Overlit: Digital Architectures of Visibility

Abstract

Despite the ubiquity of digital technologies, data-driven approaches and algorithms, organization theory so far only engages with these developments in limited ways. A deeper engagement with the organizational ramifications of a digital, datafied world is urgently needed and must start from mappings of the phenomenon and the development of better theoretical vocabularies that can guide future research. Complementing the essays by Zuboff and Power in this exchange, my essay suggests a research agenda based on how digital technologies, data and algorithms impact and shape our lives in and around organizations by making us visible in novel ways. I unpack the technological and operational underpinnings of this phenomenon in two steps. The first is a broad conceptualization of the overall shape of what I term ‘digital architectures’. The second is a more granular theorization of how data-driven, algorithmic approaches make the ‘management of visibilities’ a central concern for humans, organizations and societies, as well as some reflections on possible responses to these developments. Taken together, these discussions highlight how digital ubiquity calls for novel theoretical perspectives and research avenues for organization theory to explore.

Introduction

Our lives are increasingly ‘overlit’. Digital technologies and data-driven, algorithmic approaches make the exposure of social life easier and more invasive than ever before, and they generate new ways of seeing, knowing and governing organizational affairs. And with the ubiquity of digital technologies and data extraction across private, commercial and public settings, what we do is more visible than ever, and this state of affairs sets our times apart from the past. Increasingly, individuals and organizations are under the spotlight and must often themselves figure out what to disclose, what to keep secret and what to monitor or ignore. At the same time, these digital developments involve significant asymmetries and far-reaching inequalities when it comes to access to information and data. As a result of this ‘fast-growing abyss between what we know and what is known about us’ (Zuboff, 2020a, p. 5), people and organizations face a wealth of daunting and unprecedented challenges that organization and management studies has yet to come to grips with. Digital spaces and their operations are to date still largely uncharted territory for organization theory. For sure, organization and management scholars have focused on questions about the relationship between technologies and organizational phenomena (Leonardi, 2011), have carried out situated studies of organizational uses of technology (Orlikowski, 2000) and have studied the emergence of platforms (Cusumano, Gawer, & Yoffie, 2019). Still, many aspects of these emergent digital spaces remain in need of empirical scrutiny and theoretical attention. Our exchange in this special section about surveillance capitalism, digital architectures, data traces and visibilities calls for a deeper engagement with technological infrastructures, digital exposure and new data-driven forms of social ordering in the field of organization theory. Collectively, with these essays we seek to anchor questions about the political economy and logics of ‘surveillance capitalism’ in novel conceptualizations of knowledge, technology and human experience that may help scholars of organization study these developments. Hopefully, the exchange will pave the way for new research perspectives in organization and management studies that place these ‘digital architectures’ at the centre of research activity in the field.

In previous work, I have explored the intersection between technology, visibility and new social orderings as a way of problematizing hopes about transparency and of making sense of the dynamics and consequences of digital transformations (Flyverbom, 2019). The focus on visibility is not meant to reduce observation to a visual dimension, but rather to unpack how digital technologies, data and algorithms facilitate new forms of observation and knowledge production and, in doing so, bring about a new social order.

In this essay, I bring these ideas in conversation with Zuboff’s (2022, in this issue) interest in the nature and workings of the social order that she terms ‘surveillance capitalism’, and with Power’s (2022, in this issue) theorization of a datafied ‘economy of traces’ that has consequences not only for accounting, but also for the social sciences more broadly. In what follows, I conceptualize digital architectures of visibility in ways that capture both the outside shape of these machineries, and the inside dynamics at play – and I draw out how this conceptualization helps us better grasp the profound shift in our organizations, economies and societies that is brought about by these digital architectures.

Making Sense of Digital Spaces

Zuboff’s (2019) work on surveillance capitalism is an important starting point if we want to show how organization theory can take on the challenge of addressing questions about digital architectures of visibility and the economic logics that drive them. Propelled by data extraction, originally for the purpose of selling ads, digital developments presently take a very particular shape that may be understood as forms of data capitalism (West, 2019) or surveillance capitalism (Zuboff, 2019). The value of diagnosing contemporary digital transformations in such terms is that it unveils the commercial and political forces at play – in what may otherwise seem a natural or given state of affairs. Conceptualizations of data capitalism and surveillance capitalism remind us that the problem is not technology or digital transformations as such, but the way commercial and other forces set the conditions for such digital developments and put such technologies to work. That is, we could have technological solutions on other terms than the present ones. Zuboff’s (2019) book on surveillance capitalism is both a detailed map of a phenomenon in dire need of attention, and a powerful call to action. However, like other maps, it only captures the particular dimensions it looks at and operates at a distance from human and organizational realities (Latour, 2005). We have to move questions about digital architectures – and how they make us see, know and govern – closer to organization theory. To this end, we need theoretical vocabularies that help us make sense of these fast-changing developments and their potential impact on organizations, stakeholders and society in more concrete and granular ways.

