The Hyperreality of the Datafied Self

(1)

The Simulational Relationship between the Individual and the Datafied Self in

Contemporary Modes of Databased Surveillance

Leo van Schie 29 June 2018

(2)

Imagine having a police officer knock on your door for no apparent reason, warning you that you are being watched, and that there will be major consequences if you were to commit any crimes. This happened to Robert McDaniel, a 22-year-old Chicago resident who was unaware of the fact that he had been placed on the city’s Strategic Subjects List – or what has become more commonly known as the “heat list”.1_{This list contains the people in}

Chicago that are most likely to be associated with violent crime. It is aimed at predictive policing: “this is about saving lives”, as reported by Commander Jonathan Lewin, head of information technology at the Chicago Police Department. The list is the result of an algorithm that uses different data on individuals to calculate and assign them risk scores from one to five hundred-plus, with an astounding four hundred thousand Chicago citizens having already been branded with an official police risk score.2_{In addition to supposedly}

saving lives, these risk scores also change lives. The risk scores determine who is targeted for proactive police intervention, such as home visits by police officers, additional surveillance or invitations to community meetings – reminding people of the fact that they are being monitored and branding them as potential suspects, regardless of whether they have actually committed a crime. Critics have stated that the list contains tens of thousands of people, just like Robert McDaniel, who have gained high risk scores without any history of prior arrest for violent crimes or gun violence.

This is just one out of many examples of the ubiquity and increased pervasiveness of contemporary modes of databased surveillance. There is an increased awareness of the fact that data is collected on everyday interactions with digital technologies, and questions surrounding privacy concerns are becoming increasingly urgent and relevant in the context of the ever tightening grip of contemporary dataveillance. As more and more formerly offline practices and interactions are taken over and replaced by alternatives within the digital realm, there are hardly any aspects of everyday life left that are free from the

collection of personal data. Therefore, there are also hardly any aspects of everyday life left that are not also shaped by the consequences of practices of dataveillance.

1 Matt Stroud. “The Minority Report: This Computer Predicts Crime, But Is It Racist?” The Verge, 19 February 2014.

https://www.theverge.com/2014/2/19/5419854/the-minority-report-this-computer-predicts-crime-but-is-it-racist

(5)

Another example: in Minneapolis, a man walked into Target angrily asking why his teenage daughter received discount coupons for baby clothes and cribs. Why would they encourage teen pregnancy? The manager apologized to the man, stating that he had no idea as to how this could have happened. When they spoke again later however, it was now the man himself who apologized: his daughter indeed turned out to be pregnant. Based upon data collected on the girl’s prior consumption patterns at Target, they were able to calculate that she was indeed very likely to be pregnant. Target knew that the girl was pregnant even before her very own father did.3_{This case illustrates the invasiveness of strategies of data}

collection and the use of algorithmic profiling by companies. These processes have become key forces in informational capitalism where whole economies revolve around the collection of personal data and their use for predicting and shaping future consumption patterns of the individual.

These previous two stories are examples of how outside actors are able to target the individual through the aggregates of data collected on them. There are also practices in which the individual is able to monitor him- or herself through their own data trails. This is central to practices of self-tracking, where digital devices close to the body such as

smartphones and wearables with accompanying apps have enabled users to collect data on their bodily functioning. The data collected here quantifies and makes visible certain bodily qualities that were more opaque before, offering their users a deeper understanding of their bodies which they can act upon in order to improve their health and functioning and to prevent medical problems. For example, in a promotional video for a new series of the Apple Watch, a guy named Paul talks about how this tracking device saved his life. The self-tracking of his health led him to seeking out medical attention which ultimately got him diagnosed with a condition that was causing his liver, kidneys and heart to start shutting down: “had I not been wearing my Apple Watch I never would have sought medical attention, which in turn saved my life.”4

Our interactions with the world, and even with ourselves and our own bodies, are increasingly shaped by digital technologies. The infrastructure of so many different areas of everyday life has become digital, translating practices, identities and bodies into digital data

3 Charles Duhigg. “How Companies Learn Your Secrets.” The New York Times Magazine, 16 February 2012. https://www.nytimes.com/2012/02/19/magazine/shopping-habits.html

(6)

that can be collected, processed and stored – opening up new and more efficient ways of interaction as the individual becomes knowable as a citizen, as a consumer, and as a body. Though motivations and processes of data collection may differ, the ubiquitous generation, analysis and processing of personal data has become one of the main drivers of

contemporary information capitalism. These processes of datafication and their

consequences have sparked interest within different fields of study. Most of the critical analyses of these developments come from the fields of surveillance studies and media studies, exploring the implications of the proliferation of data generation and collection in relation to the disciplining and subjectification of the individual, their employment in strategies of national security or marketing practices, and their consequences for the individual’s rights to privacy. More recent movements such as that of the Quantified Self have focused on the possibilities that the datafication of the everyday, for example through self-tracking, may offer its users in terms of self-knowledge and self-optimization. These different fields of study have brought forward different types of discourse in the context of the ubiquity of technologies of datafication, some more theoretical and critical, others more practical and optimistic. Central in most theory regarding the datafication of the everyday are the conceptualizations of the aggregates of data that are produced on the individual. These aggregates of data are often understood as some kind of representation of the individual in digital data, that come into being through interactions with the user, as an original counterpart to the datafied self5_{produced through his or her interaction with digital}

technologies. When technologies of data generation, collection and processing become increasingly naturalized and invasive, and the aspects of everyday life which can be translated into digital data keep expanding, these aggregates of personal data, these datafied selves, become increasingly actionable.

This actionability leads to various instances in which interaction with the datafied self takes place prior to, or even instead of interaction with the actual ‘offline’ individual because of its transparent, measurable, shareable, and thus more efficient character. Targeting the individual, and exploiting his or her datafied self for commercial or governmental purposes can occur separately from its ‘original’ – from the user itself. Moreover, through interaction with the individual’s datafied self, corporations or institutions can target the ‘offline’

5 The “datafied self” will be used to refer to the aggregates of personal data that can be linked back to a specific individual – as a more neutral term in relation to the other conceptualizations that will be discussed

(7)

individual in ways that affect their present and future actions and circumstances. What takes place here is that the interaction with the actual, offline individual in a sense becomes secondary to the interaction with his or her datafied self. This raises important questions concerning the relationship between the datafied selves and the users, for because of their actionability, the datafied selves can be understood as becoming increasingly ‘real’ in relation to the individuals they can be linked back to.

