Introducing datapublics

As we walked past the Parliament Christiansborg in central Copenhagen one dusky night in October 2020, scattered groups had gathered holding handwritten signs like ‘my body, my choice’ and were clapping instrumentally to the banging of pots and pans. The signs and banners they held protested COVID-19 policies; however, they also revealed the many issues and stakes among the protestors. Some felt forced to vaccinate, some were angry at the government’s handling of the crisis, and others had shown up because they felt a loss of control over their lives. There were even some who did not believe the pandemic even existed. There was little sign of any journalists covering the events. Nonetheless, as the pandemic continued and the movement gathered in strength, with a small fringe resorting to more violent demonstrations, the media started taking an interest. They covered demonstrations and did portraits of conspiracy theorists or anti-vaxxers. On Twitter, people gathered in so-called ‘hashtag publics’ (Bruns and Burgess, 2015) to criticize or praise a programme produced on conspiracies and the people who believe them. Meanwhile, fact-checking initiatives rivalled to verify the statements on both sides of the divide with the help of flagging from Meta’s content moderation algorithm used on Facebook. COVID-19 sceptics found each other in ever-larger numbers, but increasingly gathered in closed groups and subgroups on Facebook and other social media platforms. They slowly developed their own secret language, which constantly evolved to avoid being caught by an algorithm that scanned millions of text bites every minute across the global network to automatically recognize and red-flag words like ‘covid’ or ‘vaccine’.

Just some months before this happened, something seemingly unrelated to the spread of the deadly pandemic happened. This event also raised a few eyebrows across the planet, primarily among data scientists and tech innovators. The San Francisco-based artificial intelligence (AI) research laboratory OpenAI created the third generation of an algorithm based on an autoregressive language model that uses deep learning to produce human-like text. The quality generated by GPT-3 was so advanced that it was difficult to distinguish it from a text written by a human. Because of its possibly dangerous effects, Microsoft announced on 22 September 2020 that it had licensed the ‘exclusive’ use of GPT-3. Others could still use the public Application Programming Interface to receive output, but only Microsoft has control of the source code. The company was worried about what the algorithm could be used for – and whether it had created a Frankenstein of Silicon Valley that could turn evil, corrupting and disrupting democracies with unstoppable consequences. The next generation of the GPT-3 – the ChatGPT launched in 2022 – gained much more attention for its ability to create human-like songs, texts and assignments, and it is feared how it will impact huge domains such as the music industry, media and education.

Many such AI models developed by OpenAI, Meta, Google and other similar providers of so-called large language models are not only being used in the battle against COVID-19 misinformation, flagging content produced by anti-lockdown protesters and others, but increasingly also implemented by news organizations. By using AI, news organizations hope to better distribute their content to users and reach new audiences by linking large amounts of data content with large amounts of audience data. In doing so, the traditional media also enter into a race with Big Tech companies to define the ‘public’ and cultivate publics through increasing amounts of data on these audiences.

What we can observe in these tales of turmoil in modern democracies and the race to the datafied forefront are the manifold civic practices acting and re-acting as publics in both mediated and non-mediated modalities and the massive media infrastructure supporting these practices and flows of data. The tales also highlight how democracy is fragile and susceptible to transformation, which, depending on the view and the political position of the interpreter, can be seen as either damaging, a crisis or a global moment of reckoning.

Whatever the normative position, these events are signs of publics in transformation and beg the question of how datafication processes play a role in those transformations happening in and out of data flows. From data in the form of public statistics on vaccinations, MeToo hashtags and comments to the metadata inherent in pictures of the ‘Make America Great Again’ movement’s storming of Capitol Hill, everything is used to train large language models to develop AI products and services. Subsequently that data is fed back to citizens as representations and mirroring of themselves via metrics, such as likes, clicks and shares. In turn, users respond to this mirroring as they navigate the vast woodlands of data and form in publics around various clusters of datapoints, via hashtags or debates on social media. The examples also serve to testify to the importance of researching how datafication transforms both citizens’ ability to form publics, the foundations for civic engagement practices, and the role of the media and technology in those processes in all modalities across the hybrid media system.

