2.1 Introduction

Data-driven processes extend deeply into contemporary political economies, structuring flows of economic and social activity in ways both seen and unseen. Those processes have fueled growing interest by scholars and policymakers in the idea of infrastructure, both as an analytical construct and as an object of governance. In this chapter, I use the relationship(s) between data and mobility through physical spaces to explore why the characterization of data as infrastructure might matter. I map three co-constituting processes that are reshaping flows of people and goods through physical spaces – the datafication of infrastructure, the extraction and appropriation of data, and the infrastructuralization of data – and consider the implications for both understandings of infrastructure and debates about governance of data-driven processes.

As sets of structured data flows layered over other infrastructures have multiplied and intersected with practices of data extraction and appropriation, some of those data flows – in particular, data flows maintained by entities with the resources to manipulate them in real time and at scale – have come to function as infrastructures in their own right. These shifts have destabilized existing settlements regarding the movement of people and goods. Data infrastructures (are designed to) scale both vertically and laterally; they are nimble, flexible, and adaptable to new uses. Therefore, although they lack the brute fixity of physical infrastructures such as roadways and bridges, they facilitate assertions of power in ways that physical infrastructures alone cannot, and they are also vulnerable to commandeering and cooptation. We will see that the emergence of data infrastructures has both enabled control and coordination of formerly distinct activities and afforded points of entry for new assertions of political and geopolitical power.

Conceptualizing data as infrastructure sheds useful light on contemporary legal and regulatory debates about platform power and ways of constraining it. In particular, the infrastructure lens helps to explain some of the challenges that efforts at post hoc regulatory supervision of platform activities have faced, and it also suggests some important questions about the newer, platform-specific regulatory models that are now beginning to emerge.

2.2 Data as Infrastructure?

Although for many, the term “infrastructure” conjures up images of roads, bridges, pipes, and electrical cables, the term has long been used in a more capacious way within scholarly and policy discourses to describe a broader range of standardized arrangements that facilitate flows of human activity and exchange across spaces and scales. In particular, over the last several decades, contemporary thinking about the role of infrastructure in the networked information economy has broadened beyond physical installations (now including digital communications networks) to consider the infrastructural effects of certain types of intangible resources. Today, networked information infrastructures, platforms, and datasets have emerged as important focal points for both analysis and contestation.

One starting point for studying infrastructures, familiar to legal scholars in subfields such as intellectual property and technology law, antitrust, and law and economics, has been Frischmann’s influential formulation, in which infrastructures have three characteristics: they are consumed nonrivalrously (for at least some ranges of use); social demand is driven primarily by downstream productive activities that require the resource; and they serve as inputs into a wide range of private, public, and social goods and services.Footnote ¹ One noteworthy aspect of this definition is the extent to which it detaches infrastructure from physical or tangible form; so, for example, an infrastructure might be a network of roads or fiber optic cables, or it might be a common linguistic or cultural environment. Another is the extent to which it is agnostic about management structure; that agnosticism enables Frischmann to explore how choices about ownership and management of infrastructural resources shape the patterns of downstream relationships and uses.

It is also important, however, to attend to the ways that infrastructures emerge and evolve over time and to the particular technological and sociotechnical configurations they impose – topics of pressing importance in an era of rapid sociotechnical change. Because of their scale and the investment they require, infrastructures can be the objects of deliberate construction. But infrastructures also can emerge in more piecemeal and less coordinated ways. They also can emerge via carefully strategized efforts at capture. And infrastructuring projects of all sorts have sociotechnical predicates that reflect both consciously determined priorities and unexamined choices.

Work in technology studies emphasizes these considerations and foregrounds the interplay between emergence and habituation. In Star and Ruhleder’s influential formulation, infrastructure becomes infrastructure when it sinks into the background, becoming a taken-for-granted way of mediating “the tension between local and global” in human and social relations.Footnote ² Star and Ruhleder urge, therefore, that the important question for infrastructure studies is not “what is an infrastructure?” but rather “when is an infrastructure?”Footnote ³ Factors relevant to answering that question include reach and scope, embodiment of shared standards, habituation, and transparency while working as expected.

