This article responds to recent debates in critical algorithm studies about the significance of the term “algorithm.” Where some have suggested that critical scholars should align their use of the term with its common definition in professional computer science, I argue that we should instead approach algorithms as “multiples”—unstable objects that are enacted through the varied practices that people use to engage with them, including the practices of “outsider” researchers. This approach builds on the work of Laura Devendorf, Elizabeth Goodman, (...) and Annemarie Mol. Different ways of enacting algorithms foreground certain issues while occluding others: computer scientists enact algorithms as conceptual objects indifferent to implementation details, while calls for accountability enact algorithms as closed boxes to be opened. I propose that critical researchers might seek to enact algorithms ethnographically, seeing them as heterogeneous and diffuse sociotechnical systems, rather than rigidly constrained and procedural formulas. To do so, I suggest thinking of algorithms not “in” culture, as the event occasioning this essay was titled, but “as” culture: part of broad patterns of meaning and practice that can be engaged with empirically. I offer a set of practical tactics for the ethnographic enactment of algorithmic systems, which do not depend on pinning down a singular “algorithm” or achieving “access,” but which rather work from the partial and mobile position of an outsider. (shrink)

This article considers the issue of opacity as a problem for socially consequential mechanisms of classification and ranking, such as spam filters, credit card fraud detection, search engines, news trends, market segmentation and advertising, insurance or loan qualification, and credit scoring. These mechanisms of classification all frequently rely on computational algorithms, and in many cases on machine learning algorithms to do this work. In this article, I draw a distinction between three forms of opacity: opacity as intentional corporate or state (...) secrecy, opacity as technical illiteracy, and an opacity that arises from the characteristics of machine learning algorithms and the scale required to apply them usefully. The analysis in this article gets inside the algorithms themselves. I cite existing literatures in computer science, known industry practices, and do some testing and manipulation of code as a form of lightweight code audit. I argue that recognizing the distinct forms of opacity that may be coming into play in a given application is a key to determining which of a variety of technical and non-technical solutions could help to prevent harm. (shrink)

The study of social inequality and stratification has long been at the core of sociology and the social sciences. In this article, I argue that certain tendencies have become entrenched in our dominant paradigm that leave many researchers pursuing coarse-grained analyses of how difference relates to inequality. Centrally, despite the importance of categories and categorization for how researchers study social inequality, contemporary theories of categories are poorly integrated into conventional research. I contend that the widespread and often unquestioned use of (...) state categories as categories of analysis reinforces these tendencies. Using research on colorism as an inspiration, I highlight several components of what I call the infracategorical model of inequality, which urges researchers to disaggregate difference by shifting our focus from membership in categories to the cues of categories, membership in subcategories, and perceived typicality. (shrink)

Progressing beyond the given has been a key modern tendency. Yet modern societies are currently facing the problem of how to put limits on progress, expansion, and growth, live within them, and preserve (rather than transcend) the present. Drawing on economic sociology scholarship on valuation and morality in economic life, this article develops and applies the term economization to analyze the enactment of limits on progress. The question of end-of-life care—when to stop medical efforts to prolong life, postpone death, and (...) advance the scientific frontier—serves as an illustrative empirical case that sheds light on limit-setting in general. My analysis of this case combines historical, ethnographic, and in-depth interview data on US palliative care clinicians, who specialize in making life-and-death decisions in acute care hospitals. (shrink)

Algorithms, once obscure objects of technical art, have lately been subject to considerable popular and scholarly scrutiny. What does it mean to adopt the algorithm as an object of analytic attention? What is in view, and out of view, when we focus on the algorithm? Using Niklaus Wirth's 1975 formulation that “algorithms + data structures = programs” as a launching-off point, this paper examines how an algorithmic lens shapes the way in which we might inquire into contemporary digital culture.