Most conceptualizations of these phenomena have focused on the mechanics of data collection, profiling, and targeting and their implications for the individual in terms of their rights to privacy and on the exploitation of their free labour by corporations. Often, the relationship between the datafied self and the individual has been understood as one of representation. The datafied self is perceived to be a representation that is established with the offline individual as its ‘original’ counterpart. The relationship between signifier and signified in a mode of representation implies that the individual is the primary actor that, through its interactions with digital technologies, constitutes a datafied self that offers insight into the individual which, because of this, becomes actionable. Some of the more recent conceptualizations have started to move away from this representational view, for example as Louise Amoore and John Cheney-Lippold, two authors discussed in this research have. They focus more on the workings of these processes of datafication and on the

datafied selves as tools for strategies of databased surveillance and marketing practices. However, what is lacking thus far is an alternative approach to the relationship between the individual and the datafied self, that offers a more productive, in-depth theoretical

understanding of the increased reality of the datafied selves in relation to the offline individual. This research will argue that the approach to this relationship as one of

representation is not sufficient in contemporary modes of databased surveillance and that the relationship between the datafied self and the individual is more complex. For different reasons discussed in this research, the continuous reciprocity between the datafied self and the user has led to the blurring of distinctions between a signifier and a signified, an object and a subject, a beginning and an end, or even an original and a representation. The increased ubiquity and naturalization of processes of datafication have altered the ways in which we interact with ourselves and with the world, and have changed processes of

signification. The datafied self should be understood not just as becoming increasingly real in relation to its ‘original’. Rather, there are various instances in which it becomes hyperreal: it

(8)

becomes self-referential and can function as a substitute to its original, in ways which have a very real impact on the ‘original’. This research proposes turning to French philosopher Jean Baudrillard’s theory on hyperreality and simulation as a theoretical perspective that enables a shift in thinking about the processes of signification in the datafication of the everyday. The relation of signification apparent in the datafication of the self should not be understood as a relationship of representation, but rather as one of simulation, as the datafied selves as hyperreal replicas of the individual increasingly insert themselves into reality.

This research will discuss three fields of contemporary databased surveillance in different contexts. Central here will be an exploration and comparison of the varying relationships between the subject and his or her datafied self, to understand the consequences of the increased reality of the aggregates of personal data in different contexts and to approach the hyperreality of the datafied self from different perspectives. Firstly, the data derivative, a conceptualization of the datafied self in a context of border and national security will be discussed. The second chapter will look at the datafied self and the moulding of consumers under surveillance capitalism. In the third chapter, the practices of self-tracking as a specific form of self-surveillance will be central, in a discourse where more agency is ascribed to the users of these technologies. By analysing the ways in which some of the most influential scholars in these fields of research view the relationship between the aggregates of personal data and the individual on whom they are collected, the position of the datafied self with regard to its offline counterpart will be mapped out. In each chapter the specific relationships between the datafied self and the individual will be subjected to a more in-depth theoretical reflection in order to think about the ways in which processes of datafication of the self transform the individual’s relation to reality and how the very position of reality itself has changed in relation to the constant datafication of everything. Through examining these different modes of surveillance and the different contexts in which the datafied selves are established, it will become clear how a view of the relationship between the subject and its datafied self as one of representation is no longer accurate. As demonstrated by the contemporary practices of databased surveillance discussed in this research, as well as by some of the scholars that analyse them, this representational idiom is no longer sufficient. Thus far, there is no productive alternative for theorizing this

relationship. This research will analyse these three different fields of surveillance and build upon its strategies and their conceptualizations by some of the leading scholars in the fields

(9)

of surveillance studies and media studies in order to establish a more fundamental theory on the changing interconnection between the subject and his or her datafied self. This will happen through placing the findings from the analysis of the different modes of

dataveillance into interaction with theory on hyperreality and simulation as discussed by French philosopher Jean Baudrillard (1929 – 2007).

Jean Baudrillard’s theory will be used as a theoretical perspective that enables an understanding of the changed position of reality in a landscape saturated with media technologies and processes of mediation that becomes more and more complex and opaque. In one of his most influential works, the 1981 book Simulacra and Simulation, he explores the relationship between reality (or, what is left of it) and representations. He states that we have arrived at a point at which the representation – the simulacrum – has come to precede and even substitute the original which it supposedly refers to. There is a blurring of distinctions between reality and representation, between signifier and signified and between object and subject. Baudrillard states that the representations, as signs of the

real, have come to precede their ‘real’ counterparts. When using Baudrillard’s theory written

in the 1980s to understand contemporary forms of datafication, new questions arise regarding the relationships between representation and reality and the ways in which we interact with the world and with ourselves when the pre-emptive nature of contemporary modes of dataveillance establishes models of reality that shape the individual’s present and potential futures.

Through interaction with one’s datafied self, different futures of the individual cannot only be predicted, but also shaped in a way in which the verification of an independent reality (or, of the ‘original’) becomes increasingly irrelevant. Baudrillard’s theory on hyperreality and simulation enables a shift in thinking about the reciprocal relationship between the self and processes of datafication, questioning the common understanding of this relationship as a form of representation that has an original and a ‘copy’ in digital data. According to Baudrillard, the causality of representation, as a relationship between signifiers and signifieds, has collapsed, and has become replaced by a logic of simulation.

Understanding the relationship between the individual and his or her datafied self through this logic, it can be argued that there is no clear distinction to be made between the self and its representation in data. Due to the ubiquity of digital technologies and processes of datafication, everyday life is already structured by these processes to such an extent that

(10)

there is no longer really an ‘original’ cut loose from these processes. Not only do many practices of datafication shape who you are and what you do, additionally, these datafied selves and algorithmic processes of targeting the individual increasingly impact what it

means to be you. What is lacking is an in-depth theorization of the actual relationship

between the datafied self and the individual under contemporary modes of databased surveillance, that focuses on how the datafication of all aspects of everyday life not only establishes datafied selves but how these practices have consequences for the individual’s relation to reality now that their datafied selves have become even more comprehensive than ever before and are substituting their ‘original’. This leads to shifting understandings of reality and representation and the ways in which the meaning of the individual as a citizen, as a consumer and as a body is altered not by reality itself, but by models substituting reality. Bringing existing conceptualizations of the datafied self under different modes of

contemporary databased surveillance into interaction with Baudrillard’s theory on

hyperreality and simulation leads to an understanding of the datafied self that emphasises its hyperreality, the ways in which it increasingly comes to precede and substitute its

‘original’ and how these processes may affect the individual’s present and potential futures, while simultaneously giving meaning to the individual him or herself – all based upon hyperreal models of signification operating through these practices of data processing.