Last but not least, infrastructures have profound implications for the shape of governance and for the distribution of power within human societies. Even the idea of infrastructure has a complex political and ideological history. Coined by nineteenth-century French railway engineers as a way of describing the purpose and function of the nascent networks they were building, it was appropriated by Anglophone architects of the post–World War II military–industrial complex to describe and mobilize public–private partnerships for large-scale engineering projects to be constructed in the service of an emergent liberal-democratic global order.Footnote ⁴ Globally, decisions about where and how to provide infrastructures for essential services have long mediated relations of inclusion and exclusion, “articulat[ing] social relations to make a variety of social, institutional, and material things (im)possible.”Footnote ⁵

A broader formulation that attempts to encompass such considerations alongside the process- and practice-based insights of technology studies scholars is supplied by communications scholars Hallinan and Gilmore, who define infrastructure as “a technique of governance, a means of existential support, and a site of contestation.”Footnote ⁶ In particular, they observe that infrastructures both underwrite formalized expectations about how the activities they facilitate are supposed to work and instill imaginaries that structure conscious and unconscious reasoning about the purpose(s) and social meaning(s) of those activities. As anthropologists Anand, Gupta, and Appel put it, infrastructures represent “critical locations though which sociality, governance and politics, accumulation and dispossession, and institutions and aspirations are formed, reformed, and performed.”Footnote ⁷

Turning more specifically to investigations of networked digital technologies, researchers in a number of fields have begun to interrogate the role(s) that digital platforms now play as (contested) infrastructures within the political economies of informational capitalism (in the USA) and party–state capitalism (in China). The largest digital platforms are increasingly indispensable to a wide variety of everyday activities in ways that illustrate Hallinan and Gilmore’s conception of materially and ideologically embedded governance and Star and Ruhleder’s criteria of embeddedness, habituation, transparency, and visibility upon breakdown.Footnote ⁸ In particular, digital platforms enact the themes of scale, formalization, and imaginary in ways that are organized around logics of modularization, algorithmic intermediation, and control.Footnote ⁹

Other researchers have focused more squarely on infrastructuring work performed by and through data, or what Flyverbom and Murray have called “datastructuring.”Footnote ¹⁰ The idea of data as exerting (infra)structuring effects might seem paradoxical. On the one hand, “data” are not conventionally understood as having any particular structure. On the other, it is now well recognized that the term “raw data” is a misnomer: that data always reflect both their originating contexts and more particular choices about what is relevant and important to collect and measure.Footnote ¹¹ This is equally true of the very large datasets on which machine learning and artificial intelligence processes are trained.Footnote ¹² Control of very large datasets and of the processing power necessary to manipulate them in real time also have become important factors distinguishing the largest and most dominant platform providers from others.

At a nexus between these two avenues of inquiry sit the protocols by and through which data are configured, transmitted, and processed. So, for example, platform studies researchers locate proprietary protocols for data extraction and exchange at the operational center of contemporary platform-based information and communication infrastructures.Footnote ¹³ To similar effect, work at the intersection of technology studies and surveillance studies explores the protocological optimization of infrastructures for data-driven surveillance as a strategy for ensuring a continual flow and supply of goods and services.Footnote ¹⁴

Returning, finally, to scholarship in law and cognate fields, other work on data as a modality of governance underscores the linkages between data and conceptions of appropriate political and economic ordering and highlights law’s institutional entanglements with these questions. So, for example, in earlier work I have analyzed the way that the legal positioning of data as raw material available for taking and appropriation enables the large-scale extractive practices of contemporary informational capitalism.Footnote ¹⁵ As Couldry and Mejias have shown, practices of data extraction in the contemporary, globalized information economy also reenact longstanding patterns of colonialist resource appropriation.Footnote ¹⁶ Viljoen’s important work on democratic data governance explores the possible ways that decisions about the appropriate uses of datasets might reflect and reinforce other collective choices.Footnote ¹⁷ Implicit in theories about data appropriation, data colonialism, and the possibility of democratic data governance is the notion that emergent sociotechnical formations surrounding data might do infrastructuring work in service of multiple, divergent political and economic futures.