In this essay, I suggest two moves that can open up empirical and theoretical research avenues for studies of such digital architectures of visibility. The first move is about extending the scope of the research agenda to a more encompassing conceptualization of the shape of the digital architectures that we are confronted with at present. Zuboff’s (2019) seminal work already maps and diagnoses these technological developments. Still, her account has more to say about commercial logics, political economy and the consequences for emergent forms of capitalism, than about the concrete shape of digital, datafied spaces. For organization theory, conceptual accounts of the material contexts of digitalization and datafication are however crucial, I argue, particularly in combination with the attention to the logics of accumulation that Zuboff unpacks. What we face – as individuals, organizations and societies – is a broad and fundamental set of technological transformations that condition more and more parts of organizational life. And we need to articulate their material shape in more direct ways than Zuboff (2019) does.

The second move in my essay is to deepen the analysis by focusing on ‘visibility management’ (Flyverbom, 2019) as an important set of dynamics at play inside such digital spaces. Digital architectures, I suggest, create novel forms of observation, information flows and knowledge production that we have yet to fully conceptualize. When people’s lives and other social phenomena are datafied; i.e. turned into data points and sorted out via automated forms of categorization, new forms of visibility emerge and these have consequences for the workings of organizations and for social ordering. The way technologies, data and algorithms produce visibility remains to date, however, somewhat overlooked and has not been spelled out in much detail in organization and management studies. This is problematic because processes of digitalization and datafication – which in the past were easier to consider as external forces and as more clearly circumscribed phenomena – are increasingly integrated aspects of many organizational settings and therefore in need of attention. In the text that follows, I suggest that the result is a novel situation, where managing visibilities becomes central to individuals and organizations.

These two moves together highlight the material shape and inner operations of contemporary digital architectures and point to the novel conditions that humans and organizations face in a digital, datafied age. I unpack these two moves in three sections. The first, next section offers the characterization of the emergent features and overall shape of digital architectures. The subsequent section then develops the conceptualization of the inner workings of digital architectures, with a focus on knowledge production and dynamics of visibility. The final section then offers a set of possible research avenues to explore along these lines of thinking.

First Move: The Shape of Digital Architectures

How can we describe the shape of the digital and datafied spaces that we live in at present? This simple question opens up all sorts of avenues that are difficult to navigate at first. What is important in this context is not the contents of communication, but in a sense the conduit of communication; i.e. the means and technological forms that underpin digital spaces. By shifting our focus away from the substance – what is effectively communicated, distributed or received – we become attuned to the material components and (infra)structures that condition processes of communication (Flyverbom & Murray, 2018). This interest in conduit rather than content has solid roots in mass communication and media research (e.g. Deibert, 1997; Meyrowitz, 1985) as well as other approaches that study technology in the form of communication environments, and includes a focus on audiences, producers, corporations, channels of distribution, legal systems and other material components and structures. This attention to the shape and dynamics of communication environments is central if we want to offer a more general conceptualization of digital, datafied spaces, and develop appropriate organizational and societal responses to their growing impact.

To this end, I think it is useful to describe these spaces as ‘digital architectures of visibility’. The term ‘architecture’ captures that the layout of digital spaces enable and constrain our lives and works as a subtle form of regulation (Lessig, 1999). A focus on architecture also highlights the structural design of digital spaces (Bossetta, 2018), their inside operations (Easterling, 2005) and the particular dynamics they produce or set off (Parisi, 2013). The concept of ‘visibility’ in this description then signals my interest in dynamics of observation and exposure that are central to contemporary digital developments (Brighenti, forthcoming; Flyverbom, 2019; Leonardi & Treem, 2020).

We encounter digital architectures of visibility in numerous contexts, such as when we use social media, access streaming services, or read the news online. They facilitate smooth access to products and services, but also involve extensive data harvesting that makes us exposed and visible. At their core, these digital architectures consist of entangled components such as data, algorithms and technologies (Alaimo & Kallinikos, 2020b) that make them into particular ‘machineries of knowing’ (Knorr-Cetina, 1999, p. 10). By themselves, digital architectures have a number of important technological features, including websites, networked electronic devices, protocols, applications and other interfaces and systems that make communication and other interactions possible. The most straightforward way of capturing these facets of digital spaces is to highlight that they involve an entanglement of four processes that we can describe as digitization, digitalization, datafication and connectivity (Leonardi & Treem, 2020). Digitization involves simple processes of turning analog objects into digital ones, such as when books or albums are made available in digital formats. Digitalization, in comparison, refers to broader organizational or societal transformations involving digital technologies, such as when work processes move to digital systems or when communication and collaboration takes place via distributed, digital arrangements. Furthermore, and as a result of processes of digitization and digitalization that produce digital traces or data, we increasingly talk about processes of datafication. Datafication implies that human actions are turned into data, often for strategic, optimization or governance purposes. As stressed by Leonardi and Treem (2020), digital transformations are propelled and made valuable by increased connectivity, allowing digital and datafied processes to be integrated for purposes of analysis and processing. These four features of digital architectures are important when we seek to conceptualize their shape and workings.