Surveillance Studies and the Datafied Self

Before looking at the different ways in which the relationship between the individual and the datafied self is understood in the three modes of contemporary data-driven surveillance discussed in this research, a brief discussion of some of the key concepts of surveillance studies and the most influential conceptualizations of the separation between the individual and the aggregates of data is necessary as most newer conceptualizations refer back to them. The initial conceptualizations of these aggregates of personal data are proposed as a reaction to Michel Foucault’s theory on surveillance in disciplinary society and panopticism in his book Discipline and Punish: The Birth of the Prison (1975). Disciplinary power is a system of power aimed at the individual, that establishes order through different institutions and specific regulations of the organisation of spaces, time and behaviour, which is

(11)

institutions such as the prison, the hospital, the school, and the factory. Due to technological innovations such as computers and databases, in his “Postscripts on the Societies of Control” (1992), Gilles Deleuze argues that these disciplinary societies have been replaced by the societies of control (3), signifying the most important shift in practices of surveillance leading to the contemporary modes of databased surveillance discussed in this research. Due to technological developments and networks, the enclosure of physical spaces has become less defining for surveillance and the exercise of power. Subjects are free to roam from one network of monitoring to another: continuous surveillance that logs the subject’s behaviour through their freedom (Deleuze 1992, 6). Deleuze’s theory on the societies of control has led to one of the most influential conceptualizations of the distinction between the individual and the data that is collected on them. Deleuze states that there has been a transformation from individuals to dividuals (1992, 5). The individual is no longer an ‘indivisible’ entity. Rather, the individual can now be divided and subdivided endlessly. Particular information on a specific person can be used apart from the individual him- or herself and can be recombined in various ways outside of the individual’s hands, in ways beneficial to outside actors such as governmental institutions or corporate marketeers (Williams 2005, n.p.). This is where the most significant aspect of Deleuze’s concept of the dividual to this research comes forward: he recognized how the individuals themselves became less relevant as subjects of surveillance. No longer is it the individual as an actual person and body that is central, no longer the individual that has to be subjected and disciplined. Rather, it is now the individual’s representation: the divided individual (Galič et al. 2017, 20). What has changed in the societies of control, as opposed to the previous disciplinary societies, is that the making docile of bodies is not what is central anymore. Rather, it is now about the moulding of consumers and the controlling of access (at airports, borders, etc.) through the dividuals, whose data bodies have become dominant over their real bodies (Galič et al. 2017, 20).

These observations have become fundamental in the conceptualizations of the datafied self that followed the Deleuzian understanding of the societies of control and the dividual. The most influential concept for discussing the aggregates of data collected on the individual is that of the data double as proposed by Kevin D. Haggerty and Richard V. Ericson in “The Surveillant Assemblage” (2000), where they conceptualize the human body in a context of increased surveillance during the rise of computerized databases. In a similar

(12)

fashion to Deleuze, they move away from Foucauldian understandings of disciplinary society. They argue that a new set of analytical tools is necessary to conceptualize ‘contemporary’ surveillance. Part of this new toolbox is Deleuze and Guattari’s notion of the assemblage, which refers to “a multiplicity of heterogeneous objects whose unity solely comes from the fact that they work together as a functional entity”, that “comprise discrete flows of an essentially limitless range of other phenomena such as people, signs, chemicals, knowledge and institutions” (Haggerty and Ericson 2000, 608). These assemblages can turn into systems of domination (Haggerty and Ericson 2000, 609). This is how they conceptualize surveillance. They use the concept of the assemblage to grasp all of the heterogeneous elements into a flexible phenomenon, that is multiple, unstable and is lacking of clear boundaries or

relations to specific governmental departments (Haggerty and Ericson 2000, 609). They view surveillance as a continuous, flexible and mobile monitoring of everyday behaviour and patterns (Haggerty and Ericson 2000, 609).

Haggerty and Ericson state that the human body under surveillance has turned into a flesh-technology-information amalgam – it has become of a distinctively hybrid composition (2000, 611). Under surveillance, the body is ‘broken down’ by being removed from its territorial setting, after which it is reassembled in different settings via a series of data flows. This results in the establishing of a “de-corporealized body, a ‘data double’ of pure virtuality” (Haggerty and Ericson 2000, 611). The surveillant assemblage can be understood as a visualizing device, that visualizes previously opaque flows of information that pertains to the human body, which previously were beyond our normal range of perception (Haggerty and Ericson 2000, 611). What thus happens is that the surveillant assemblage naturalizes the capture of this human-technology hybrid and produces a new type of individual – one that consists of pure information (Haggerty and Ericson 2000, 614). This new type of

decorporealized and virtual individual is what they refer to as the data double. Rather than targeting the human body as a single entity “to be molded, punished, or controlled” (Haggerty and Ericson 2000, 612), as was the case in disciplinary society, it first has to be

known, by fragmenting it into series of ‘discrete signifying flows’ (Haggerty and Ericson 2000,

613) after which the body can be transformed into pure information – into the data double. This data double is a multiplication of the individual, an additional self (Haggerty and Ericson 2000, 613). In other words, the data double is an additional self that consists of pure

(13)

more opaque aspects of the individual – enabling the individual to be known. It is thus a form of representation of the individual into digital data – since this entity can be linked back to specific individuals and offers all the information available on him or her. Haggerty and Ericson acknowledge this, yet state that the data double transcends a purely

representational idiom (2000, 614), since it is not just a representation of an individual, but is a form of pragmatics, a means for institutions to particularize subjects. In other words, the data double is not just a representation, but is an actionable representation. The data double can thus be understood as the product of the surveillant assemblage as a visualizing device – reading and interpreting the fragmented observed body, and producing an additional and actionable replica of the self out of this interpretation that due to its mobile and comparable character is increasingly turning into the prime target of both governmental as well as marketing practices (Haggerty and Ericson 2000, 614).