In the balance of this chapter, I bring these lines of inquiry about infrastructures, platforms, data, protocols, and governance into conversation with one another, with particular focus on the infrastructural character of the massive datasets amassed by dominant, platformized providers of services for facilitating the movement of people and goods.

2.3 Mapping Mobilities

During the first two decades of the twenty-first century, the patterns that structure flows of people through physical spaces have changed dramatically, in ways that reflect three interlocking processes. The first process is the datafication of existing infrastructures for physical mobility, facilitated by networked interconnections between trackers in mobile devices and sensors installed at fixed locations including businesses and highway and transit systems. The second process involves the extraction and appropriation of mapping and mobility data by dominant providers of mapping and geolocation services. The third is the resulting infrastructuralization of mapping and mobility data: the massive datasets amassed by dominant providers of mapping and mobility services now perform the scale-spanning, existential support, and governance functions of infrastructures. At the same time, the development of data infrastructures for mapping and mobility as privatized, centrally controlled resources has shaped the kinds of mapping and mobility they afford.

Scholarship in multiple fields has explored various effects of digitally mediated transformations in human mobility. Some scholars focus on the ways that digital mediation changes the lived experience of physical spaces for individuals and communities.Footnote ¹⁸ Others explore shifts in patterns of driving, riding, and transit use, juxtaposing the ascendancy of piecework ride services like Uber and Lyft with declining use of both regulated taxicabs and public transit services.Footnote ¹⁹ Digital technologies alone did not produce these changes in the look and feel of cities; they also reflect the accumulating effects of a half century of neoliberal economic policies that fueled economic inequality and urban stratification, followed by the knockout punches of pandemic lockdown and a more durable shift toward teleworking and hybrid work arrangements.

Underlying all of these developments, however, is an infrastructural transformation that involves mapping data overlaid on roads and walkways and correlated in real time with mobility data extracted from mobile digital devices. In developed Western political economies, the principal actor in this story is Google. For nearly two decades, Google has aimed to comprehensively map the world’s public roadways and provide seamless navigation assistance that includes real-time information about traffic and road conditions. Most visibly, millions of individual drivers and pedestrians use Google’s flagship Maps app and the route-planning app Waze, which Google acquired in 2013, on a daily basis to locate destinations, determine probable driving and walking times, and plot routes. Uber has enterprise-level licenses to incorporate Google’s mapping data into its matching and routing algorithms, as do other enterprises that need to coordinate vehicle fleets.Footnote ²⁰ And a vast assortment of other mobile apps that incorporate location services – for transit, parking, location-based reminders, and many more – use software developer kits (SDKs) supplied by Google that embed mapping functionality. Mobile users of Apple devices rely principally on mapping functionality developed by Apple, and Apple supplies its own SDKs to app developers.

In China, the principal actors are different. Although most smartphones used in China run Google’s Android mobile operating system, Google’s other services, including Maps, are blocked in China. Android users wanting to locate destinations and plan driving, transit, or walking routes most typically choose between the mapping and mobility apps provided by dominant search engine Baidu and dominant e-commerce and mobile payment provider Alibaba Group.Footnote ²¹ (In countries across the Global South where both Google and Chinese platform companies operate, Google supplies mapping and mobility services to Android users.) Apple’s mapping functionality works in China, for the most part. The dominant ride-hailing company, Didi Chuxing, has its own rich troves of mapping and mobility data covering Chinese roadways.Footnote ²²