However, we do not need to probe very deeply to realize that these are not merely technical features. For one, it is increasingly difficult to separate the technological systems from the data they produce and rely on, such as the digital traces that are both by-products of everything we do when using digital technologies, and a valuable resource for those who can use them to understand and modify behaviour (Power, 2022; Zuboff, 2019). Also, algorithms and other automated ways of sorting out information are inseparable from the data and contexts they work with and are shaped by (Faraj & Pachidi, 2021; Glaser, Pollock, & D’Adderio, 2021). Adding further to the complexity, strong commercial and political forces such as powerful tech companies (Nadler & Cicilline, 2020) and national policies and digitalization strategies (Gillespie, 2017) further propel and shape the digital solutions and transformations that we live with at present (Flyverbom, Deibert, & Matten, 2019).

In other words, a conceptualization of digital architectures captures more than technical features so as to encompass the complexity and richness of technological, social and commercial forces that propel and shape them. However, despite already extensive theorizing and a solid literature seeking to articulate technological and societal transformations (Castells, 1996; Couldry & Hepp, 2018), we still have few integrative and precise conceptualizations of the relationship between digital systems, data, algorithms and the commercial and political forces that shape these technological spaces at present. The challenge is, of course, that such a conceptualization will have to include both technological and social dimensions (Leeonardi, 2011; Orlikowski & Scott, 2008), and both the minutiae of data analysis as well as broader processes of datafication and digital transformations (Amoore & Piotukh, 2015). A full understanding of the make-up of such architectures will arguably only be possible in hindsight, and as a result of theoretical work and accumulating empirical insights. Still, some pieces of the puzzle can be put together at this point.

By way of analogy, a short detour to the societal importance of the printing press may be illustrative. Similar to digital architectures, the significance of earlier technological developments was not apparent when they emerged. The printing press, invented in the middle of the 15^th century, was a relatively simple technical innovation making it possible to reproduce texts mechanically. Copying books and pamphlets had previously required labour-intense work, done meticulously and word by word with a pen in hand, but could now be sped up and done at scale. At the time, the ramifications of this technological innovation were unclear to those developing, using and benefitting from it. However, in hindsight, we know that the printing press paved the way for an extensive and fundamental transformation of society. Much of what we associate with democracies, modern societies and science have ties with this technological advance, which made it possible to print, copy and distribute information. The ability of the church to control what could be read, and by whom, became contested, ordinary citizens could engage in public debates about politics and other matters, scientists could disseminate their findings, and national cultures and languages could flourish as a result.

In a similar manner, we will only fully understand the consequences of digital technologies and processes of datafication in probably 50 or 100 years from now. The technological components of present-day digital architectures may also be vastly more complex than the printing press. Still, their significance will not only result from technology, but also from the societal dynamics that they give rise to and are entangled with. If the printing press paved the way for the ‘Enlightenment’ period, I wonder if we may think of our times as an ‘Overlightenment’ period. The printing press and other sources of insight and ways of sharing information made it possible for more people to gain insights and unveil otherwise opaque phenomena. Digital technologies have extended and accelerated these developments that we associate with the Enlightenment. But they have also created new and at times perverse dynamics and developments that take us way beyond a state of affairs where information is searched and shared for the benefit of all. As Leonardi and Treem (2020, p. 17) suggest, we are now living in digital communication spaces marked by visibility, where ‘the rapid increase in digital connectivity enabled by both social and technical infrastructures means that people and organizations are at risk of becoming hyper-visible’. The ubiquity of digital technologies and data makes it possible to expose and share more parts of our social life. Just like the invention and ubiquity of electrical light led to what Bogard (2014) describes as the ‘end of night’, our digital communication spaces are not just enlightened, but rather ‘overlit’. Zuboff (2019) forcefully shows that this situation is the result of political and commercial forces that seek to dominate our human future, and we urgently must explore its shape and the human and organizational conditions it gives rise to. In the following, I highlight some theoretical building blocks that such a conceptualization may start from and which I use to offer a more encompassing definition of the shape of digital spaces.

Building blocks for a conceptualization of digital architectures

As already mentioned, the conceptualization of contemporary digital architectures needs to consider the relationship between technical features and the human, organizational, institutional and societal contexts they are used in. Such more encompassing approaches are well-known in the literature on socio-materiality (Latour, 2005; Orlikowski & Scott, 2008) and affordances (Hutchby, 2001; Leonardi, 2011) that studies how technologies, humans and social forces are interrelated. As a result of work in these areas, we understand why technologies become used in ways that those designing them may have never envisioned, and we approach the consequences of technological transformations in turn with more nuance and curiosity. Still, these approaches have mainly been used in situated, often ethnographic studies (Orlikowski, 2000), rather than to theorize the overall makeup of digital architectures – which is a much broader phenomenon.

