surveillance as Social Sorting proposes that surveillance is not simply a contemporary threat to individual freedom, but that, more insidiously, it is a powerful means of creating and reinforcing long-term social differences. As practiced today, it is actually a form of social sorting - a means of verifying identities but also of assessing risks and assigning worth. Questions of how categories are constructed therefore become significant ethical and political questions.Bringing together contributions from North America and Europe, Surveillance as Social Sorting offers an innovative approach to the interaction between societies and their technologies. It looks at a number of examples in depth and will be an appropriate source of reference for a wide variety of courses.

Part I
Orientations

1 Surveillance as social sorting
Computer codes and mobile bodies

David Lyon

This first chapter explores some of the key themes involved in “surveillance as social sorting.” The first four paragraphs state the argument in brief, before I suggest a number of ways in which social sorting has become central to surveillance. In what follows I look at some implications of surveillance as a routine occurrence of everyday life; focus on the emergent “coding” and “mobile” aspects of surveillance; and conclude by suggesting some fresh directions for surveillance studies in the early twenty-first century.¹

Surveillance has spilled out of its old nation-state containers to become a feature of everyday life, at work, at home, at play, on the move. So far from the single all-seeing eye of Big Brother, myriad agencies now trace and track mundane activities for a plethora of purposes. Abstract data, now including video, biometric, and genetic as well as computerized administrative files, are manipulated to produce profiles and risk categories in a liquid, networked system. The point is to plan, predict, and prevent by classifying and assessing those profiles and risks.

“Social sorting” highlights the classifying drive of contemporary surveillance. It also defuses some of the more supposedly sinister aspects of surveillance processes (it’s not a conspiracy of evil intentions or a relentless and inexorable process). Surveillance is always ambiguous (Lyon 1994: 219; Newburn and Hayman 2002: 167–8). At the same time social sorting places the matter firmly in the social and not just the individual realm – which “privacy” concerns all too often tend to do. Human life would be unthinkable without social and personal categorization, yet today surveillance not only rationalizes but also automates the process. How is this achieved?

Codes, usually processed by computers, sort out transactions, interactions, visits, calls, and other activities; they are the invisible doors that permit access to or exclude from participation in a multitude of events, experiences, and processes. The resulting classifications are designed to influence and to manage populations and persons thus directly and indirectly affecting the choices and chances of data subjects. The gates and barriers that contain, channel, and sort populations and persons have become virtual.

But not only does doing things at a distance require more and more surveillance. In addition, the social sorting process occurs, as it were, on the move. Surveillance now deals in speed and mobility. In the race to arrive first, surveillance is simulated to precede the event. In the desire to keep track, surveillance ebbs and flows through space. But the process is not one-way. Socio-technical surveillance systems are also affected by people complying with, negotiating, or resisting surveillance. Now let me spell this out, a little less breathlessly.

A key trend of today’s surveillance is the use of searchable databases to process personal data for various purposes. This key is not “technological,” as if searchable databases could be thought of as separate from the social, economic, and political purposes in which they are embedded. Rather, the use of searchable databases is seen as a future goal, even if, at present, the hardware and software may not all be readily available or sufficiently sophisticated. The point is that access to improved speed of handling and richer sources of information about individuals and populations is believed to be the best way to check and monitor behavior, to influence persons and populations, and to anticipate and pre-empt risks.

One of the most obvious examples of using searchable databases for surveillance purposes occurs in current marketing practices. Over the past two decades a huge industry has mushroomed, clustering populations according to geodemographic type. Canada, for instance, is organized by Compusearch into groups – from U1, Urban Elite to R2, Rural Downscale – which are then subdivided into clusters. U1 includes “The Affluentials” cluster: “Very affluent and middle-aged executive and professional families. Expensive, large, lightly-mortgaged houses in very stable, older, exclusive sections of large cities. Older children and teenagers” (TETRAD 2001). U6, Big City Stress is rather different: “Inner city urban neighbourhoods with the second lowest average household income. Probably the most disadvantaged areas of the country . . . Household types include singles, couple, and lone parent families. A significant but mixed ‘ethnic’ presence. Unemployment levels are very high” (TETRAD 2001).