One might observe that these preliminary descriptions of data-driven mobility services feature public physical infrastructures working as they should – that is, supplying nonrivalrous inputs into downstream patterns of production and use that have shifted in ways responsive to changing societal demand. From that perspective, drivers, transit riders, and pedestrians are simply using publicly maintained roads and sidewalks as they were meant to be used, and information services such as Google’s, Baidu’s, Alibaba Group’s, and Didi Chuxing’s simply facilitate their usage more efficiently. One also might observe that consolidation of real-time mapping data under the control of dominant service providers reflects both the winner-takes-all character of network-and-standard-based processes and the background political economic conditions of informational capitalism.Footnote ²³ Shared conventions about mapping that are decades and even centuries old might support multiple mapping services organized around shared but open protocols. However, ever-increasing reliance on both online search and mobile connectivity, along with the emergence of dominant providers for such services, have fueled the emergence of dominant providers of maps and mapping. Additionally, manipulating very large quantities of data in real time and across geographic scales requires processing power that very few entities are in position to command.

But Google, Apple, and their Chinese counterparts are not simply facilitating usage of public infrastructures; instead, they are playing out complicated dances of datafication, data extraction and privatization, and infrastructural transformation. Real-time mapping by and through platformized controllers of mapping and mobility data has become an infrastructural service in all of the senses that the previous section described. It is both a nonrivalrous demand-side input into many downstream productive activities and an embedded and taken-for-granted means of mediating mobility across local, regional, national, and global scales. It is also a technique of governance that shapes mobility throughout physical spaces. As it operates, real-time mapping constitutes and iteratively reinforces formalized expectations about always-on navigation assistance and seamless interconnections between transit options; simultaneously, it constitutes a powerful imaginary that structures conscious and unconscious reasoning about how mobility works and should work in highly informationalized societies.

In the case of Google, these processes have unfolded at the intersection of two recursive dynamics that are hallmarks of capitalist political economies. One involves systematic extraction and appropriation of public domain resources as inputs to private productive activity – typically, without acknowledgment of the value of the public contribution. As Mariana Mazzucato has shown, well-rehearsed narratives about private innovation tend to downplay (or ignore entirely) the role that public innovation has played in catalyzing technological development.Footnote ²⁴ Data infrastructures are no different. Although Google now relies on updates from billions of user devices to refine its real-time mapping services, it initially developed those services using troves of government data provided under open access arrangements.Footnote ²⁵ The second recursive dynamic involves systematic application of concepts of “public” and “private” to delineate resource ownership and enclosure of those resources deemed private or proprietary.Footnote ²⁶ In particular, Google (and Apple) continually harvest, enclose, and mine data about individual and social mobility, situating the datafication of mapping and mobility firmly within logics of extraction and capital accumulation.

In China, the party-state is a more active partner in processes of datafication, data extraction, and privatization. Use of government surveying and mapping data requires government authorization, and dominant economic actors also are more directly accountable to the party-state for their decisions and actions. In particular, as dominant technology platform companies have emerged, the party-state has both supported their global expansion and periodically asserted its ultimate authority over the design and deployment of data-driven services and data infrastructures. So, for example, in 2020 concern over the growing power of e-commerce giant Alibaba Group led the party-state to order postponement of the company’s planned Shanghai and Hong Kong initial public offerings and impose unspecified personal sanctions on its CEO, Jack Ma.Footnote ²⁷ Similarly, in 2021, after Didi Chuxing’s initial public offering on Nasdaq prompted concern about the company’s growing clout and its control of data related to the mobility patterns of party officials, the party-state imposed hefty penalties including a two-year moratorium on enrolling new users.Footnote ²⁸ More recently, the party-state has enlisted Alibaba Group’s and WeChat’s support in its effort to develop a state-sponsored ride-hailing service that might compete with Didi’s.Footnote ²⁹