Media studies is the scholarly tradition that most consistently has sought to broaden its scope of inquiry to include cultural, political and other social forces in conceptual and empirical explorations of communications technologies (Morley, 2009). The attention to new technological developments in this tradition means that such accounts continue to produce important insights about the intersection of social and technological forces. In recent years, media studies has, for example, theorized key dimensions of digital spaces, including how they mediate social reality (Couldry & Hepp, 2018), and produce new ‘programmed socialities’ (Bucher, 2018). These are important building blocks. But as a field, media studies has tended to focus on one medium at a time and in separate discussions – moving from radio to television to computers, and so on. In a similar vein, the different components of digital spaces have been scrutinized separately in recent years. For instance, scholars have explored the shape of digital transformations under increasingly specific headings, such as platform studies (Cusumano et al., 2019), social media studies (Fuchs, 2017), algorithm studies (Pasquale, 2015), software studies (Manovich, 2013) and critical data studies (Iliadis & Russo, 2016). These are all important facets of digital spaces, but they fail to recognize that these facets increasingly intersect. Zuboff’s (2019) account of the all-encompassing nature of surveillance capitalism highlights the problem with such siloed approaches that oftentimes miss the underlying logics and forces at play. The focus on the specific facets of digital spaces becomes increasingly problematic in our times where technologies both subdivide and converge. It seems plausible that, in the near future, it will be futile to discuss any medium as separate from the architectures of entangled data, algorithms and technologies that both facilitate and constrain everything from music, news and films to information access, communications and business. The value of a conceptualization that focuses on the architecture as such is that it can connect these many building blocks and insights about different technologies and features of digital transformations to help us develop a more encompassing picture and pave the way for more informed responses.

The ‘architecture’ image, as already mentioned, points to work on infrastructures that focuses on digital spaces as built arrangements that, just like electricity grids or railway systems, ‘facilitate the flow of goods, people, or ideas and allow for their exchange over space’ (Larkin, 2013, p. 327). Work by Larkin and others reminds us that digital spaces are however both like and unlike other infrastructures. They play organizing roles like other infrastructures, but are also shaped by particular technological, commercial and political forces, such as surveillance capitalism and the push for hyper-visibility. Understanding these particular forces helps us grasp the shape of digital architectures and the surveillance logics at play, and reminds us that in our times, these infrastructures are both vehicles of communicative practices as well as political and economic forces that seek to order societies and dominate our human future (Zuboff, 2019, 2022).

Indeed, digital architectures can perhaps best be understood as assembled and as the result of design choices, business models, concrete uses and ideational forces (DeNardis & Hackl, 2015). Alaimo and Kallinikos (2020b) suggest the term ‘apparatus of data-technologies-algorithms’, and this description is nicely broad and captures the architecture-like configuration of data, algorithms and technologies. However, it is a largely descriptive, neutral label and does not emphasize the key operations of ‘visibility management’ that are, I argue, at the core of current data and technology architectures. Hence, I here define digital architectures as an entangled configuration of technologies, data and algorithms that make social life increasingly visible and knowable. In the next section I elaborate the core operations of exposure and visibility inside these architectures that turn human action into data objects for purposes of optimization and governance.

Second Move: The Inside Operations of Digital Architectures

My next move in this essay focuses on the forms of data analysis and automated sorting that happen inside digital architectures and make social life visible and knowable in new ways. In particular, my ambition is to articulate that ways of categorizing data and managing visibilities are important dynamics of digital architectures that have profound consequences for humans and organizations. I show this by both articulating how datafied forms of visibility management affect us, and by highlighting some ways of reacting – individually, organizationally and politically – to these dynamics.

As suggested at the outset of this essay, insights from mass communication and media theory (McLuhan, 1964; Meyrowitz, 1994) help us start to understand the workings of digital spaces by focusing on their technological forms and operations, rather than their contents. Along these lines, digital architectures can be seen as the convergence of ‘message-circulating technologies’ and ‘data-extraction-and-analysis technologies’ (Turow & Couldry, 2018), and as the ‘apparatus of data-technologies-algorithms’ that Alaimo and Kallinikos (2020b) describe. Recently, a growing number of contributions highlight the roles of digital architectures in the reconfiguration of social affairs, and do so by looking at the concrete operations and dynamics of the data, technologies and algorithms involved (Alaimo & Kallinikos, 2017, 2020b; Amoore & Piotukh, 2015). My analytical focus on the inside operations of digital architectures, including algorithms and the data sorting work they do, also follows this line of enquiry, but brings in a much more direct focus on questions about visibility management.

Refractions, visibilities and social ordering

Digital architectures digest human experience and social life by turning them into data and return them to us in algorithmically sorted forms that shape our conduct in fundamental ways. This is so because many social activities – human movements, social relations, cultural consumption, production processes, and so on – are or can be made into data objects, and can be analysed with an eye to similarities and patterns that are used to understand, commercialize and shape particular forms of human action.

