Abstract
Online tracking in its various forms is a backbone of digitalization that has sparked hopes and fears alike: It opens up new opportunities for users and businesses as it enables individually targeted content. At the same time, the encompassing tracking of often unaware and ill-informed users and the opaque practices of data procession has alarmed critics from multiple sides. How can we better understand but also proactively and constructively shape the emerging Tracked Society? Our special issues seek to shed light on these questions from various perspectives and disciplines. In this introduction, we give a brief overview of the topic in general and our special issue in particular.
Online tracking in its various forms (from simple web cookies to more sophisticated technologies that follow users across different devices and application contexts) is a backbone of digitalization that has sparked hopes and fears alike: It opens up new opportunities for users and businesses as it enables individually targeted content. At the same time, the encompassing tracking of often unaware and ill-informed users and the opaque practices of data procession has alarmed critics from multiple sides. In recent years, tracking-related issues have gained increasing attention by experts but also the public. Heated debates on Big Data, privacy, filter bubbles or the challenges of governing algorithms are now part of regular news. The global crisis connected to the COVID-19 disease has added another layer to the discussion: Now, tracking appears to be a matter of life and death. Governments and health experts around the globe experiment with tracking technology in an attempt to contain the spread of infections. This existential dimension makes the common concerns around privacy and data protection almost look pedantic to some – how can it be justified to reject tracking if it is able to save lives? This type of reasoning is not new. Similar points have been made in regard to Big Data’s promises to improve diagnostics or in the context of law enforcement. Also, the criticism to such arguments has been expressed before: Privacy is an unalienable right and there are disproportional threats of various kinds associated to tracking practices. What may be new about the debate on tracking since COVID-19 are the higher stakes and the increased urgency. Tracking has been framed as a question of life and death before but now it immediately and directly affects billions of people – at least, that is how it seems. This is true for any positive or negative effects of the hastily implemented tracking measures and in fact even for decisions against them, too. With a daily rising death toll, there is simply no time for lengthy and nuanced deliberation. At the same time, what has been decided in a state of emergency is likely to persist much longer as the experience from other historic events such as the 11 September attacks have shown.
But even the pre-COVID-19 discourse on tracking has left many open questions. Since online tracking is vital to digital innovation, it will probably remain a cornerstone of the Internet and beyond, regardless of its associated problems. So far, attempts to restrict it (e.g. through the European Union’s General Data Protection Regulation; GDPR) have been scarce and their effectiveness unclear, leaving the market highly under-regulated (see below).
Therefore, the question arises, how can we better understand but also proactively and constructively shape the emerging Tracked Society? This is necessarily a multi-disciplinary endeavor as so many perspectives are relevant to grasp this phenomenon. Our special issue brings various disciplines and research angles together in an attempt to cover this wide field. We also conducted an authors’ workshop in cooperation with the University of Amsterdam in which the papers were critically assessed by invited discussants as well as the other participants. In this way, the authors were able to exchange their very different perspectives with each other directly face-to-face. Already during the workshop, the wideness and complexity and also the urgent importance of the topic became apparent. There is a multitude of ways we can – and should – talk about the Tracked Society. As in all interdisciplinary endeavors, there is an inherent danger of getting lost on the meta-level, focusing more on how to talk about it than the actual issue itself. Where exactly does the process of tracking start, where does it end? Should we address it from a techno-centric perspective, aiming at clear definitions based on the various tracking technologies involved? Would it be more fruitful to put the practices and interactions of and around tracking at the center of our attention? What precisely qualifies a Tracked Society to carry this label? There is no doubt that these questions are important if we want to gain a nuanced understanding of the Tracked Society. At the same time, there is no doubt that there are dozens of ways to answer them. We neither pretend that we have definite answers, nor do we believe that attempts of finding such answers are not worthwhile. Rather, we aim at respecting the diversity of perspectives and possibilities to contribute meaningfully to this topic. In this way, we hope to provide a wide as well as deep (but necessarily also fuzzy and incomplete) picture of the Tracked Society, suitable to spark further research on the pressing questions emerging in this context. We start with a brief look into the foundations of tracking before we outline some of the intensively discussed associated challenges and end with our specific view on the Tracked Society in this special issue.