This book investigates datapublics by asking how the formation of publics is challenged, formed, cultivated and transformed in datafied democracies. In this context, datapublics should not be understood as yet another public concept – this time with ‘data’ added – but as a lens through which we can examine datafication processes and how certain publics are constructed in and out of data flows.

We argue however, that publics are also constructed through the agency and civic practices of users and citizens as they navigate the hybrid media environment (Chadwick, 2011). As we trace how logics of datafication are transforming practices of media and journalism, while also profoundly affecting the many ways in which citizens engage in datafied democracies, this book bridges the gulf between audience studies and media production studies.

The book also considers the dynamic interactions between different aspects of data and how they intersect at disparate junctures with the public. Importantly, we bring nuances to discussions of datafication and its impact on modern democracies by not taking publics or datafication for granted but examining them through historical, cultural and sociological, and techno-materialist approaches. In doing so, we also reveal some of the discourses and normative assumptions regarding what publics are, highlighting the discursive dimension of public formation and the role of technologies in the shaping of those discourses. For example, we show how technologies and publics are imagined and implicitly seen as inherently linked to discussions of specific ideals of democracy.

As a highly digitalized society characterized by high internet penetration and a traditionally strong media system, Denmark represents an exemplary case for many of the explorations in this book. While the Nordic region is our point of departure, we do venture on journeys taking us afar to other media systems, inquiring for example into global infrastructures of data, albeit from a Nordic media perspective. The purpose here is not to present empirical data from Denmark as a singular case, although the book does provide a unique contribution on this ground, but to study the many junctions between data and publics from a variety of perspectives that speak to each other and shed light on a complex, yet interrelated, set of phenomena.

Datapublics beyond the rise and fall narrative

The coming of new technologies and their role in shaping publics has been met with both hopes and concerns, both by the media and in public discourse, for example praising the discursive and mobilizing effects of hashtags for both democratizing processes in the revolutions in Iran and Egypt to campaigns like #blacklivesmatter or #metoo. But warnings have also been raised in connection with spreading of disinformation, the storming of Capitol Hill and a rise in alternative, populist news media. This ambivalent relationship between public formation and technology is nothing new. Ever since the dawn of media research, scholars have been drawn to explore the immense possibilities of new technologies or, conversely, to critically examine the potentially harmful effects of the same technologies. In these tales of technologies, there is a constant balancing of utopian and dystopian visions – or in Anderson’s words, a ‘rise or fall’ narrative (2013, p 1007) – of how such technologies are essentially reconfiguring the relationship between human and machine.

The dialectics between dystopian and utopian tales of man versus machine can be found dating back to the invention of the printing press in the 15th century, which on one hand was perceived as a promise to revolutionize access to written texts, while scribes and the Church, on the other hand, feared its impact on their livelihood and culture (Chadwick, 2013, p 28). Equally, the introduction of the World Wide Web and the way it revolutionized news through the transition to online news has been met with high hopes for increased access to information, the possibility of interaction through hyperlinks and improved engagement with news by its audiences. A decade later, online news was considered superficial, clickbait and the monster that would drive readers into the hands of social media for good. Suddenly, the internet was also to blame for the economic problems of legacy media and for falling subscription rates for print media (Domingo, 2005; Møller Hartley, 2011). In current public debates, the word ‘internet’ seems to be used interchangeably with the terms ‘AI’ or ‘algorithms’. However, the tale remains the same. It is a tale that touches upon the ancient conflict of man’s ambitions ‘to achieve mastery over nature’ (Ingold, 2000, p 312) through the use of technology, but simultaneously a tale of how the corruptive powers of the same technology might destroy the very foundation of humans and nature. The corruptive powers, we argue, are currently in much public debate and academic literature linked to the processes of datafication.