This article examines the stratifying effects of economic classifications. We argue that in the neoliberal era market institutions increasingly use actuarial techniques to split and sort individuals into classification situations that shape life-chances. While this is a general and increasingly pervasive process, our main empirical illustration comes from the transformation of the credit market in the United States. This market works as both as a leveling force and as a condenser of new forms of social difference. The U.S. banking and (...) credit system has greatly broadened its scope over the past twenty years to incorporate previously excluded groups. We observe this leveling tendency in the expansion of credit amongst lower-income households, the systematization of overdraft protections, and the unexpected and rapid growth of the fringe banking sector. But while access to credit has democratized, it has also differentiated. Scoring technologies classify and price people according to credit risk. This has allowed multiple new distinctions to be made amongst the creditworthy, as scores get attached to different interest rates and loan structures. Scores have also expanded into markets beyond consumer credit, such as insurance, real estate, employment, and elsewhere. The result is a cumulative pattern of advantage and disadvantage with both objectively measured and subjectively experienced aspects. We argue these private classificatory tools are increasingly central to the generation of "market-situations", and thus an important and overlooked force that structures individual life-chances. In short, classification situations may have become the engine of modern class situations. © 2013. (shrink)

In business and government, databases contain large quantities of digital transactional data (purchases made, services used, finances transferred, benefits received, licences acquired, borders crossed, tickets purchased). The data can be understood as ongoing and dynamic measurements of the activities and doings of people. In government, numerous database devices have been developed to connect such data across services to discover patterns and identify and evaluate the performance of individuals and populations. Under the UK’s New Labour government, the development of such devices (...) was part of a broader policy known as ‘joined-up thinking and government’. Analyses of this policy have typically understood joining up as an operation of adding together distributed data about subjects, which can then be used in the service of government surveillance, the database state or informational capitalism. But rather than such technical or managerialist analytics, I argue that topological analytics capture what these database devices enact and do: they materialize the ‘individuality’ of subjects in intensified, distributed and fluctuating ways and materialize and intensify a logic of what Deleuze describes as modulating controls. Through examples of UK New Labour social policy initiatives over the past decade, I argue that topological analytics can account for these as immanent rather than exceptional properties of database devices and, as such, part and parcel of a governmental logic and ontology of subjects. (shrink)

The welfare state has come under severe pressure internationally, partly for the well-known reasons of slowing economic growth and declining confidence in the public sector. According to the influential social theorist Pierre Rosanvallon, however, there is also a deeper and less familiar reason for the crisis of the welfare state. He shows here that a fundamental practical and philosophical justification for traditional welfare policies--that all citizens share equal risks--has been undermined by social and intellectual change. If we wish to achieve (...) the goals of social solidarity and civic equality for which the welfare state was founded, Rosanvallon argues, we must radically rethink social programs. Rosanvallon begins by tracing the history of the welfare state and its founding premise that risks, especially the risks of illness and unemployment, are equally distributed and unpredictable. He shows that this idea has become untenable because of economic diversification and advances in statistical and risk analysis. It is truer than ever before--and far more susceptible to analysis--that some individuals will face much greater risks than others because of their jobs and lifestyle choices. Rosanvallon argues that social policies must be more narrowly targeted. And he draws on evidence from around the world, in particular France and the United States, to show that such programs as unemployment insurance and workfare could better reflect individual needs by, for example, making more explicit use of contracts between the providers and receivers of benefits. His arguments have broad implications for welfare programs everywhere and for our understanding of citizenship in modern democracies and economies. "For more than two decades Pierre Rosanvallon has been analyzing the development and the crisis of the 'welfare state,' combining precise, specific knowledge with philosophical and historical depth in a way that is rare among social policy analysts.... [A] subtle and informed book."--From the foreword by Nathan Glazer. (shrink)