Foundations of tracking
While certain tracking technologies may be new, the practice of gathering information about objects, individuals, and their behavior for a variety of reasons is certainly not. In fact, tracking (in this broad sense of the term) seems to be an activity, which is as old as humanity itself: Humans, like all animals, are surveillance machines by the very nature of our sensory organs and self-preserving instinct. Our ability to decipher environmental clues and to narrate events may even be traced to prehistoric hunting [. . .]. (Lauer, 2012: 570)
But also on a societal level, we need to gather and process information about other individuals to be able to interact with them. As sociologist Georg Simmel (1950) has noted, “Obviously, all relations which people have to one another are based on their knowing something about one another” (p. 370). Such knowledge provides us with the ability to make interactions to some extent predictable. In the most primary sense, this is simply a question of life and death: Does the person I am interacting with have the intent and the ability to harm me? On a larger scale, it is a question of power and control. Therefore, it comes as no surprise that forms of surveillance can be traced back to at least as far as ancient Greece. Lauer (2012) even argues, “The history of surveillance [. . .] is as long as human history” (p. 571). For obvious reasons, early surveillance practices such as in ancient Greece were rather rudimentary and relied on unsophisticated methods such as oral reports from informants (Lutz and Lutz, 2018). More complex societies increased the need to monitor behavior and technology provided new possibilities to do so. The field of surveillance studies has focused mostly on rather recent developments tied to advances in information technology, especially since the 1960s. Many researchers from this area regard modern societies as surveillance societies, “[. . .] characterized by increased investments in bureaucracies and techniques to systematically – and over longer time periods – collect, store and use information” (Boersma et al., 2014: 1). While it is not hard to identify such characteristics in contemporary societies, it is less clear when exactly the label of a “surveillance society” is justified and when these transformations precisely took place. Josh Lauer (2012) identified a shift toward an “evidential paradigm” already for the late 19th century when three technologies laid the foundation for extending and documenting communication: photography, the phonograph, and the telephone. Interestingly, some of the core issues we discuss today in the context of tracking have already been subject of heated debates then. Acknowledging the limitations of the format of a special issue but also of what principally is assessable about the Tracked Society, we choose to take the perspective of problem-oriented research (Bechmann and Frederichs, 2005). Thus, we focus on the already observable concerns related to the Tracked Society rather than specific technological questions.
Core issues of the Tracked Society
Privacy
An evident issue now as well as in the 19th century is the threat to privacy posed by media technologies’ immanent ability to decontextualize personal information. However, although many scholars agree that privacy is an important value (and thus, regard its violation as problematic), they have struggled to pinpoint what privacy actually is.
Currently, privacy is a sweeping concept, encompassing (among other things) freedom of thought, control over one’s body, solitude in one’s home, control over personal information, freedom from surveillance, protection of one’s reputation, and protection from searches and interrogations. (Solove, 2008: 1)
Accordingly, there is a wide consensus that media technologies pose a threat to privacy but the views on what exactly is problematic here are equally wide and diverse. It seems to be hard to determine privacy as a value in itself, especially in a (potentially) global environment with all its different cultures. Helen Nissenbaum (2009), therefore, suggested analyzing privacy in concrete social/technical contexts. Instead of alleging a dichotomy between the private and the public, she argues, “something considered public in relation to one realm may be private in relation to another” (p. 215). We willingly present details of our financial life to our accountants but usually would not reveal our naked bodies to them. At the same time, most of us will hesitate to disclose our incomes to our doctors despite their intimate knowledge on our health. Privacy can only be understood through contexts but it would be wrong to limit ourselves to the perspective of particularities: Viewing privacy more contextually alone often fails to provide sufficient direction for making policymaking or legal decisions, which depend upon making generalizations. To have a useful theory of privacy, we must generalize. Too general a theory, however, will prove vague, simplistic, and reductive. Therefore, we must navigate the tension between generality and particularity, between abstraction and concreteness. (Solove, 2008: 48 f.)
The birds-eye perspective enables researchers to gain a more encompassing understanding of privacy. Larger transformations of society become visible, for example socio-cultural questions such as what is regarded as a private and what as a public matter (Westin, 2003). It also allows us to grasp the scale of transformation.
Ubiquity of data
While the “new” media of the 19th century already raised many of the concerns we discuss today (Lauer, 2012), we may wonder whether the sheer ubiquity of data leads to changes that are not only quantitative but also qualitative. Already in the mid 1990s – a time when the Internet was still in its infancy – David Lyon (1994) saw the emergence of a surveillance society which he describes in this way: “Precise details of our personal lives are collected, stored, retrieved and processed every day within huge computer databases belonging to big corporations and government departments” (p. 3).