Inspired in part by Ziewitz’s notion of ‘the algorithmic drama’ (2016), we have identified a rise and fall narrative regarding the impact of datafication on the formation of publics, a drama comprised around the central narrative that public formation is in crisis, largely due to forces of datafication. The villain, Big Tech (with datafication as its corrupting power), is taking advantage of a passive and defenceless victim, the public. The hero of this compelling drama is of course journalistic media, who battles the villain, protecting the kingdom of democracy. While there is always a grain of truth in all dramas, the narrative is of course too simplistic and obviously misrepresents actual states of affairs. However, good stories have a tendency to be told and retold, attracting attention to the tellers, enchanting those ready to listen. As arguments get told over and over again, they become stable and at times doxic.

The concept of doxa can be helpful to understand the technological drama and the discursive effects it produces. Aristotle usefully distinguished episteme from doxa and endoxa (Aristotle, 1996). The term doxa refers to the domain of opinion, belief or probable knowledge – in contrast to episteme, the domain of certainty or true knowledge. In the tradition of rhetoric, doxa is constructed through argumentation, and can be used in argumentation, and as such also paves the way for certain solutions to the problem defined. When doxa has become more stable we can talk of endoxa, claims that were previously debated but that have now become common sense. The drama surrounding public formation in datafied democracies run the risk of becoming common sense, without solid backing in evidence, and thus this book also aims to influence and inform this debate with nuanced detailed empirical and conceptual analyses. We believe the story is currently being retold in many different versions, yet always involving our three protagonists. We see the story shaping many conversations about the place and role of technology in our societies.

This drama and its characters are strongly represented in public discourse, shaping, for example, policy making and politics, which naturally respond to the narrative of a democracy and public formation in crisis. The story can be heard in the words of EU Commissioner for Competition, Margrethe Vestager (2020), when she says: ‘We have to take our democracy back. We cannot leave it to Facebook or Snapchat or anyone else.’ The drama is also infused by some academic literature that have proved influential for public discourse, for example, the work of Shoshana Zuboff (2019) or the notion of filter bubbles (Pariser, 2011). The drama is often implicitly present in much academic work, or at least not explicitly challenged and hence passively endorsed and reproduced.

All dramas invoke a fight between good and evil and affirm the relevance of morality. But this should not come at the price of oversimplification, ignorance and Manicheism. The drama of public formation, in all its simplicity, presents significant blind spots and further lack crucial connections to wider historical developments on datafication and public formation. This is problematic, because certain solutions are often prescribed as a medicine to safeguard democracy against the problems identified discursively (but not supported empirically) in the public debate concerning technology, Big Tech and datafication. By prescribing inadequate solutions, we might even make things worse, for example by reinforcing old media structures and not recognizing the hybrid nature of today’s media landscape. Responding to this development, we believe that what we need now, perhaps more than ever, is not more technology-blaming for the failures of public formation. Instead, we see a need for empirically supporting our understanding of public formation as it takes place nowadays with the help of media, technology, data and beyond. In the following we expands on the different parts of this drama. We begin with the victims, the passive publics.

Defining datapublics

How we perceive the extent of a crisis of public formation for democracy depends on how we define publics and their importance for democratic processes. Historically, classic theoretical developments concerning the concept of publics emerged in the Lippmann–Dewey debate, centred on Walter Lippmann’s provocateur’s view of the public as an impossibility, or in his own (1993 [1925]) words, ‘a phantom’. Dewey (2012 [1927]) responded with a more positive view of the public as real and as a necessity for well-functioning societies and democracies. Dewey argued that publics were essential to decision-making processes; he went as far as to argue that state action results from the process of organizing publics and, thus, does not oppose publics since they appear when issues are deemed unmanageable by existing institutions (Dewey, 2012 [1927]). Dewey, therefore, had a processual view of publics and saw them as emergent and multiple phenomena formed as part of an entanglement with issues. He argued that the ‘lasting, extensive and serious consequences of associated activity bring into existence a public’ (Dewey, 2012 [1927], p 35). Thus, the public was formed by all those affected by indirect consequences: individuals who became subjectively effective once organized and self-aware and were helped along by modern communicative technology. While we stand with Dewey in recognizing the importance of issues as a shaping factor and publics as processual and multiple, we want to emphasize the role of the media. Further, we argue that publics are not only emerging in the void of state action, but often in tandem with state legitimization of issues and politics in general.