I am interested in the forms of ‘knowing’ that digital architectures rely on and propel, and how these ‘machineries’ (Knorr-Cetina, 1999) make us grasp and govern social affairs in new ways. Producing knowledge via data means that social phenomena are taken apart and reassembled. This happens, for instance, when music or fashion is reworked (Alaimo & Kallinikos, 2017, 2020b), and treated not in terms of the properties of an album or a piece of clothing, but by capturing how people interact with music or clothes, how they share and comment on social media, or how users appear to resemble other people. When we focus on the inner workings of digital architectures, these questions about knowledge production and social ordering move to the fore. The point is that the work of dealing with social phenomena through data points transforms not only objects, but also the very nature of social realities. As Alaimo and Kallinikos (2020b) put it: ‘categorization delivers the basic blocks of reality’, and this helps us grasp the workings of datafication and refraction as particular forms of social ordering. With the term refraction, I suggest that digital architectures constitute particular symbolic environments that produce new social realities. When human and organizational practices and other social phenomena come to rely on or travel through digital architectures, they are turned into data, they are reordered and they have performative effects.

My conceptualization of refraction and visibility management extends insights about the kinds of ‘information reductionism’ (Tsoukas, 1997) that all digital systems rely on. Early on, Tsoukas (1997, p. 829) described how digital technologies make it possible for ‘everything to be viewed as information (especially digitized information), as some-thing (an object) that can be processed, stored, sent over, retrieved’. This reduction of social phenomena into digital information is, as already mentioned, central to processes of digitalization and datafication. As individuals or social processes are reduced to data points and sorted algorithmically, they take on new shapes; their richness and origins may be backgrounded or go missing altogether, and the historical, political and other social forces shaping them may be overlooked or taken for granted (Power, 2022; Zuboff, 2019). Also, when our lives are refracted through the prism of technology, we take the shape of ‘digital doubles’ – understood as the sum of our digital traces – and these aggregations of data may be used as proxies (Mulvin, 2021) when others make decisions about us.

Organizations and other social entities and groups are similarly reconfigured by digital architectures, such as when newspapers use tools from Google and Facebook to sell ads or reach audiences, or when public institutions rely on data-driven, algorithmic systems in decision-making. This kind of digital doubling or what I define as ‘refraction’ creates a range of associated dynamics that affect interactions between individuals and organizations. We have yet to understand such dynamics and figure out what happens to individuals and to social processes that are churned through digital architectures with particular technological, political and commercial logics.

In my view, when digital architectures refract, they harvest and order data in aggregate form as patterns of similarity and difference, they provide and push information for application in other domains, they enact human actions and identities in new ways, and they enable and constrain even more fundamentally what we attend to and act on. These refraction dynamics cut to the core of our human condition and remind us that digital architectures are all about power, knowledge and social ordering. Explorations of such dynamics have to take place inside digital architectures, where data are first selected, organized, moderated and mediated, and then cast as insights, truths or certainties with performative effects. My understanding of these developments is indebted to Foucauldian (1983, 2007) insights about ‘pastoral power’ and ‘governmentality’. Such work stresses how all forms of governance rely on practices of seeing and knowing that turn humans and social phenomena into governable subjects and objects. Whereas Foucault’s work focused on the governance of entire populations, we increasingly see technological and political orientations towards the governance of individuals or of proxies that stand in for the individual. Such governance plays out in commercial settings where products and services, such as insurance, are targeted very precisely, in health care where individualized, precision treatment is in focus, or when public sector institutions prioritize their interventions based on fine-grained data-driven, algorithmic profiling of their citizens.

To make sense of these performative dynamics, our conceptual vocabulary has to, first of all, capture how social phenomena are processed through architectures of data, algorithms and technologies. Beer (2019) offers the concept of a ‘data gaze’ to explain how digital technologies and data enable particular forms of visualization, observation and knowledge production. As Beer (2019, p. 11) puts it, the concept of a data gaze can help us understand ‘new forms of data-led knowledge and examine what these will mean’ and ‘explore how data are seen and how we come to be seen, unendingly, through data’. Algorithms are central to these refractions, because they can distinguish, categorize, and otherwise sort out information in automated ways, and this means that they have both commercial and cultural ramifications. They not only underpin data-based business models, largely focused on advertising, but they also shape human action, cultural developments and social orders (Gillespie, 2018). For instance, when recommender systems use patterns in data to suggest new contents to users, and to anticipate user behaviour and emergent trends, they also again refract and regulate human action. These kinds of recommender systems underpin Spotify, YouTube and Netflix as well as many other platforms, and are used to push content that grabs our attention through algorithmic operations that sort and classify data and users, often in opaque ways, producing much more than commercial advantages for social media and tech companies. Refractions may take the shape of negative consequences, such as when YouTube operates as a ‘radicalization machine’ that uses recommender systems to drive users into ever more radical – so-called ‘engaging’ – content (Tufekci, 2018). Other problematic refractions include what O’Neil (2016) terms the spread of data-driven ‘weapons of math destruction’ that undermine democracy and justice, in line with what Zuboff (2019) highlights. When apparatuses of data, algorithms and technology are used to evaluate performance, for example in schools, or to make decisions about individuals, they may have negative effects that are both hard to digest and address unless we understand how they work as refraction devices, or ‘prisms’. Such apparatuses can thus propel, amplify and encourage social transformations of all sorts because they not only feed back on our behaviour but by doing so fundamentally alter our identities, thoughts and behaviours. These reflections about digital prisms and refractions align with Zuboff’s (2019) point that tech companies are in the ‘reality business’ – they extract human realities, commodify human experience and modify human behaviour in fundamental ways.