This characterization could easily be applied to many present-day societies. Yet, all elements in this sentence have to be carefully re-thought from our current perspective: the quality of what is portrayed here as “precise,” the depth of what is seen as “personal,” the size of what is considered “huge,” the extent to which corporate and governmental power is increased. While we may use the same attributes to characterize the developments, we need to also recognize that the relatively basic capabilities and actual practices of surveillance of the 1990s do not have much in common with the encompassing around-the-clock monitoring in the era of smartphones, social media, and the Internet of Things. Many of the heated debates of the last few years would have not been possible without today’s ubiquity of data. No concerns around filter bubbles due to hyper-personalized content, no fear of manipulation of the electorate through micro-targeting, no discussions on influencing citizen behavior through scoring systems and digitally mediated nudging. Again, some of the underlying concerns are not new. For instance, Putnam (2000: 177) already warned of a fragmentation of the public through what he called “cyberbalkanization” at the turn of the century. But at this time, there simply were not enough Internet users and the technology was not sophisticated enough to make this a widely discussed issue. Only the scale and ubiquity of data in combination with technological advances fueled the hopes and fears connected to terms such as Big Data. Today, there is hardly a sector or an aspect of our lifeworld that is not more or less deeply affected by Big Data (Kolany-Raiser et al., 2018), prompting scholars to speak of a “datafied society” (Schäfer and Van Es, 2017). None of this would have been possible without tracking technology.
What makes all of this so attractive to governments and businesses is of course the promise to make behavior of citizens and clients more predictable and thereby controllable. While many critical scholars and activists ring the alarm of the emerging system of “surveillance capitalism” (Zuboff, 2019), others praise the opportunities to monitor “customer journeys” from their initial interest in a product to its actual purchase (Lemon and Verhoef, 2016). Netflix VP of product, Todd Yellin, did not only point out the increased efficiency of behavioral data, he also categorized established demographic factors such as geography, age, and gender as “garbage” to predict movie taste (Morris, 2016). This is in line with Chris Anderson’s bold prediction that Big Data would end the requirement of theory (Anderson, 2008). Why theorizing if a complete picture can be generated through sophisticated data analysis? Critics reminded us that numbers do not “speak for themselves” (boyd and Crawford, 2012: 666) and that we should not be “seduced by the promises of big data to render theory unnecessary” (Graham, 2012).
Nevertheless, no matter how overrated the possibilities of data analysis actually are, companies and governments invest heavily in this technology (lately especially, under the umbrella of Artificial Intelligence). Data are widely regarded as a commodity, often described with metaphors referring to it as a natural resource (e.g. oil). However, such metaphors are not very adequate: Data is not a natural resource that replenishes itself, but in social media platforms it is created by users with intentions entirely unrelated to its use as a valued commodity. It is created by humans and recorded by machines rather than being discovered and claimed by platform providers or third parties. (Puschmann and Burgess, 2014: 1699)
No matter which metaphors may or may not be suitable to characterize the ongoing datafication of society, it is clear that questions regarding ownership, access, and control of the data arise, leading us to the next section.
Power asymmetry
Although users are the main generators of data, they usually do not participate directly in important decisions of the platform providers. Accordingly, there is a massive power asymmetry between data subjects and data processors (including third parties). It is hard to gain a comprehensive picture of the precise actors and their practices. As Bau et al. (2013) have noted, “there are hundreds of companies, dozens of business models, and myriad tracking technologies in use” (p. 1). Tracking reaches from simple cookies implemented to maintain basic website functionality to complex socio-technical systems with the sole intention of creating detailed user profiles for commercial purposes. Many trackers come from the advertising industry, designed to deliver targeted personalized ads. Although there are still countless companies and services, this market has experienced significant concentration processes. To name just two examples, the advertising company aQuantive has been purchased by Microsoft and Google has bought DoubleClick (Schneider et al., 2014: 32). At the same time, the platformization of the web (Helmond, 2015) has transformed the Internet to an infrastructure that allows for easy and far-reaching tracking (for a broader look at the societal impact of platformization, see also Van Dijck et al., 2018). While the early Internet was built around the hyperlink (e.g. as an important ranking factor for Google’s web search), there is a trend toward what Gerlitz and Helmond (2013) have called the “Like economy.” By implementing social media plugins (e.g. Facebook’s “Like” button) on external websites, platforms are able to extend their ecosystem beyond its original borders of the site itself, resulting in a centralizing effect: Simply because the platform can expand some of its key features into the entire web and integrate ever more objects into the social graph, it can recentralise and monetise the created connections and data flows, as they all direct back to Facebook. The dynamics of de- and re-centralisation are not only interconnected, they form a prerequisite for the Like economy. They enable Facebook to maximise its data mining activities while at the same time keeping control over the key entities of exchange—data, connections, traffic and [. . .] user affect. (Gerlitz and Helmond, 2013: 1355)
Through these and other methods, already powerful actors become even more powerful. Moreover, it is increasingly difficult for average Internet users to control the innumerable ways in which they (intentionally or unintentionally) contribute data to the tracking systems and assessing the consequences is even harder.