The interlocking dynamics of data extraction and privatization and the emergent infrastructuralization of mapping and mobility data have reshaped human relationships to the physical environment. Data infrastructures for mapping and mobility suggest and reinforce route choices based on their own optimization parameters without regard to users’ own knowledge of their communities or to embodied experiences of physical geographies.Footnote ³⁰ Traffic flows mediated by data infrastructures also can be insensitive to local planning considerations. So, for example, the tendency of both Waze and (albeit to a lesser extent) Maps to route traffic around major thoroughfares during peak congestion periods has provoked repeated disputes with local authorities trying to minimize traffic through residential areas and keep traffic through smaller commercial districts to manageable levels.Footnote ³¹

At the same time, infrastructuralized mapping and mobility data now implicate a range of activities that formerly may not have been understood as linked or adjacent. Some activities correspond to imperatives of capital accumulation. To maintain and sharpen their real-time mapping capabilities, dominant platform providers of mapping services offer mobile app developers SDKs that embed location tracking functionality, enabling them to capture and harvest additional traces of individual movement through physical spaces and push location-linked information, including advertisements, back to mobile devices.Footnote ³² They combine the data flows harvested in real time from users’ devices with other data collected via their search, social networking, online shopping, and advertising services. In particular, dominant platform providers of mapping services like Google, Apple, Alibaba Group, and Baidu also have rich datasets covering local businesses that pay for advertising and/or accept mobile payments through services offered by the same dominant providers. This enables them to offer mobile users pointers to conveniently located purveyors of goods and services, which in turn underwrites expectations about seamless, invisible integration of mobility and commerce and instills an imaginary within which mobility and commercial convenience are both seamlessly entwined and infrastructurally provisioned. Even when users have withheld consent to merge their own data in this way, the provider can make predictive judgments about what users might want to see along their routes based on its data-driven profiling of similar users.

Nascent in the fusion of infrastructuralized mapping and mobility data with location-driven advertising, moreover, is potential for assertion of a broader range of logics and priorities. For example, given the other kinds of data about user preferences and affiliations that dominant platform providers already possess, it is not too difficult to envision mapping and mobility infrastructures services capable of routing users toward or away from minority-owned businesses or businesses friendly to their particular political tribe, or toward or away from political protests based on their revealed affinities.Footnote ³³ Assuming that such capabilities were disclosed, moreover, they are arguably consistent with an imaginary of digital mobility within which convenience – now more broadly defined to encompass a range of personal and political priorities – is infrastructurally provisioned. As a different example, although Western governments do not directly dictate the design of mapping and mobility services, their wishes and priorities regarding data retention and production do shape the design of underlying data infrastructures for mapping and mobility.Footnote ³⁴ The now-routine production of so-called geofencing information recording all mobile devices present at particular locations and times has already destabilized longstanding legal settlements regarding fundamental rights against targeted state surveillance.Footnote ³⁵ It is not too difficult to imagine experiments in so-called smart city governance that have the unintended (or intended) effect of enabling deepened and/or selectively targeted controls of mobility.

In China, where the party-state has supported a more pervasive datafication agenda infused with its own domestic and geopolitical aims, these capabilities are no longer only nascent. So, for example, during the COVID-19 pandemic, the dominant platform providers of mobile payment services, Alibaba Group and WeChat, developed so-called health codes that functioned as mobility passports. Links between public health databases, government ID databases, and transit system data infrastructures facilitated a new regime of permissioned mobility, in which a green code was required even to move around near one’s home and yellow and red codes signaled progressively more draconian restrictions. Party-state officials have repurposed these mobility controls to anticipate and disable planned protest gatherings.Footnote ³⁶

In sum, contemporary data-driven mobility is both constituted and shaped by emergent data infrastructures for mapping and mobility, and data infrastructures embody distinct kinds of functionality and flexibility. Data infrastructures for mapping and mobility extend vertically, in stacks that superimpose data-driven logics and imaginaries over preexisting physical infrastructures such as roadways and transit systems, and laterally, in webs that extend tendrils of infrastructural logic into domains that formerly were more distinct, such as advertising and political protest. Their capacity to scale vertically and horizontally makes them amenable to configuration, reconfiguration, and control in ways that underlying physical infrastructures typically are not. This destabilizes preexisting, often tacit, settlements regarding the relationship(s) between infrastructure and governance.