Using such clusters in conjunction with postal codes – zip codes in the USA – marketers sift and sort populations according to their spending patterns, then treat different clusters accordingly. Groups likely to be valuable to marketers get special attention, special deals, and efficient after-sales service, while others, not among the creamed-off categories, must make do with less information, and inferior service. Web-based tools have broadened these processes to include other kinds of data, relating not only to geodemographics but to other indicators of worth as well. In processes known variously as “digital redlining” (Perri 6 2001) or “weblining” (Stepanek 2000), customers are classified according to their relative worth. So much for the sovereign consumer! The salesperson may now know not only where you live, but details such as your ethnic background (Stepanek notes that in the USA Acxiom matches names against demographic data to yield “B” for black, “J” for Jewish, “N” for Nipponese-Japanese and so on).

Already one may see how off-line and on-line data-gathering may be matched or merged. As the Internet has become more important as a marketing device, so efforts have increased to combine the power of off-line (mainly geodemographic) with on-line (mainly surfing patterns and traces) databases. This was behind the purchase of Abacus (off-line) by Doubleclick (on-line) in 1999, that resulted in a lengthy court case following an outcry. When marketers merge individually identifiable information pertaining to postal or zip code characteristics with evidence of purchasing habits or interests gleaned by tracking Internet use into a searchable database, they create a closer relationship with relevant customers. In a striking case, an American physician was recently offered a list of all her perimenopausal patients not on some estrogen replacement therapy (Hafner 2001). On-line and off-line data may be combined to produce fine-tuned sales.

Another field in which searchable databases have become more important for surveillance is policing. During 2001, Toronto, Ontario, Canada police introduced upgrades in their patrol vehicles that extended the scope of information-based activities. The e-Cops – Enterprise Case and Occurrence Processing System – was adopted, which uses wireless data communications to connect police officers using laptops in their cruisers to Web-based tools for crime detection and prevention (Marron 2001). Not only can officers now connect directly with Toronto Police Service files as well as the Ontario Ministry of Transport drivers’ license records and suspect lists held at the Canadian Police information Centre (CPIC), but also with an IBM database and business-intelligence software.

This initiative, like database marketing, makes use of geodemographic information. In this case, it identifies geographical patterns of crimes with a view not only to detection but to pre-empting crime by indicating where a particular offender may strike next. The new systems automate tasks that previously required clerical staff as information-processing intermediaries, and connect tools that used to be used in relative isolation. Thus, the system is more fully integrated, and, it is argued, more cost-effective. Background information on suspects is now instantly available and retrievable by officers at their car seat laptops. And the searchable database may be used to indicate whether the suspect is likely to be a serial offender, on a “crime spree” or a novice.

As with database marketing, the policing systems are symptomatic of broader trends. In this case the trend is towards attempted prediction and pre-emption of behaviors, and of a shift to what is called “actuarial justice” in which communication of knowledge about probabilities plays a greatly increased role in assessments of risk (Ericson and Haggerty 1997). How certain territories are mapped socially becomes central to police work that is dependent on information infrastructures. But such mapping also depends on stereotypes, whether to do with territory – “hot spots” – or social characteristics such as race, socio-economic class or gender. As Ericson and Haggerty observe, these categories cannot be impartial because they are produced by risk institutions that already put different value on young and old, rich and poor, black and white, men and women (Ericson and Haggerty 1997: 256).

The two examples, from marketing and from policing, clearly indicate how searchable databases have become central to surveillance. If surveillance is understood as a systematic attention to personal details, with a view to managing or influencing the persons and groups concerned, then the searchable database may be seen as an ideal in other emerging areas as well. Risk management and insurance assessment in particular tend to encourage the quest for greater accuracy of identification and faster communication of the risk, preferably before the risk is realized. New technologies, such as biometrics, using fingerprints, handprints, iris scans, or DNA samples, are harnessed for accuracy of identification (see Nelkin and Andrews in this volume), and the networking of these to increase speed of communication are thus part of an increasingly common pattern.