Some 35 years later, Habermas proposed the notion of the ‘public sphere’, which he defined as ‘a realm of our social life in which something approaching public opinion can be formed. Access is guaranteed to all citizens’ (Habermas, 2006 [1989], p 73). Habermas perceived the public sphere as a more stable zone of publicness ‘located between the civil society and the state, grounded in the former addressing the latter’ (Calhoun, 2017, p 24). This sphere, he argued, was made possible by new patterns of social organization that could sustain such a zone of publicness: newspapers, coffeehouses and the state being increasingly more attentive to public opinion, creating the infrastructures of publicness (Habermas, 2006 [1989]; Calhoun, 2017). The public was theorized as a mode of connection rather than as a process and body of people forming around an issue. This mode was characterized by rational debate and the ‘mutual willingness to accept the given roles and suspend their reality’ (Habermas, 1989 [1962], p 131), emphasizing how private interests (including those of the market) were suspended in search of the public good. Habermas’ concept of the public sphere has been critiqued and further developed, including by Habermas (1988) himself, who argued for pluralism and for multiple publics that could advance democracies as long as they would engage in a rational search for truth.

Thus it is now widely acknowledged that there is not one public sphere, but rather multiple publics that are loosely connected and carry the potential to influence each other (Chadwick, 2011), as stressed here in the latest definition provided by Habermas. Furthermore, online spaces such as social media platforms afford high-speed communication and multimodality and are generally structured around weak ties, which enable the transition of communication to higher levels at a more rapid pace than in traditional offline public spheres (Kaiser and Rauchfleisch, 2019, p 242). Nancy Fraser’s (1990) seminal work on counterpublics is an important addition to the analytical approach we take in this book, in that it allows a conceptualization of publics as hegemonic and part of and at times in opposition to wider publics. She argues that counterpublics ‘function as spaces of withdrawal and regroupment [and] as bases and training grounds for agitational activities directed toward wider publics’ (Fraser, 1990, p 68). With the emergence of social media, it is now possible for people to express themselves in public using a range of modalities (written word, spoken word, body language, pictures, video, sound, graphics, memes, and so on) in an ongoing, real-time process. In this process, fragments of communication can be taken up and circulated by others (liked, shared, commented, and so on) and gain a collective force, power and weight by the ‘quantification of displayed adherence’ (Fraser, 1990, p 68). Hence, social media is potentially a space where the public can form without traditional gatekeeping from journalistic mass media (information sharing, identity building, debating, and so on).

To capture some of these digital transformations to public formation, but focusing more on everyday connective action, the concept of ‘networked publics’ was developed (boyd, 2008, 2010; Ito, 2008). boyd argues that networked technologies restructure the public through distinct affordances that shape which forms of participation arise and how participation unfolds (boyd, 2010, p 39). Further networked publics are a form of mediated public, which both share characteristics with non-mediated and other mediated publics (boyd, 2008, 2010). However, they are also unique due to the structural affordances of the networked technologies – persistence, searchability, replicability and scalability – which are not typically present in face-to-face public life or become amplified in comparison to past media technologies. Following the popularity of Twitter as a platform and the increasing use of hashtags, Bruns and Burgess have suggested ‘ad-hoc issue publics’ or ‘hashtag publics’ (Bruns and Burgess, 2015). Highlighting the affective side and discursive elements of social movements forming on social media, Papacharissi (2015), suggests the notion of ‘affective publics’, grounded in empirical analysis of the Arab Spring and the Occupy movement.