2.4 Cartographies of Consumption

During the first two decades of the twenty-first century, flows of retail goods through physical spaces also have changed dramatically in ways that reflect the three processes identified in the preceding section. The datafication of consumption and production and the extraction and appropriation of consumption- and production-related data by two increasingly dominant providers of e-commerce services – Amazon and Alibaba Group – have fueled an increasingly pervasive infrastructuralization of commerce-related data that is remaking the commercial world. As with shifts in patterns of human mobility, the results illustrate the profoundly destabilizing potential of data infrastructures, which inheres in their capacity for lateral expansion and vertical extension and in their amenability to centralized control. The emergence of commerce-related data infrastructures is fueling far-reaching vertical and lateral reconfigurations of production and consumption relations. These shifts, covering the interlinked categories of product search, promotion, sale, and order fulfillment, have begun to reshape what is produced, how it is exchanged, and how complex sets of human activities surrounding production and consumption are organized.

The contemporary history of data-driven transformations in flows of physical goods is a story of two e-commerce giants: Amazon, which dominates retail traffic in the US, and Alibaba Group, which dominates retail traffic worldwide. In the US, Amazon now controls more than one third of overall retail sales traffic for a bewildering variety of goods ranging from books and consumer electronics to clothing and sporting equipment to personal care products to groceries.Footnote ³⁷ It is routinely linked to the stark decline of retail sales in shopping malls and downtowns throughout the country. Here again, changing retail sales patterns reflect many causal factors, including a half century of neoliberal trade policies designed to facilitate global supply chains for mass-produced goods sold through “big box” stores, growing economic inequality, rising urban rents, and changes in consumption patterns.Footnote ³⁸ Amazon, however, has been uniquely positioned to capitalize on these developments by virtue of the very large datasets it has amassed and the logistic capabilities it has used the data to develop.

Alibaba Group, meanwhile, operates in many parts of the world where there was less of a preexisting brick-and-mortar retail footprint to displace. Throughout its expansion, it has focused on tapping the consuming potential of customers without credit cards and unbanked customers and businesses traditionally operating in cash-based, local economies.Footnote ³⁹ The AliPay mobile app, designed principally for use by Chinese nationals, allows consumers to make purchases via Alibaba Group’s Chinese language Taobao platform using a linked Chinese bank account. AliExpress, the Alibaba Group’s global e-commerce platform designed for business-to-consumer transactions with consumers outside China, accepts both credit cards and prepaid cards denominated in a variety of currencies. (The flagship Alibaba platform, designed primarily for higher-volume business-to-business transactions between Chinese merchants and purchasers around the globe, requires at least a bank account.) AliExpress and Alibaba are used widely throughout South Asia, Australia, Africa, and the Middle East and have growing footprints in the Global North. Together, the Alibaba Group’s services account for nearly half of retail traffic in China and control an estimated 25 percent share of retail sales worldwide.Footnote ⁴⁰

Both Amazon and Alibaba Group have invested heavily in datafication and data extraction. The two enterprises track individual and business users of their websites and mobile apps extensively, in ways that follow the patterns described in Section 2.2 above. The massive datasets amassed by the two e-commerce giants in turn have become increasingly indispensable preconditions for directing and receiving flows of goods across local, regional, national, and global scales. For consumers, they underwrite formalized expectations of next-day availability and constitute accompanying imaginaries of networked, always-on efficiency and convenience that characterize contemporary consumption patterns. For merchants, they underpin increasingly formalized expectations around the conditions of access to customers, coupled with a potent imaginary of geographically limitless commercial reach.