In contrast to what I argue is a fundamental process at the heart of many digital architectures, social media and tech companies describe themselves instead as neutral platforms and as mirrors of society. This implies that they take little responsibility for what people share or say on their platforms and they downplay the role they play when it comes to instigating or perpetuating social and political dynamics. Facebook’s vice president of global affairs and communications, Nick Clegg, recently put it this way:

When society is divided and tensions run high, those divisions play out on social media. Platforms like Facebook hold up a mirror to society — with more than 3 billion people using Facebook’s apps every month, everything that is good, bad and ugly in our societies will find expression on our platform. That puts a big responsibility on Facebook and other social media companies to decide where to draw the line over what content is acceptable. (Clegg, 2020, p. 1)

To say that social media are mere reflections of society seriously downplays their role as forces that create or otherwise facilitate political and social transformations. To grasp these forces, we need, as mentioned, to think differently about digital architectures and their relation to reality. Rather than mirroring existing realities, digital architectures produce new realities by extracting, categorizing and highlighting parts of social life (Loxley, 2007). Digital architectures do not simply reflect social phenomena and realities, but refract them.

These refraction dynamics create in essence an entirely novel human condition – one where we are digitally and visually made and remade – and is one that requires new strategies of resistance. Such strategies involve constant and extensive activities that I refer to as forms of ‘visibility management’. In digital architectures marked by increased visibility, humans and organizations are motivated to control and curate their digital traces by cleaning up, hiding and managing what otherwise gets exposed about them, often by default and via automated forms of data harvesting. This is a daunting task in the face of complex commercial systems set up to maximize attention grabbing and data extraction. Managing our visibilities is also difficult because it requires technical, legal and organizational skills that very few possess individually. Particularly children and other vulnerable groups are largely left to their own devices, and the lack of regulation and standards for conduct in this domain has created a situation where normal expectations about corporate responsibility, accountability and openness often do not seem to apply.

Conceptualizing the inside operations of digital architectures

At present, a number of conceptualizations seek to describe how data processing actually happens and aims to account for the steps involved. Such work often speaks of data analytics in terms of life cycles (Lury, 2020) and value chains (Alaimo & Kallinikos, 2020a; Flyverbom & Madsen, 2015). These accounts are important starting points when we want to understand the inside operations of digital architectures. Just like scientific results and other forms of knowledge production (Latour & Woolgar, 1979), datafied knowledge production cannot be separated from the contexts and conditions that shape it (Thylstrup, Flyverbom, & Helles, 2019). For instance, Amoore and Piotukh (2015, pp. 360–361) highlight processes of ‘ingestion, partitioning and machinic memory [that] reduce heterogeneous forms of life and data to homogenous spaces of calculation’. These operations not only reduce social life to data but may also institutionalize problematic forms of management and logics of governance. In their account of the ‘computing’ of everyday life, Alaimo and Kallinikos (2017) distinguish between processes of ‘encoding’, whereby social phenomena are reduced to data points; ‘aggregation’, whereby calculative operations and classifications compile data into more abstract categorizations; and ‘computation’ whereby scores and measures are crafted. These steps pave the way for the kinds of output we associate with data-driven approaches, such as user recommendations and other forms of action taking place at the user interface (Alaimo & Kallinikos, 2017).

Other ways of accounting for these knowledge production processes include typologies that distinguish between the ‘datafication of daily life’, algorithmic operations focusing on ‘patterns and predictions’, and the ‘calculative rationalities’ and forms of governance at play (Madsen, Flyverbom, Hilbert, & Ruppert, 2015). Datafied, algorithmic knowledge production involves particular ways of seeing, knowing and governing social affairs that can be broken into distinct analytical steps. Such deconstruction demystifies the very nature of datafied, algorithmic operations and moves us beyond the fear that they are ‘black boxes’ which cannot be accounted for (Bucher, 2018). In short, by conceptualizing the inner workings of digital architectures as processes of extraction, sorting and classification, we get a sense of the different steps and of the concrete operations at play when social life is refracted through the digital prism.