The problem with these concepts, as noted by Møller Hartley et al (2021), is that they are to some degree platform-centric, without really defining publics. They add a digital or platform modality, claiming that digital public formations differ radically from offline spaces or to some degree are separate from offline spheres. In this book, we would like to emphasize the links between these spheres, and not focus alone on the digital, just because we are dealing with datafication and publics.

In some of these works, we have noticed that the concept of publics is used instead of related, perhaps more accurate, concepts such as, for example, audiences, segments or communities, without offering reflections or arguments for choosing to do so. For example, we might ask if Tarleton Gillespie’s (2014) concept of the ‘calculated public’ qualifies as a public. With concrete examples, such as ‘customers like you’ and ‘friends, and friends of friends’, Gillespie’s understanding of publics is primarily a commercial and not a political or democratic construct, not even in the broadest sense, although it might become so later or overlap with publics that have similar intentions and gather around a common issue. The same can be said of Christin’s (2020) ‘algorithmic public’, which refers to the work of metrics inside newsrooms that might be powerful in guiding, constraining and controlling journalists at work but that does not refer to an intentional, collective entity. What we find useful in Christin’s work is that she highlights the discursive role of data in the construction of ‘a public’ inside the newsrooms.

What unites these approaches is that, by adding ‘public’, they highlight how the commercial and datafied profiling of audiences is in opposition to, and at odds with, the normative understandings of publics as collectives, gathered around the deliberation of a certain issue. Applying the notion of the public to all sorts of contexts in which technology is said to be doing something to these publics might only feed into this vague impression that democracy is in disarray and that technology is responsible for it. The notion of the public should not be used to raise the stakes for criticism addressed to technology. Rather, a central question in this book is how citizens, audiences or communities are formed into publics and how these public formations are diversely intertwined with the technologies of data and the media (Møller Hartley et al, 2021).

Situating datafication in relation to public formation

In this book, we use the concept of publics to explore the modalities associated with their formation and to investigate their multiple intersections with data. Hence, instead of suggesting the notion of datafied publics – implying that data does something to publics – we suggest the concept of datapublics to explore the reciprocal influences that one may have on the other without presuming hierarchy or directionality. At the same time, we consider the notion of datapublics to provide analytical entry points for empirical investigations. This allows for careful empirical illustrations of how publics are constructed (differently) by multiple actors and practices in relation to data. The drama, if still a drama, is an interactive one, with several possible endings.

In our approach audiences and users of media or citizens are not publics per se and should not be conceptualized as such. However, they are all part of distinct but overlapping constructions of publics, defined by their position in the space of symbolic struggles for legitimization and visibility. We understand publics as characterized by the overlap between, on the one hand, the constructions of publics by the media often in the shape of audience imaginaries and as an abstract ‘othering’ and, on the other hand, citizens’ civic engagement and agency struggling to legitimize certain publics.

This leads us to the other part of our concept, namely data. We suggest investigating data not only as a technological process – the dominant position in the literature, often labelled ‘datafication’ – but also as a discourse and a value system articulated through technology and various actors, including the public itself. The concept of datafication was initially used by Kenneth Cukier and Viktor Mayer-Schönberger (2014) to describe how data have become a resource that could be harvested through digital technologies, implying a highly commercial logic, as these data have become a newfound economic resource. In continuation, Helen Kennedy, Thomas Poell and José van Dijck emphasized the processual nature of datafication by defining it as the ‘process of rendering into data aspects of the world not previously quantified’ (Kennedy et al, 2015, p 1). Datafication has also been approached as a colonization of our lifeworld (Mejias and Couldry, 2019) to emphasize how datafication moves beyond simple practices of producing data. Rather, it changes multiple aspects of everyday life. In this book, we lean on these previous definitions in understanding datafication as processes of quantification enabled through digitalization, where aspects of human interaction are counted and made accessible as a new source of information that is both valuable in itself and fundamentally transformative to how publics can form, take shape, be cultivated and legitimized.