This conceptualization of the data double, upon which most recent understandings of the datafied self are based, follows a representational idiom. Though Haggerty and Ericson state that it transcends a “purely representational idiom” due to its actionability, the concept itself is still based upon a relationship of representation: on the data double as an entity that is established with the subject as its original, as an additional self through which the individual can be known. However, the relationship between the datafied self and the individual has become increasingly complex, now that the separation between the two is blurring due to the ever increasing pervasiveness of processes of data generation and collection, as well as the dominant position this datafied self takes on with regard to its offline counterpart. In a context where strategies of databased surveillance are becoming more and more specific to the multiplicity of technologies involved and the opaqueness of these processes, this research proposes another basis for conceptualizing the relationship between the datafied self and the subject of strategies of surveillance. These strategies are becoming more focused on determining potential futures of its subjects, which in turn shapes the way in which they are targeted in the present – emphasizing potentiality rather than actuality. Additionally, these technologies also become more and more determining in processes of signification. It is not merely about the production of an additional, digital and actionable self, but by abstracting what it means to be, for example, of a certain gender, race, sexuality, social class, et cetera, based upon data collected on others, algorithms target the individual according to their presupposed identity in ways that shape the life chances

(14)

and choices of individuals. This results in a self-affirming process of signification that impacts the ways in which individuals understand the world, but also how they relate to themselves and even to their bodies – now that their very disposition has been altered by the

potentiality and invasiveness of these technologies. This research proposes an

understanding of the constant datafication of everything from a framework of hyperreality and simulation, enabling a shift in thinking about the causal relation of signification, which will lead to an alternative conceptualization of the relationship between the individual and his or her datafied self.

Chapters one to three of this research will analyse three partly overlapping, yet fundamentally differing modes of contemporary information-based surveillance: firstly, “National Security and the Data Derivative”, secondly, “Surveillance Capitalism and the Moulding of Consumers”, and finally, “Self-Surveillance and the Quantified Self”. The

dynamics between the individual and the datafied self within these three fields of discourse will be examined, placing the focal point upon the ways in which these different types of datafied selves relate to the individuals that they can be linked back to, and how that in turn affects their relation to reality. The findings in these chapters will be brought into an

interaction with Baudrillard’s theory on hyperreality and simulation, in order to analyse the blurring distinctions between reality and representation, between object and subject and between beginning and end. The relationship between the datafied self and the individual will be reconceptualised not as a form of representation, but as a form of simulation. By understanding this relationship as one of simulation, the complex relation of reciprocity between the data and the physical body of the individual can be understood more properly: the increased reality of the aggregates of personal data is central here, and will be used to establish more fundamental theory on the ways in which the datafied selves have

increasingly come to live our lives.

1. WHEN? NATIONAL SECURITY AND THE DATA DERIVATIVE

(15)

One of the central themes in surveillance studies is the way in which the monitoring of the human body and the information that this produces provides insight into the population. This insight is essential to contemporary strategies of national security through which the population is controlled and protected. Evelyn Ruppert argues that the data that is collected on individuals in part solves the general problematic of governing: to know the nature of the population. This insight is used to govern and regulate the forces of the population, the referent object of biopolitics (Ruppert 2012, 127). The biggest aim of surveillance processes is to know the population in order to govern properly, and secure the nation against

potential threats. In disciplinary society, surveillance involved the spatially-bound observation of the individual and his or her body in different institutions. Whenever misbehaviour was observed, the ones in power could intervene and repress: verification preceding judgement. However, in a post-9/11 world, where the fear for terrorist acts has impacted the sense of security, and the potential threat of another attack is always lurking around the corner, the system of surveillance has changed – and intensified. Due to

technological advancements, the continuous logging of everyday behaviour as discussed in the Deleuzian societies of control leads to huge databanks of accessible data on the

individual.

It is within this context that surveillance has increasingly become a system of pre-emption. Due to the emergence of big data practices, surveillance operations have shifted their focus onto the future, rather than on the present or past. This emphasis on

anticipation has led surveillance’s strategies to be centred more around the managing of

(potential) consequences (Lyon 2014, 6). In the disciplinary mode of surveillance,

intervention took place after the individual had shown misconduct. In contemporary modes of surveillance, the focus is more and more placed on the scanning for potential threats, for potential misconduct, using algorithms to calculate and set up risk profiles in order to prevent something from happening, or to manage the consequences whenever something does indeed happen. Within this system, an important shift has taken place: rather than verification preceding judgment, it is now the other way around. Judgment precedes verification: an individual, based upon the results of the continuous logging of one’s

everyday behaviour, can be deemed a threat for national security regardless of having acted in a way that justifies this suspicion – as was the case with Robert McDaniel, discussed in the introduction. In an earlier phase of surveillance studies, the role of the data double in

(16)

determining the life chances and choices of the individual was a major concern (Haggerty and Ericson 2000; Lyon 2001), and after its magnification through big data strategies, these concerns should be taken even more seriously (Lyon 2014, 6). When one can be judged based upon his or her datafied self, without the need for actual interaction with the flesh and blood individual that is linked to that data, what happens to the relationship between the self and the datafied self?

1.2 The Data Derivative

Different analyses have been conducted of the use of personal data in the context of national and border security. A very important contribution is Louise Amoore’s

conceptualization of the data derivative, which she proposes in her analysis of the eBorders programme in the United Kingdom, a risk-based system that employs processes of data mining and analysis to flag potentially ‘risky’ individuals entering the country (Amoore 2011, 26). Her concept of the data derivative refers to the aggregates of data collected on the individual that enable a specific form of abstraction employed in current risk-based security calculations – a system that acts on and through people and populations in new manners (Amoore 2011, 26). The data derivative – similar to both Deleuze’s concept of the dividual, and Haggerty and Ericson’s data double – is formed by individuals first being broken down, being disaggregated, after which they are re-aggregated, this time through algorithm-based association rules through which a risk score can be calculated and connected to the

individual (Amoore 2011, 27). Where most of the conceptualizations of the datafied self are understood as some kind of representation of the individual in data, the data derivative is different. According to Amoore, the data derivative’s key focus is not placed on who the individual is, and not even on what the data can say about the individual. Rather, the data derivative is centred around who the individual might be – “on our very proclivities and potentialities” (Amoore 2011, 28). The data derivative is the product of a system of pre-emption, which aims at the “making present of future consequences” (Amoore 2011, 29) – or, as Massumi put it: how “pre-emption brings the future into the present” (2005, 8). This system of pre-emption is not merely aimed at predicting the future. Where in, for example, strategies of data processing focused on marketing, the goal is to determine the future