Two further developments also illustrate these surveillance shifts. One refers to the rapid proliferation of Closed Circuit Television (CCTV) or “video surveillance” and the other to a growing range of locational devices that not only situate data subjects in fixed space, but also while on the move. Again, these are not merely technological innovations with social impacts. They are technologies that are actively sought and developed because they answer to particular political-economic pressures. The political pull factors have to do with neo-conservative governments wishing to contract out services and to cut costs, especially labor costs. In so doing, they are also attempting to reduce public fear of crime and create spaces for “safe” consumption in the city. Pull factors on the commercial side include narrowing profit margins and the desire to capture markets through relationships with customers. The push factors, on the other hand, relate to the drive to sell (companies) and to adopt (agencies, organizations, governments) new technologies.

The UK is the currently unrivalled world capital for video surveillance in public places, but other countries are rapidly following the British example. Major cities in North America, Europe and Asia are using CCTV as a means to control crime and to maintain social order. For example, Sudbury was the first city in Ontario to install public video cameras in 1996, in a move inspired directly by the example of Glasgow, Scotland. Sudbury police obtained help from rotary clubs and Canadian Pacific Rail to put up their first cameras which, it is now claimed, have led to significant reductions in crime rates – which are falling faster than those in other Canadian cities (Tomas 2000). In most cases, searchable databases are not yet used in conjunction with CCTV, though the aim of creating categories of suspicion within which to situate unusual or deviant behaviors is firmly present (see Norris in this volume).

In some cases, however, searchable databases are already in use in public and private situations, to try to connect facial images of persons in the sight of the cameras with others that have been digitally stored. In Newham, London, CCTV is thus enhanced by intelligent systems capable of facial recognition (Norris and Armstrong 1999). In a celebrated case in 2000, the turnstiles at the annual Superbowl events in Florida were watched by such a CCTV system, that compared the 100,000 plus images of those entering the stadium against stored images of the faces of known offenders (19 matches were made). This was a test-run by a camera system company, to demonstrate the capabilities of the machines, which at least suggests the nature of technology push factors in this case (Slevin 2001).

Much more commonplace than street-level facial recognition systems – at least before 11 September 2001 – are the facial recognition technologies used at casinos to catch cheats. As with the turnstiles, the casino entrances offer the opportunity to capture relatively clear images. These may then be matched with database images and used to apprehend known offenders (CNN 2001). The increasing use of digital security cameras is likely to encourage this trend (Black 2001). Since 11 September 2001, however, widespread interest has been expressed in many cities for facial recognition CCTV systems to reduce the likelihood of “terrorist” attacks. The new political will and public willingness to countenance the spread of such systems in public places has been more than matched by confident “expert” announcements about available technology, even though their capacity actually to perform as required is unproven (Rosen 2001).

Sophisticated CCTV systems, such as those in Newham, London, may be used to follow people from street to street if they are of interest to the operators. Thus not only fixed sites, but moving targets may also be subject to surveillance. Cameras, however, are not the most common kinds of devices used for keeping track of persons on the move. Other locational technologies that use Global Positioning Satellites (GPS) and Geographic Information Systems (GIS) in conjunction with wireless telephony provide much more powerful surveillance potential. There is already a popular market for such mobile phone and satellite technologies among parents wishing to keep track of their children, but broader commercial interest is found among car rental companies, emergency, and security services.

Selected new cars from Ford offer the On-Star service that enables, for example, hotels or restaurants to alert users when they are near by. This is a predictable extension of electronic business. Following a federal ruling in the USA, cell phones will carry wireless tracking technology to permit the pinpointing of persons making emergency calls (Romero 2001). This, too, is predictable, and the benefit to persons in trouble, palpable. But such systems also permit other agencies – insurance companies, employers – to discover the whereabouts of individuals and it is only a matter of time before they will develop the means of profiling them too (see Bennett, Raab, and Regan in this volume). In the UK, recent legislation, the Regulation of Investigatory Powers Act, allows police unparalleled access to new-generation cell phones (“mobile” phones in the UK or Australia) for tracing the location of cal...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Contributors
• Preface and Acknowledgements
• Introduction
• Part I: Orientations
• Part II: Verifying Identities: Constituting Life-Chances
• Part III: Regulating Mobilities: Places and Spaces
• Part IV: Targeting Trouble: Social Divisions