A focus on knowledge production highlights questions about the consequences of digital architectures for organization and social ordering, such as how they churn social phenomena and human lives into digital traces and extract insights (Power, 2022). Addressing such questions allows us to explore more broadly the consequences of digital architectures consisting of data, algorithms and technological systems, and how we may identify their limits and consider possible alternatives. The focus on the operations of datafication is also part of Zuboff’s analysis, but her book is primarily oriented towards a novel mutation of capitalism and the human condition it creates, and not to the more granular operations of digital and datafied spaces and their consequences for organizations. Digital architectures will only become more ingrained in social life once their ways of working seep out of social media platforms and apps, and into our public institutions and organizations of all sorts. Even if we imagine a world where tech giants were regulated heavily or broken up, the digital architectures of visibility they have developed and propelled would still thrive as prisms refracting our organizations and societies.

Societal responses to digital architectures

Finally, the conceptualization of the inner workings of digital architectures is a way to point to the political and regulatory choices that we face collectively if we want technological transformations on other terms than the present ones. To manage our visibilities in digital architectures is an unfair, even impossible, task for individuals, simply because we have no way of setting limits to data extraction, even knowing when or where it happens. This is the basis for the ‘epistemic inequality’ of contemporary digital architectures that Zuboff’s (2019, 2020b) work seeks to uncover and critique, such as when she reminds us that we have a ‘right to sanctuary’. At present, fighting for that right is largely an individual task that involves various forms of digital self-defence and careful management of our own data. Others also stress that the increase in visibility that I have highlighted is marked by deep and problematic chasms and divides. Hong (2020, p. 2) puts it this way:

The promise of better knowledge through data depends on a crucial asymmetry: technological systems become increasingly too massive and too opaque for human scrutiny, even as the liberal subject is asked to become increasingly legible to machines for capture and calculation.

In light of this, the management of visibilities is not simply an individual problem or a practical matter. It is also a pressing political and societal concern that we need to address. Digital architectures are ‘overlit’ and saturated by surveillance logics because they have been colonized by a handful of tech giants with primarily commercial aims (Zuboff, 2019). As the strategic public relations fog created by these tech companies evaporates and the public and political demands for rights and responsibility grow, we can start to develop the institutions and regulatory frameworks that may give us digital technologies and solutions on better terms.

Indeed, political choices and interventions will be central in years to come. The European Commission and a number of European states are leading these efforts to both enforce existing regulation and develop new regulation. Such efforts have yet to be ratified and adopted by member states, but they may potentially set limits to surveillance capitalism and pave the way for digital transformations on fairer and more democratic terms by limiting certain business models and protecting the rights of citizens (European Commission, 2020). Similarly, in the United States, there is a growing focus on regulatory interventions to limit the dominance and rogue operations of tech giants (Nadler & Cicilline, 2020). However, and in line with my arguments in this essay, I believe that we need a clear focus on digital architectures and their data operations, rather than separate components, individual companies or media contents if we want to effectively understand and regulate our digital future (see Vergne, 2020 for a similar argument).

Research Avenues for Organization and Management Studies

As suggested above, processes of digitalization and datafication impact organizations in multiple ways, and these transformations can be studied from both established and novel starting points. This essay’s conceptualization of digital architectures and dynamics of visibility emphasizes how technologies condition how social phenomena are refracted and reconfigured through ‘digital prisms’ and ‘data gazes’. If digital technologies operate more like refracting prisms than neutral windows, their significance and ramifications cut to the core of organizational and social life. When social phenomena are refracted through digital architectures, it has consequences not only for the shape of culture, such as music, fashion and taste, or for politics, such as elections, political movements and worldviews. Digital ubiquity also regulates, refracts and reassembles fundamental aspects of organizations and organizing processes. As organizations become increasingly enmeshed with digital architectures, their logics and ways of datafying, categorizing and visualizing set new conditions for organizational life. However, questions about how the apparatus of data, algorithms and technology makes us see, know and govern organizational affairs have yet to gain traction in organization and management studies. In this section, I unpack some of the research avenues and questions that this approach offers.

Phenomena that have for long been central to organization theory, such as structure, interaction, technology, power and knowledge are deeply affected by digital transformations (Ahrne, Aspers, & Brunsson, 2015; Plesner & Husted, 2020). With the growing digitalization, fundamental aspects of organization take on new shapes and get entangled with data, algorithms and technologies. Even before the coronavirus pandemic hit in 2020, many organizations and workplaces already relied on digital solutions, and tools and technologies from Silicon Valley tech companies continue to seep into public organizations (e.g. schools), state agencies and democratic institutions of all sorts. As digital architectures become critical infrastructures for both private companies and public institutions, multiple organizational questions about legitimacy, power and knowledge become salient. Such questions are only partly about technology and cut to the core of how organizations engage with the world around them and make sense of social realities. Organizational research may explore the practical ways in which processes of datafication reduce work practices, human interactions and organizational environments to digital traces that can be processed and visualized algorithmically. To this end, the vocabulary of ‘prisms’ and ‘refractions’ that I have offered may be useful, especially in combination with Power’s (2022) conception of the organizational implications of an ‘economy of traces’ and Zuboff’s (2020b) concern with ‘epistemic inequality’. Such work could unpack fundamental organizational developments that at present are clouded by the hyperbole around ‘artificial intelligence’ and fears about the dominance of ‘tech giants’. The organizational transformations we face are both more concrete and far-reaching when digital architectures of visibility make their way into the core of organizations, and come to shape how they produce knowledge, manage their affairs and engage with the world around them.