(17)

desires and needs of potential consumers of a certain product so that they will eventually buy it, in this context of border security, the future consequences of an eventuality are made present – of an eventuality that may or may not occur. What is different here, is that the system of the data derivative is indifferent to whether this eventuality actually occurs or not. Its aim is not to predict the future, but, according to Amoore, rather to be capable to act in the face of uncertainty – to render the data actionable. Rather than a prediction, it is a

projection (Amoore 2011, 29) – a projection that simulates the potentialities of the

individual into the present, through its data derivative. Moreover, Amoore states that the data derivative does not seem to care about who you are, but rather about what you are capable of, what kind of actions you might carry out – and when these proclivities and potentialities seem to indicate behaviour that could form a threat, the system of the data derivative allows for preventive intervention. This focus on temporality is distinctive for the data derivative: it makes present all the future potentialities of the individual, regardless of whether they occur or not. By defining the individual behind the data derivative by its potential futures, they become realized, in the sense that these potential futures have consequences for the individual in the present, without a necessity for their actual

occurrence. These future actions of the individual are treated as if they are real, and in this sense, they become hyperreal: they do not even have to take place in order to be significant, in order to be acted upon. It is thus not only about the capacity to act when something might actually happen, but also about already having acted – regardless of the occurrence of a certain eventuality.

As stated before, the data derivative is used in order to ‘flag’ individuals that may be a risk to national security. Huge aggregates of data are used to establish a curve of what is ‘normal’ or ‘safe’ behaviour to which, consequently, the data derivatives can be tested in order to detect individuals that are deviant from this norm. The data derivative becomes actionable through searching across items of data looking to find correlations (Amoore 2011, 30), through which a risk score can be calculated. Certain combinations of data about a specific individual may lead to an increased risk score. For example, some factors that in combination may lead to an increased risk score in the context of border security may be someone’s nationality, past travel destinations, mode of payment (cash, or by a third-party, e.g.), meal choice, tickets booked but not seated together, et cetera (Amoore 2011, 27). In other words, this system is based upon both an ontology of association, as well as on

(18)

discrimination: when certain associations can be made between data that are deviant from ‘the norm’, one may receive a higher risk score. This system enables security workers to judge an individual merely based upon data that is collected on them, without having to interact with the actual flesh and blood person behind it. However, these interactions with the virtual data derivative can have some very real consequences. For example, they may lead to the individual being subjected to additional security screenings, or to problems with visa requests, and even to innocent people becoming suspicious merely because of certain associations made between datasets collected on them.

1.3 The Hyperreality of the Data Derivative

The first step in understanding the relationship between the datafied self and the individual within this context is thus to recognize that the data derivative, compared to the other two types of datafied selves discussed in this research, is the least interested in forming a

datafied personification or representation of the individual, nor should be understood as the ‘electronic footprint’ that the subject leaves behind as a digital residue. Amoore states that it is a means, a means of “dividing, separating and particularizing subjects”, to bring them to attention and to be able to make them into the subjects of interest (Amoore 2011, 35-36), thus focusing on its actionability – its efficiency as a tool for strategies of national security. However, this approach is still not completely distanced from an idiom of representation since the data derivative is understood as a separate entity that is established through the actions of the individuals, and offers insight into them. This research proposes an alternative approach to the understanding of the relationship between the subject and its data

derivative, that places emphasis on the increased reality of the data derivative and how it fundamentally alters the disposition of the individual – complicating the causality inherent to a representational relationship between an offline individual which can be known through its datafied self. Additionally, this alternative approach emphasizes how the derivative is more than a means, more than mere pragmatics, since it has become part of the individual him- or herself under databased surveillance. This is where the Baudrillardian logic of simulation becomes relevant in understanding how, if not as a representation nor as just as a means, the data derivative should be understood. Baudrillard states that representation

(19)

itself has become replaced by simulation. The equivalence of the sign and the real and the causality between signifier and signified, characteristic to a relation of representation, no longer exist in simulation (1981, 31). This causality becomes obsolete because the

simulacrum is defined by its hyperreality. This means that the simulacrum does not need the verification of an external reality in order to be real: its reality becomes self-referential. In other words, the signs do not refer to something in reality, but rather refer back to

themselves. The data derivative does not aim to represent the individual. Rather, when it is acted upon, the data derivative functions as a substitute for the individual: it is targeted instead of the individual him or herself. In this sense, a relationship between the two is simulated: the data derivative presents itself as if it were its ‘original counterpart’, while there is no clear relation of signification between the two. This idea will become more concrete throughout this chapter.

In these instances, in which the data derivative is acted upon in the context of national and border security, the datafied self comes to precede its original. Most of the filtering and analysis processes take place prior to the actual identification of individuals (Amoore 2014, 111). The datafied self is judged before the individual is actually known or has done something to ‘deserve’ a higher risk score. Baudrillard states that “simulation is characterized by a precession of the model, of all the models based on the merest fact” and that these models constitute the event (1981, 16). “The facts no longer have a specific trajectory, they are born at the intersection of models, a single fact can be engendered by all the models at once” (Baudrillard 1981, 16-17). From this Baudrillardian view in which reality becomes increasingly shaped not by facts, but by models that precede and shape reality, the deceptive nature of the data derivative can be understood more properly. The data

derivative does not deal with facts; it deals with hypotheticals. What are perceived to be, and are acted upon as facts, are merely the result of models – models that calculate the likeliness of an individual being a potential threat, derived from aggregates of data that have been instilled with hypothetical meaning. Hypothetical, because the meaning given to certain associations between datasets are based upon generalization, bias and speculation: when someone’s books a, for example, halal meal with their flight, this is registered and can together with someone’s nationality and travel history lead to a potentially higher risk flag. What is crucial to understand here is that these hypotheticals are also based upon practices of making meaning without a clear relation of signifier-signified: someone’s meal choice

(20)

does not have any real relation to someone being a potential threat – it merely presupposes somebody’s likeliness to be of a certain religion, which once again does not equal being a threat to national security. It is only within a specific context that a specific meaning is

ascribed to for example someone’s religious beliefs: not based upon facts, but on

contextually bound preconception. The model, with its own facts and truths, is taken up as legitimate and as real in its use. However, there is a “confusion of the fact with its model” (Baudrillard 1981, 17). These models that are not based upon facts, produce their own self-referential facts that do not need an external reality in order to be true. This is the

precession of the model, that engenders events in reality.