Of course, these developments do not take place without resistance. Efforts to contain the expanding tracking systems and address power asymmetries reach from informational campaigns, 1 browser plugins to block or obfuscate usage tracking 2 to legislative regulation. For example, the European Union has inscribed a “Right to be Forgotten” and other measures in the General Data Protection Regulation (GDPR). Now the individual Member States are called upon to adapt their laws to the realities of the digital age (Pohle, 2018a: 97 ff.; Pohle, 2018b: 133 ff.). In particular, the regulation of platforms and intermediaries is currently under intense discussion. As in so many places in the legal profession, the respective interests of all parties involved must be taken into account and a regulation must be made in this regard. In addition to the GDPR, an ePrivacy Directive should actually have been adopted some time ago. However, the Member States failed to reach agreement in 2019, which means that many questions about tracking and cookies remain unanswered. Due to the cross-border impact of tracking measures, it is obvious that rules need to be established in a uniform manner across the European Union. A uniform level of data protection is required. Equally important for companies is the issue of a level playing field. However, the arms race between trackers and anti-trackers is rather uneven (Beales and Muris, 2019).
Raised awareness and attempts to block tracking through technical countermeasures are met with increasingly sophisticated innovations, for example, hard to delete “Evercookies” or “browser fingerprinting” techniques which exploit individual browser characteristics for profiling purposes (Peacock, 2014: 6; for a more comprehensive overview of current tracking techniques see Ermakova et al., 2018: 4735 ff.). Regulations such as the GDPR require the users’ informed consent to tracking practices but the legally established way to achieve this is flawed: Users are supposed to confirm the privacy policies and terms of service when they sign up to a new platform but empirical insights reveal that most of them spend little or no time with these documents (Obar and Oeldorf-Hirsch, 2018; Steinfeld, 2016). Thus, the phrase “I agree to the terms and conditions” has also been called the “biggest lie on the Internet” (Obar and Oeldorf-Hirsch, 2018). Already before the big hype of social media and smartphone apps, it was found that an average Internet user would need to devote about 40 minutes per day just to study the terms of service they agreed to (McDonald and Cranor, 2008: 560). In the age of data ubiquity, this number must have increased significantly. Moreover, regulations such as the GDPR might actually be counterproductive as they seem to result in even longer legal documents (Becher and Benoliel, 2019: 11). Given the monopoly-like situation in data-driven markets such as social media platforms and the urge to participate in them, it has also been questioned how much of a choice users really have. Peacock (2014: 7) has argued that the relationship between users and data-driven corporations can be characterized as “unconscionable contracts” which are “exploitative, unjust, unavoidable and put the burden of an economic transaction wholly on one side, and in this case the online user”. Whatever the next steps in these arms race, it seems unlikely that they will turn the current power asymmetry in favor for the data subjects.
(Mis-)Representation and discrimination
The concerns addressed earlier lead directly to another major problem related to tracking practices: Powerful actors gather detailed information about individuals as well as groups across different contexts, and thereby, gain the ability to impact their lives severely. Many scholars have pointed out that the practices by platforms and other important actors are not neutral but serve particular functions, mostly toward capitalist ideologies (e.g. Fuchs, 2011; Mager, 2012; Zuboff, 2019). Hence, there is an ongoing concern that—contrary to earlier visions of the Internet—the Internet may contribute to more inequality, especially as processes become automatized and thereby scalable (Eubanks, 2018). At the same time, most of the relevant decisions are black-boxed, and therefore, hard to notice and even harder to control (Pasquale, 2015). In this context, data subjects from vulnerable groups face a dual risk of being either made invisible or misrepresented as Taylor (2016) has argued. Tech companies have used the Global South as a large experimentation ground and local populations have been subjected to especially deep privacy infringements (Arora, 2019).