However, we are also inspired by other understandings of datafication. In describing the current datafication of today’s society, José van Dijck (2014) referred to ‘datafication’ as a discourse, or what she calls the ‘ideology of dataism’, which describes a ‘widespread belief in the objective quantification and potential tracking of all kinds of human behaviour and sociality through online media technologies’ (p 198).

Particularly useful for our work has been the conceptualization by Göran Bolin (2022) of data as value and processes of valorization from a Deweyan perspective. In this, he notes the marked distinction between when we ‘through valuation, assign value to objects around us. [W]e do this in the form of either nominal values (good/bad, ugly/beautiful) or ordinal values (1, 2, 3 etc.)’ (Bolin, 2022, p 171). Accordingly, Bolin argues that datafication is also the process of transforming quality into quantity, or, if exemplified with the distinction between private and public value, transforming ‘soft’ value forms such as equality and knowledge into numeric form (for example, equal numbers or grades [Bolin, 2022]).

This means that we pay attention to how such valorizations are being normalized, how such processes of valorization are contested and to what extent and how they are made possible by the data infrastructures. We are interested in the objective, symbolic forms that such valorizations take and the autonomy or loss of such results, for example, how some forms of civic engagement are valued more and for different reasons, or how systems of valorization technologies are imagined or resisted. It also means asking how the affordances of digital platforms allow for specific valorization and for publics to form in certain ways. Most importantly, we need to question how we, as researchers, have distinct and normative ideals of how a democracy should work (that is, what a ‘good’ public looks like and the role of the media in that process). The risk is that we overlook certain modalities of public formation or promote elite forms of participation, while marginalizing those in opposition to the elite. We need to step outside the technological drama so to speak.

These concepts point to the difficulty of distinguishing datafication from other related concepts such as digitalization, deep mediatization, metrification, platformatization, and so on. Notably, while digitalization enables datafication, the latter has a longer history, not least in the Western world. Data collection by the media is also nothing new; the media has collected demographic data on their audience, and later psychographics, without giving rise to major concerns. In the book Beyond Measure, Vincent (2022) brilliantly shows how metrics and standardizations have been around for centuries, which has historically produced and organized societies, as well as made it easier to trade across borders and keep track of citizens. In that perspective data becomes an organizing principle in societal processes with great importance for public formation.

Therefore, we emphasize historical approaches and avoid assuming that emerging technologies are dramatically new or have a dramatic impact just because they are introduced or their use is generalized. Instead, we should pay attention to how these technologies produce changes in means of power, in our capacity to imagine certain audiences as publics or in imagining certain effects of technology. Rather than taking datafication for granted, we focus on how processes of datafication are implemented, represented and valued and what effects these phenomena are producing in contexts in which publics are constructed and cultivated.

Another challenge concerns the role of the media as a prerequisite for political acts and orientations, as scholars argue that the digital is blurring the boundaries between the private and public domains (Marwick and boyd, 2014). The blurring is foremost a consequence of digital platforms increasingly becoming ingrained into our daily lives, serving no longer merely to link audiences to information platforms, but acting as personal communication platforms, debate forums for user-generated content and billions of ‘small acts of engagement’ (Picone et al, 2019) every second across the globe. Thus, digital traces in the form of data on civic and mediated practices is an increasing and continual presence across the sites, activities and relationships of everyday life, also creating a feedback loop (Mathieu and Pruulmann Vengerfeldt, 2020) to those same practices. As Dourish and Bell (2013) have pointed out, technology is not simple, but involves the ‘mess’ of its constituent or related parts, as well as those of the institutions, power relations that govern its use and the conflicting discourses that define it.