In this way, the individual takes on a secondary place in relation to its data derivative. Amoore conceptualizes the data derivative as a combination of elements of possible

associations, links, and threats that, due to its emphasis on pre-emption and prevention, does not care who the individual was or is, but rather what he or she may become in the future – on the capacity to act and intervene when necessary. However, as discussed, this emphasis on the individual’s future state may also alter the present of the individual significantly. Additionally, though Amoore argues that the data derivative is not a

representation, she does understand it as producing some kind of insight in, and knowledge on the individual that is taken up as legitimate and that can be linked back to specific

individuals. She understands the data derivative as produced by data that is generated by the individual, as its ‘original’ counterpart. If not as a representation nor as a mere tool, how then can the datafied self within a context of pre-emptive strategies of national security be understood, taking into account not only its focus on the individual’s potentialities and proclivities, but also on how it impacts the individual him or herself in its present and in its very disposition – engendered by these models defined by their self-referentiality?

As argued, the data derivative may be more productively understood through a logic of simulation, because it is precisely in the instances in which the data derivative is put to use that it becomes real and gets its meaning: when someone’s data derivative is employed, for instance, before the individual that it is connected to wants to go through customs – these aggregates of data are handled as if they are true and as if they are a sufficient

approximation of the individual upon which real decisions can be based. In other words, it is in these instances that the datafied self comes to substitute the individual. As these data become increasingly real, the individual’s reality becomes increasingly irrelevant. In his

(21)

chapter “Clone Story” from Simulacra and Simulation, Baudrillard talks about the process of cloning and he states that it puts an end to totality. What he means by this is that, when looking at the genetic code of a being, just a small fragment of code can offer all the

information necessary to understand the body. If just a fragment of this code contains all the information there is to understand a body, then the whole body, the totality, becomes irrelevant – it loses its meaning (Baudrillard 1981, 97). Although the data derivative does not offer all the information there is on the individual, this view can be used to illustrate how the bits of information that are available here, within this context, are sufficient in order to judge the whole of a person on whether they should be subject to restrictive measures or

additional surveillance – whether they are suspicious. The totality of the individual here becomes irrelevant, in the sense that the data derivative does not care who the individual is, or whether they actually commit a crime, but it does substitute itself for the individual as it becomes the prime point of access in processes of pre-emptive surveillance.

Within this Baudrillardian logic of simulation, ‘real’ facts no longer matter (Baudrillard 1981, 31), because the model precedes, and engenders events in reality (Baudrillard 1981, 32). In the context of the data derivative, the model consists of the risk profiles based upon aggregates of data collected on thousands of individuals, that indicate which correlations between specific datasets would indicate an increased potential for risk. When the model is taken up as being real (which is the case here, since the risk profiles are used in order to screen for threats) and it indicates that somebody is a potential threat, it does not matter whether this person actually turns out to be in violation of the law, as the outcome of these processes can be understood as being the same. This becomes clear through Baudrillard’s example of psychosomatic disorders which he uses to describe the complexity of the distinction between simulation and reality. He states that there is a difference between someone who fakes an illness, and someone who simulates an illness. When someone fakes being ill, he can stay in bed all day to make others believe that he is actually ill. This is a deliberate effort, which does not result in this person actually feeling ill. However, when one simulates an illness (which is what happens with psychosomatic

disorders), when someone believes that he might be ill, he can actually produce some of the symptoms in himself (Baudrillard 1981, 5). In other words, the different symptoms can be understood as signs, that supposedly signify something: in this case, a certain disease. However, the symptoms can be there without the person actually being ill – and in this

(22)

sense, the reality of this situation can be understood as being the same as if the person were, in fact, ill: the symptoms are still there. Therefore, there is a blurring of the

distinctions between true and false, between real and imaginary, and between reality and representation (Baudrillard 1981, 5). It is not about simulation imitating or masking reality, but rather about simulation substituting reality. The same goes for the models that

determine what makes an individual ‘risky’: they consist of all these predetermined

correlations and characteristics that have been instilled with meaning. They are taken up as legitimate, and when they match with one’s data derivative, one can be labelled as

suspicious, regardless of whether they actually have committed a crime or will do so in the future. In other words, the ‘symptoms’ are already produced in the individual due to the interaction with his or her datafied self – exemplifying the self-referentiality of the model and the changed position of reality in these processes.

1.4 The Data Derivative as an Operational Prosthesis

What can thus be concluded is that the data that is used in these strategies of national security is performative: “the composition of flecks and bits of data into the profile of a terror suspect, the re-grounding of abstract data in the targeting of an actual life, will have the effect of producing that life, that body, as a terror suspect” (Raley 2013, 128). In other words, when somebody’s datafied self is flagged as suspicious, this may actually lead to a person in ‘real life’ becoming suspect, even when they were not before, and there were no cases of misdemeanour on behalf of this person. The relationship between the individual and what is understood here as the data derivative is thus a rather complex one. Within this form of security, this system is “less interested in who a suspect might be than in what a future suspect may become; less interested in the one-to-one match of the watch list or alerts index database, and more interested in the signals of real-time predictive analytics” (Amoore 2014, 109). Again, Amoore thus states that the data derivative is not a

representation of the individual but that it should be understood as a means, a means for rendering individuals into subjects of interest (Amoore 2011, 36). This research proposes an approach for understanding this type of datafied self that focuses more on how it inherently