Various attempts have been made to address these different concerns. Suggestions reach from often vague ethic guidelines to more concrete measures such as algorithm audits (Sandvig et al., 2014). Many call for more fairness through a form of “data justice” (Dencik et al., 2016; Taylor, 2017) or “design justice,” for example, by making decision-making processes more open and participatory (Costanza-Chock, 2018). What exactly is regarded “problematic” and how it should be tackled is of course just as controversial as the larger debates around social justice and other related topics. Depending on political and cultural leanings, the arising questions can be answered very differently. Some may seek for extended governmental control; others will regard such measures more as a problem than a solution. This also reminds us that the impact of tracking technology is not unidirectional or deterministic. The relationship between technology and society is now widely understood as co-evolutionary and depending on the concrete context, technologies may have very different societal impacts (Dolata, 2013).
Overview of this special issue
This brief introduction gives an impression of the diverse and complex problems the Tracked Society faces. Obviously, our special issue will not be able to cover all of these concerns, let alone those beyond of what has been described here. It should be clear by now that tracking technology has already left a wide as well as deep societal impact. It is equally obvious that society itself shapes this technology. Many of the topics we glanced at are subject of heated and ongoing debates among experts as well as the public. We do not claim to reinvent what has been discussed under related terms such as surveillance, datafication, and platformization. Rather, we aim at connecting and contributing to these discourses.
Nevertheless, we believe that a focus on the specific technology of tracking and its societal impact is justified and fruitful. Although online tracking builds the foundation of today’s digital economy, it remains poorly understood. Tracking is embedded in wider societal practices (e.g. surveillance, trade) and in larger socio-technical infrastructures (e.g. platforms) but to develop a nuanced understanding of its impact, it seems worth to focus particularly on this technology and its applications.
The papers assembled in this special issue tackle various aspects of the Tracked Society. Rasmus Helles, Stine Lomborg, and Signe Sophus Lai provide an overview of the ecology of third-party services across top sites in the EU. Mark Rosso, ABM Nasir, and Mohsen Farhadloo take a macro-economic look on the question whether the revelations by Edward Snowden have led to an increased usage of Privacy Enhancing Technology (in this case, the search engine DuckDuckGo which avoids the tracking practices of its competitors). Nadine Bol, Joanna Strycharz, Natali Helberger, Bob van de Velde, and Claes H de Vreese focus on the question who actually gets tracked with regard to socio-demographics such as gender and education level, thereby challenging the notion that these categories become less significant in the context of individualized tracking practices. Elena Maris, Timothy Libert, and Jennifer R Henrichsen also remind us that certain demographics are especially vulnerable to tracking, in this case with a closer look at pornography websites whose tracking activities may allow for identifying individuals’ sexual preferences. Luke Stark conducts a Values in Design analysis of a rather under-researched and relatively novel field: mood tracking. Finally, Johannes Breuer, Libby Bishop, and Katharina Kinder-Kurlanda direct our attention to how tracking may be used for research purposes through public-private partnerships.
The latter perspective also opens the debate for a rather rare point of view: While criticism of tracking is vast and frequent, there is also no doubt that it could be useful beyond its current predominantly economically motivated purpose. In any case, tracking will likely remain a cornerstone of the digital economy. Therefore, researchers should not only study its impact from a critical perspective, but also how it may be utilized in a way that serves the common good. The current debates around tracking as a measure against COVID-19 have only underlined this.
Footnotes
Acknowledgements
The editors of this special issue would like to thank Richard Rogers, Anne Helmond, Fernando van der Vlist and Esther Weltevrede from the University of Amsterdam for their support. They are grateful for all contributing authors as well as for the discussants’ critical insights at their associated workshop. Moreover, thanks go to Tim Bornkessel and Nils Wehkamp for their administrative help.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This special issue was edited within the framework of the ABIDA project (Assessing Big Data). The ABIDA project is a research project funded by the German Federal Ministry of Education and Research (funding code: 01IS15016A-F). The content of the special issue reflects exclusively the results and views of the authors. These do not necessarily correspond to those of the ministry or the individual project partners.