Not only should we consider the dialectics between data and publics, but we also encourage examining those in a variety of contexts, settings and situations that can be understood as hybrid (Chadwick, 2011), crossing over legacy and new media, offline and online contexts, and private and public forms of engagement. Such a dialectic is a rearticulation of the classical dichotomy structure versus agency, which in media studies often translates into a dichotomy between media and technology versus the audience. It is important to open up for the possibility that actors, including the public and journalists, can align with these structures or oppose them, evade them or change them to win the struggle over who has a right to be a public and how.

Outline of the book

In a hermeneutic fashion, this work has been assembled as a puzzle. Each individual study has been used to inform the whole picture that the book provides on public formation in datafied democracies, while providing distinct pieces of the puzzle from a unique empirical or conceptual angle. We articulate different methods aimed at unpacking the complex processes of public formation in the hybrid media system. We do not solely rely on qualitative investigations but use quantitative survey data, interviews, participant observations, protocol readings, walkthroughs, focus groups and document analysis. We situate our analyses and conclusions in newly gathered empirical data; however, we also draw on previous studies we have conducted individually and collectively.

The book is divided into three parts, each addressing different parts of the drama we have identified. Hence, the first part of the book is concerned with the role of publics, whom we perceive as agentic. We turn to audience research to explore the formation of publics from a sociocultural perspective, the mundane lives of citizens, the resources they engage and their concrete experiences with datafied media.

Chapter 2 develops an argument against the soft determinism implied in the way publics are said to be affected by data and algorithmic control. It introduces a sociocultural approach, focusing on media consumption as a distinct moment from media production, the actual locus of algorithmic control. A distinction is made between our ‘uses of media’ – what is essentially captured as data and turned into (limited) knowledge – and our sociocultural practices – which remain to date largely invisible and incomprehensible to the data episteme. The gap between uses and practices provides a reflexive space of agency for users to escape algorithmic control, inserting a narrative of hope after a decade of research that has traced a rather dark picture of the datafied society and its consequences for democracy.

Chapter 3 explores how the dynamics of datafication intervene and influence the public formation processes and presents a typology of public formation tactics. Using COVID-19 sceptic groups as a paradigmatic case study, it analyses how these marginalized and often ridiculed groups assert themselves in today’s datafied, hybrid media system through going from an inward to an outward focus framed as mobilization tactics, counter-tactics and publicity tactics. The chapter discusses how quantification logics permeate the processes of publicness, how the collective is not necessarily connected through an issue but a common struggle for legitimacy as anti-mainstreamers, and how traditional mass media still plays a central role in recognizing the counterpublic as such.

Chapter 4 uses a mixed-method approach, drawing on the Multiple Correspondence Analysis of large-scale survey data and netnographic mapping of Facebook groups. It argues that the activities in these groups, which tend to attract the seemingly apolitical and publicly disconnected, on a broad scale maintain the simmering potential of the publicness of civil society. Challenging the classic notion of the private/public distinction, the chapter presents empirical evidence suggesting, first, that public formation occurs in seemingly ‘private’, mundane and non-political spaces, and, second, that it does so while linked to processes of social stratification, the resources of those citizens who engage and their habitual preferences (or civic ‘lifestyles’), which does not enjoy equal social recognition.

The second section of the book investigates the imaginaries and values as a vital part of understanding what is at stake when new technologies enter the hybrid media landscape. Chapter 5 illustrates through a case study of the New York Times how data logics and imaginaries of digital technologies often construct conflicting and opposing narratives of ‘the news users as publics’ and the role of journalistic media as cultivators of such imagined publics. Further, by linking this case study to a historization of the increasingly datafied distribution and audience measurement technologies, the chapter explores how audiences and thus publics are constructed differently as new measurement technologies emerge, from democratic collectives to segmented consumers, and finally, with the introduction of personalized recommendations as aggregated datapoints. Thus, the chapter shows how such opposing narratives have always been present, but different weights have been given to user constructions at different times.