(23)

that are increasingly replacing the understanding of reality. The data derivative functions more as what Baudrillard terms as an operational prosthesis of the individual. As mentioned earlier, in “Clone Story”, he argues how the whole of a body loses its meaning since a mere fragment of the individual’s genetic code can offer all of the information necessary for understanding it. He looks at the genetic code of a being and states that cells, as fragments of the whole, contain all of the information on the individual – and that they become

prostheses of the body that enable its indefinite mechanical reproduction (Baudrillard 1981, 98). The same could be said for the personal data of the individual within a context of surveillance. The data derivative forms a cybernetic prosthesis that contains all of the information necessary for interacting with the individual within this context of databased surveillance. It renders the individual operational: the individual becomes transparent and his or her potentialities and proclivities are projected into the present. Because of this, the individual becomes more efficient to handle as the data derivative can be targeted as a substitute for the original, referring to the etymological meaning of the word prosthesis: a substitution. The data derivative can be viewed as an “abstract and autonomized part of a whole” that “becomes an artificial prosthesis that alters this whole by substituting itself for it” (Baudrillard 1981, 98). As an operational prosthesis, the data derivative does not only provide access to the individual, but it also alters the whole of the individual. On the one hand, it makes present future potentialities of the individual as if they had already

happened, thus shaping the individual’s life chances and choices. On the other hand, as an addition to the individual, the data derivative renders its disposition transparent, knowable, shareable, actionable – making the subject of these strategies of dataveillance an inherently different entity as it would be outside this context of databased surveillance.

Understanding the data derivative not as a representation, nor merely as a means, but rather as an operational prosthesis emphasizes how it has become part of the individual: as an addition that has altered the entity of the individual as a citizen, by rendering it

transparent and actionable. This prosthesis is one that cannot be cut loose from the

individual under databased surveillance and has thus, on the one hand, become naturalized as part of the individual: it is always there, without much that the individual can do about it, and people have become more or less accustomed to it; accustomed to the processes of the collection and analysis of their personal data. Simultaneously, the approach of the data derivative as a prosthesis underlines its artificiality. It is not a natural part of the individual. It

(24)

is not ‘real’ – not even as a representation of the individual’s reality. Rather, it is established through models that produce their own self-referential facts that do not need a relation to an external reality in order to be true as they are taken up by these processes of databased surveillance and have become actionable. The data derivative in this sense has become hyperreal because of the self-referentiality of the ‘facts’ it consists of, and its efficiency as the main access point for strategies of pre-emptive databased surveillance, while the reality of the individual it is linked to has become more or less irrelevant within the context of these practices. The data collected on the individual is decontextualized and instilled with meaning produced through associations between datasets. From this point of view, the data

derivative can thus not be understood as based upon the individual as an actual person. Because of this, the risk-calculating algorithms can already produce symptoms of, for example, suspicion, in the individual, regardless of actual misdemeanour. Within this

context, the self-referential models can come to substitute reality, and the data derivative as an operational prosthesis can come to precede the individual it is connected to within his or her interactions with strategies of national security.

2. WHO? SURVEILLANCE CAPITALISM AND THE MOULDING OF CONSUMERS

(25)

The transition from disciplinary society to the societies of control is inextricably intertwined with a restructuring of capitalism. Where previously profit was made by the individual performing manual labour, in most developed countries, the industries have shifted towards being more centred on ideas, creativity and communication. Manuel Castells describes how through a major social, technological, economic and cultural transformation, a new type of society came to rise: the network society (Castells 2010, xvii). Within this society, labour markets and work processes have become drastically altered by information and

communication technologies (Castells 2010, xxiv). Castells argues that capitalism has

restructured itself through these technological innovations and organizational changes, and that this has led to what can be understood as informational capitalism (Castells 2010, 18) – signifying a discontinuity in the material basis of economy, society and culture (Castells 2010, 29). Absolutely central within these transformations are the technologies of

information processing and communication. Information itself has turned into the product of informational capitalism (Castells 2010, 78) and the technologies that are central in this research thus form a fundamental part of its production process.

This shifting emphasis onto information corresponds to the shifts that have taken place within the systems of surveillance. In disciplinary society, the individual was subjected to processes of discipline through the use of long-term surveillance aimed at making the body docile, a characteristic of industrial capitalism. But in the societies of control whose infrastructure is that of networks and computing technologies, this emphasis on information rather than on physicality comes forward in the way in which the individual as a person and body is no longer the subject of surveillance, but how this has shifted onto their

representation in information: onto their dividual (Galič et al. 2017, 20). Interestingly, it is the individuals themselves that, through their everyday interaction with digital technologies, produce the data that constitutes these representations, blurring the lines between

production and consumption, as famously discussed by Alvin Toffler and his notion of the

prosumer (The Third Wave: The Classic Study of Tomorrow, 1980) and, more specifically

connected to digital technologies, in Axel Brun’s conception of produsage (Blogs, Wikipedia,

Second Life and Beyond: From Production to Produsage 2008). The specific type of labour in

which the users of digital media freely produce the products of informational capitalism’s knowledge economy has also been discussed through conceptualizations of immaterial

(26)

2013) – constituting the lucrative business model of extracting surplus value from the free ‘labour’ as performed by the users of digital media that has become a defining characteristic of informational capitalism.

Within the field of surveillance studies, a closer look at the ways in which the individual is monitored and how that benefits the information economy offers insight into the ways in which the relationship between individuals and their datafied selves becomes increasingly complex due to the actionability of the datafied self and the use of predictive analytics. One of the most important sources of revenue in informational capitalism is the trade in personal data collected on the user. These data are sold in order to enable more personalized types of marketing as marketing becomes increasingly based upon consumer profiling and surveillance (Koskela 2012, 52). Because of the insight this data offers into the population as consumers, the market is no longer unknowable (Galič et al. 2017, 25). By knowing the population, advertising companies can target subjects individually and

personally in order to bring specific products together with likely consumers of that product. This knowledge of the market also enables the production of new consumer goods,

specifically for a certain target audience, knowing that there will be a demand for them. The datafied selves that are constituted through one’s everyday behaviour online, within the context of marketing are treated as if they are representations of the individual. These consumer profiles are valuable because of the very fact that they are understood as offering insight into the individual and his or her personality, taste, needs and desires. This

knowledge of the market and the population is based upon the belief that the individual can be known through their personal data, at least well enough in order to approach them with personalized advertisements and to produce consumer goods for specific target audiences. This knowledge of the market and the population also signifies another shift in the logic of accumulation. Shoshanna Zuboff talks about this new logic as surveillance capitalism – a new form of information capitalism that aims to predict and modify human behaviour in order to produce revenue and to gain market control (2015, 75). This prediction and modification of human behaviour is what especially makes the relationship between the datafied self and the individual within this context so complex, since it is here that questions of

representation, authenticity and reality come to the fore.