Although overarching if not foundational conceptualizations of digital governance in the field of critical data studies aptly account for and explain subjection, calculated resistance is left conceptually unattended despite case studies that document instances of resistance. I ask at the outset why conceptualizations of digital governance are so bleak, and I argue that all are underscored implicitly by a Deleuzian theory of desire that overlooks agency, defined here in Foucauldian terms. I subsequently conceptualize digital governance as encompassing subjection as well (...) as resistance, and I cast the two in relational perspective by making use of the concepts “affordance” and “assemblage” in conjunction with multiple subjectivities and Foucault's view of power as productive as well as his view of resistance as an “antagonism of strategies” in his late scholarship on resistance, ethics, and subjectivity. I offer examples of salient modes of what I call “productive” resistance, and from a Foucauldian perspective I explain how each mode targets and subverts technologies of repressive power to produce new elements of the digital environment and construct new truths. I conclude by recognizing the agency embodied in resistance as an end in itself, but I also consider that modes of productive resistance can have extrinsic value as they affect the fluid interaction among elements of the digital environment, potentially disrupting the presumed structure of dominance and dependence, and opening our conceptualization of algorithmic life to hopeful possibilities for change. (shrink)

Machine learning algorithms reshape how people communicate, exchange, and associate; how institutions sort them and slot them into social positions; and how they experience life, down to the most ordinary and intimate aspects. In this article, we draw on examples from the field of social media to review the commonalities, interactions, and contradictions between the dispositions of people and those of machines as they learn from and make sense of each other.

While it is often claimed that the emerging digital governmentality functions as a new apparatus of surveillance, the aim of this paper is to characterise this regime in relation to Foucault’s disc...

What makes insurance special among risk technologies is the particular way in which it links solidarity and technical rationality. On one hand, within insurance practices ‘risk’ is always defined in technical terms. It is related to monetary measurement of value and to statistical probability calculated for a limited population. On the other hand, and at the same time, insurance has an inherent connection to solidarity. When taking out an insurance, one participates in the risk pool within which each member is (...) reciprocally responsible for others’ risks. The combination of technical controllability and solidarity has made insurance a successful tool for governing welfare societies during the twentieth century. From the point of view of business ethics, it is interesting that, as we argue in this article, the connection between insurance and solidarity is not limited to social welfare assemblages, but is evident in relation to private insurance as well. At the same time, however, it is important to understand that insurance does not advance all forms of solidarity. Hence, this theoretical article analyzes the specific conceptions of solidarity that the different forms of insurance practice produce. Particular emphasis is put on the distinction between ‘chance solidarity’ and ‘subsidizing solidarity’. The main questions of the article are: What kinds of conceptions of solidarity are built in the insurance technology? And how are the limits of solidarity defined and justified in different forms of insurance? (shrink)

The aim of this article is to assess the impact of Big Data technologies for insurance ratemaking, with a special focus on motor products.The first part shows how statistics and insurance mechanisms adopted the same aggregate viewpoint. It made visible regularities that were invisible at the individual level, further supporting the classificatory approach of insurance and the assumption that all members of a class are identical risks. The second part focuses on the reversal of perspective currently occurring in data analysis (...) with predictive analytics, and how this conceptually contradicts the collective basis of insurance. The tremendous volume of data and the personalization promise through accurate individual prediction indeed deeply shakes the homogeneity hypothesis behind pooling. The third part attempts to assess the extent of this shift in motor insurance. Onboard devices that collect continuous driving behavioural data could import this new paradigm into these products. An examination of the current state of research on models with telematics data shows however that the epistemological leap, for now, has not happened. (shrink)

The use of algorithmic prediction in insurance is regarded as the beginning of a new era, because it promises to personalise insurance policies and premiums on the basis of individual behaviour and level of risk. The core idea is that the price of the policy would no longer refer to the calculated uncertainty of a pool of policyholders, with the consequence that everyone would have to pay only for her real exposure to risk. For insurance, however, uncertainty is not only (...) a problem – shared uncertainty is a resource. The availability of individual risk information could undermine the principle of risk-pooling and risk-spreading on which insurance is based. The article examines this disruptive change first by exploring the possible consequences of the use of predictive algorithms to set insurance premiums. Will it endanger the principle of mutualisation of risks, producing new forms of discrimination and exclusion from coverage? In a second step, we analyse how the relationship between the insurer and the policyholder changes when the customer knows that the company has voluminous, and continuously updated, data about her real behaviour. (shrink)