In Chapter 6, we enter news organizations to investigate the introduction of personalized content distribution as a next step in a greater datafied evolution. With rich empirical examples from ethnographic fieldwork and qualitative interviews, we show how publics are cultivated through what we label ‘personalization logics’: individualism; dataism; and binarity and pre-determinedness. These ‘personalization logics’, we argue, become drivers for how media organization (re)construct their audiences, namely as aggregated, predictable and controllable datapoints. Thus, this reconstruction of the audience allows the media organizations to engage in new form of publics cultivation – publics by design – as they now materially begin to shape and design the publics they wish to cultivate into these systems. Finally, we discuss the broader implications of these processes and the cultivation of publics as personalized aggregated ‘dividuals’ by news media (Deleuze, 1992).

The third section of the book investigates the role of large-scale providers of technological infrastructure in shaping, validating and creating publics. It does this through the lens of what could be termed the ‘old’ or ‘traditional’ leading characters in this endeavour: the institutions of journalism and media.

Chapter 7 zooms in on metadata as an infrastructure in the news distribution ecology linking news media to the demands of intermediaries, by examining the metatags of 260 European news websites and determining the extent to which they apply metatags that allow webpages to be featured in Google search results. Furthermore, the chapter examines how media organizations are compliant with the metadata library of Schema.org, which is organized by, among others, Yandex, Google and the World Wide Web Consortium. While a majority of media are compliant with Google and Schema.org libraries, only a few large organizations utilize these to the full extent. Finally, the chapter discusses the impact on publics in the context of access to information via the news delivered through search engines.

Chapter 8 examines the digital infrastructure connected to media website architecture and discusses effects of such infrastructures on journalistic ideals of providing the information needed to enable a democratic public (see Strömbäck, 2005; Fenton, 2010). By using the media tech stacks as our point of departure, the chapter presents a mapping and categorizing of the systems sustaining journalism in the efforts of cultivating publics.

Next, by mapping third-party web services found on 361 European media websites, we uncover that the media in general rely heavily on Big Tech-provided systems for the production and distribution of journalistic content. In addition, we observe a large quantity of ‘small tech’ which tend to be overlooked in narratives around the dominance of Big Tech.

Chapter 9 concludes the book by discussing how datafication processes are transforming how the public takes shape, as well as the journalistic values and the dilemmas this brings about. However, we argue for a view that recognizes stability in that those technological affordances tend to reinforce already existing logics and discourses in journalism as well as existing public perceptions concerning the role that news media should play. Such assembling involves increasing dependence on large amounts of data and tools for analysing such data, which transforms not only the ways that news organizations cater to different publics but also the ontology of news and publics. This development questions which news is enhanced by audiences’ (hybrid) activities. These audiences influence, contradict, oppose and discursively construct the tales of how citizens stay informed and democratically engage in datafied societies and the role of news in the formation and cultivation of publics. Knowing more about the formation of publics across spheres, we suggest, enables us to understand the multiple realities of datafication, and how these relate and coexist.

Taken together, the chapters offer different versions of how and when datafication matter for public formation and the multiple actors and practices involved in public formation. An important contribution and suggestion in this book is that publics are materialized and constructed over time when people pay attention to the discursive issues and injustices brought to light and represented by the media and direct this attention to engagement, calling for (political) action. In other words, publics are discursive clusters of stratified position takings by citizens (the plural is important here because publics do not always gather as a unit with a common goal) overlapping with clusters of media attention and visibility, legitimizing those position takings to a varying degree. And datafication matters for many of these modes of publicness. In the concluding chapter we revisit these core arguments and suggest that there is a need to blend our understanding of technologies and media with that of audience agency and look at how datafication is bringing about certain forms of power to specific actors, if we are to move beyond the drama and the rise and fall narrative of technologies as being something that is imposed upon us.