As stated earlier in relation to the shift from disciplinary society to the societies of control, in surveillance capitalism, subjects are not viewed as (physical) individuals. Similar

(27)

to the Deleuzian dividuals, they are rather understood as patterns and propensities of behaviour that are calculated from large datasets. Companies use these technologies to tap into the individuals’ consumption patterns and inclinations to act in a certain manner (Palmås 2011, 347-348). Through everyday online interactions, a datafied self is established and that entity acts back on those associated with the specific datasets, on the individual behind the data, telling them who they are, what they should desire or hope for and who they should become (Lyon 2014, 7). In other words, it is not just about knowing who the individual is in real-time. What simultaneously happens here is a specific kind of shaping or conditioning performed by algorithms (Lyon 2014, 7). Once again, there is a system of pre-emption at work in surveillance capitalism, not merely in the calculation of future needs and desires of the individual as a consumer, but also in another aspect. As discussed in the first chapter, the data derivative’s pre-emptive character is one based on restriction. It is a long-term type of surveillance that is more indifferent as to who the individual is, but that is rather interested in calculating the proclivities and potentialities of specific individuals, enabling intervention and repression when one might become a threat. In contrast, the datafied self under surveillance capitalism has a more positive character. Its pre-emptive character is not based upon restriction, but rather on a more positive feature – positive in the sense that surveillance capitalism has a productive disposition. Firstly, through an after-the-fact reconstruction of behaviour, habits and actions, collected through the monitoring of the individual, consumer profiles are constructed (Galič et al. 2017, 22). Then, through the use of the statistical power of large numbers, new knowledge can be created based upon what is already known about the individual, in order to grasp fragmented details of the individuals’ lives (Lyon 2014, 7). Through predictive analytics, this system aims to predict human behaviour. Similar to the data derivative, it focuses on who the individual may become and what his or her desires and needs may be in the future, so that marketing strategies can be specifically targeted towards individuals with particular futures. What sets the datafied self apart within this context is how its actionability can be used in order to

modify future behaviour. By establishing consumer profiles, and focusing on future needs

and desires of consumers, marketing cannot only be used to tap into those proclivities, but can also instil new needs and desires, and thus shape the future choices and behaviour of the individual. This is where surveillance capitalism’s productive disposition comes into being: instead of making bodies docile or tracking one’s behaviour for enabling the

(28)

possibility to intervene in the event of a potential threat, what is happening here is the

moulding of consumers through the interaction with the individual’s datafied self that has

come to precede the actual individual (Galič et al. 2017, 20). Rather than adjusting

production to changing consumption patterns, marketing through personal data collapses the division between production and consumption. Due to the increased automation of data processing strategies, surveillance-based simulation no longer merely entails a means to discipline or control consumption, but also to manufacturing customers as valuable

information commodities (Zwick and Denegri-Knott 2009, 224). In other words, the

aggregates of personal data are used to set up actionable consumer profiles which can be interacted with in ways that may determine the behaviour of the offline individual. The datafied self is thus constituted in a reciprocal relationship, in which the actions of the individual fuel his or her datafied self, but that also feeds back to its users, through

personalized advertising and content, to alter future behaviour which once again informs the datafied self, and so on.

In most of these conceptualizations, the individual linked to a specific datafied self is understood as its original. The datafied selves are viewed as the products of the individuals’ interactions with digital technologies and their reaction to the data that is fed back to them, making the individual the point of departure in these processes. In a sense this is true: the datafied self is constituted through an interaction between the individual and digital technologies, and the datafied self’s value lies in the idea that it offers knowledge on the individual. However, this understanding of the datafied self as the product of a particular type of relationship of representation of the individual, as its original counterpart, is no longer accurate due to the increased ubiquity and naturalization of the processes of datafication of the everyday that have altered our interaction with the world. There are different problems with this view. Firstly, this representational relation implies that the individual can be known through its datafied self. Secondly, this point of view ascribes more reality towards the individual than it does towards the datafied self. Thirdly and finally, it implies that there is such a thing as an ‘original’ within these processes of databased surveillance capitalism. A Baudrillardian framework enables a more suitable understanding of these processes through its shift in the logic of representation, illustrating how it is not just the distinctions between production and consumption that are collapsing here, but also those between original and representation, between object and subject and between reality

(29)

and simulation – in ways that externalize processes of identification as the production of identity is increasingly performed by actors outside of the individuals themselves.

2.2 The Precession of Digital Personifications

To address the problems with the conceptualization of this relationship between the

datafied self and the individual, John Cheney-Lippold’s theory on the algorithmic identity will be used as it offers insight into the processes of identification within the digital realm. These processes of identification often take place preceding the individual’s awareness of them: according to Cheney-Lippold, “we are entering an online world where our identifications are largely made for us” (2011, 165). He argues that one’s online behaviour results in the

production of a new algorithmic identity, an identity that is ascribed to the individual. This new algorithmic identity is established through the networked infrastructure of the internet, in which the user is tracked across different platforms, websites and servers, producing data that is processed by web analytics firms that use this information on the individual and employ algorithms to make sense of this data (Cheney-Lippold 2011, 165). This new

algorithmic identity is established through mathematical algorithms that infer categories of identity upon the individual in order to make the anonymous known. It traces certain statistical commonalities through which it deduces characteristics such as class, race and gender automatically. Simultaneously, through the associations made between datasets, the actual meanings of these characteristics themselves are defined. What is taking place here is the transition of the practice of identification into the digital realm, rendering identification

measurable (Cheney-Lippold 2011, 165). The configuration of identities here is thus no

longer in the hands of those inscribed themselves, but in the hands of those in control of the systems of digital mediation (Zwick and Denegri-Knott 2009, 232-233). Identity is

exteriorized, separated from the interiority of the individual’s consciousness and moved into the realm of the information machines. This enables advertising agencies to manufacture consumer simulations with specific desires and needs already built in (Zwick and Denegri-Knott 2009, 233). Through these processes, meaning is also given to the characteristics of certain personalities themselves. For example, what does it mean to be male? What are things that men are generally interested in? If generally, a lot of men seem to be interested

The Hyperreality of the Datafied Self - The Simulational Relationship between the Individual and the Datafied Self in Contemporary Modes of Databased Surveillance