Such forms of knowledge production and intervention seek to identify patterns, highlight deviations and institute particular sets of norms and values, often by the use of proxies such as ‘digital doubles’ that stand in for the person or object under scrutiny (Mulvin, 2021). With digital refractions upon refractions via proxies, organizational life may become increasingly distant from lived experiences and immediately observable phenomena. We have yet to study more systematically – and with conceptual finesse – what these developments mean in the context of organization and management studies. In particular, it will be important to show how such ways of knowing and governing involve alternative orientations to time, such as predictive approaches, unsettle established hierarchies between those with and without digital competencies, and install new logics and politics, such as hopes about ‘total certainty’ and hidden forms of behavioral modification (Zuboff, 2019). The skills at play and in need of development are also very different from those of the past. In organizations shaped by digital architectures of visibility, people have to manage and curate data carefully, both to cater to the logics of algorithmic systems, and to make sure that their digital doubles and other proxies are accurate enough to represent who they are and what they do. Also, people must be aware of how algorithms operate and shape our understandings of organizational phenomena in new ways. These ‘machineries of knowing’ require new forms of knowledge production and assessments of what counts as knowledge that organizational research has yet to describe and conceptualize. Finally, in organizations marked by hypervisibility, questions about security and privacy move to the fore. When everything becomes datafied, sensitive information can leak and private matters can made public or commercialized in ways that were unimaginable previously. Such organizational practices have remained largely unaccounted for in research and may help us think differently about competencies in a digital era (Flyverbom, Leonardi, Stohl, Stohl, & Treem, manuscript).

Furthermore, digital, datafied refractions of organizational processes not only have consequences inside organizations, but also institutionalize novel forms of organizing across different settings. Future work may consider what happens when digital architectures and their logics travel from one domain to another, and when refractions are stacked on top of each other. What once was primarily contained in particular types of digital spaces, such as Google’s search engines and ad-business, or Facebook’s social media platforms, are now travelling to multiple other domains. Not only are their services and tools becoming the infrastructural conditions of news, music and other cultural domains, but these technologies are also seeping into the public sector, such as when police forces and security agencies use digital systems to develop new data-driven approaches to their work, or when technological systems developed for one purpose travel to others. Such examples include the use of earthquake detection systems to predict crime and the use of Google’s Ads system to predict and prevent online radicalization (Flyverbom & Schade, forthcoming). Organizational research may investigate how digital architectures, including data sets, algorithms and systems travel across organizational and institutional settings, when commercial systems become integrated into the public sector or when different types of organizations rely on similar ways of working with data and algorithms. Such situated or comparative work may attune us to the consequences of technology for new forms of institutionalization and standardization that will shape organizations in years to come. A better understanding of how data-driven, algorithmic approaches transform organizational domains is central, if we want to both rely more on digital architectures and set limits to their influence.

These are important issues to theorize in organization and management studies, both because they may reinvigorate well-established concerns with knowledge production, power and control, accountability and technological changes, and because they spark innovative theoretical reflections and debates at the intersections of organizing, technology and society.

Concluding Reflections

Digital architectures surround us and fundamentally impact organizations and processes of organizing. Data, algorithms and technologies are increasingly woven into the fabric of our personal lives, our organizations and society more broadly. Still, we are only beginning to conceptualize their shape and workings, and how they condition organization and management. In an attempt to ignite more theoretical and empirical work on these topics, this essay developed an encompassing conceptualization of contemporary digital spaces. This conceptualization builds on existing work on datafication, algorithms and digital transformations, including Zuboff’s account of surveillance capitalism. Focusing on what I term ‘digital architectures of visibility’, the essay has made two conceptual moves: The first move is to extend the scope of analysis to capture the components and contours of digital architectures. In this manner, it brings together different dimensions – often treated separately in specialized disciplines – in an encompassing conceptualization of digital architectures. The second move digs into the inner workings of digital architectures by articulating how ideas about digital prisms and visibility management can be used to make sense of processes of data extraction, algorithmic sorting and logics. The point of this part of the conceptualization is to elaborate on how digital architectures condition new forms of seeing, knowing and governing that make the management of visibilities more important than ever. My hope is that this conceptualization will serve as a source of inspiration and as a trigger for research and theory development that zooms in on the shape and workings of digital architectures of visibility as these affect organizations and processes of ordering and organizing in society.