Also following the general pattern described in Section 2.2 above, as data infrastructures for production and consumption have coalesced, they have extended lateral tendrils into formerly distinct activities. In the case of production- and consumption-related data, these extensions have unfolded in ways enabled and constrained by different patterns of regulatory scrutiny and laxity. As competition regulators in the US and Europe have long recognized, Amazon’s extensive datasets have enabled it to diversify its product offerings and to develop powerful, data-driven strategies for cross-promotion of its own products and services. In particular, it has leveraged its proprietary Prime membership program to engage in price discrimination that privileges its own offerings.Footnote ⁴¹ Alibaba Group engages in similar patterns of diversification and cross-promotion. However, as it has drawn the scrutiny of party functionaries worried about the power and transnational reach of its data harvesting and processing operations, it now must conduct its domestic operations under data protection constraints similar to those that Amazon has confronted in Europe, plus additional data nationalization and security review requirements that serve the party-state’s data sovereignty goals.Footnote ⁴²

From the beginning, credit provision and payment processing have played central roles in Alibaba Group’s growth. In particular, as noted above, it has emphasized data-driven strategies for extending credit and payment services to unbanked individuals and small businesses. Due largely to the fact that the Global North jurisdictions in which it primarily operates regulate credit provision and payment processing more comprehensively, Amazon has lagged behind in financial diversification efforts.

Amazon, meanwhile, which operates in a home jurisdiction lacking modern data protection regulation, has pursued lateral expansions involving sensors that extend datafication and data extraction capabilities more deeply into customers’ homes and lives.Footnote ⁴³ Amazon’s Echo home speaker systems incorporate a voice-enabled assistant, Alexa, through which users can issue purchase orders that Amazon’s shopping portals will fulfill, as well as requests for streamed audio and video content to be delivered via Amazon’s Prime Music and Prime Video services. The Ring doorbell monitors comings and goings and also doubles as a concierge capable of superintending secure package delivery.Footnote ⁴⁴ Echo and Ring, however, are stationary installations. Amazon’s acquisition of iRobot, which manufactures the Roomba automated floor sweeper, enables it to take in-home consumer surveillance and voice recognition mobile, mapping the layouts, contents, and uses of living environments more comprehensively.Footnote ⁴⁵

Other lateral expansions facilitated by infrastructuralized data are merchant facing. For merchants in developed Western economies, opening a store on Amazon Marketplace has become an indispensable means of reaching customers increasingly habituated to and dependent on online shopping. In turn, operating within Amazon Marketplace is easier to do if one purchases back-end hosting and support from Amazon Web Services and if one uses Amazon’s order fulfillment mechanisms to facilitate delivery.Footnote ⁴⁶

Alibaba Group, meanwhile, has taken the Chinese model of petty capitalist production global, linking family- and village-based production enterprises directly with customers at the far ends of global supply chains formerly mediated exclusively by global manufacturing and retail concerns.Footnote ⁴⁷ In this effort, it has benefited from party-state efforts to dis- and re-intermediate legal infrastructures for global trade in the service of its larger geopolitical ambitions. China has constructed special trade zones to facilitate the outward flow of Chinese produced goods to buyers all over the globe. Parties transacting within the trade zones can expect relatively efficient, seamless customs procedures and can call upon special trade courts, distinct from the rest of the Chinese judicial system, to resolve trade-related disputes involving non-Chinese parties.Footnote ⁴⁸ Alibaba Group’s proprietary logistics hubs are both geographically proximate to and operationally linked with the special trade zones.Footnote ⁴⁹

As data infrastructures for production and consumption have scaled vertically and horizontally, they have begun to drive reconfigurations of underlying physical infrastructures for order fulfillment and delivery. Initially, Amazon relied on the US postal service and private shippers for these functions and used data provided by those entities to compute delivery times and shipping costs, while Alibaba Group relied on global shipping networks to connect local petty capitalist enterprises within China to global markets. But Amazon and Alibaba Group are no longer just users of preexisting physical shipping infrastructures. The datafication of processes for order fulfillment has enabled them to underwrite massive rearrangements of those processes designed with their own priorities in mind. Through sophisticated logistics divisions, they coordinate large and growing order fulfillment infrastructures comprised of giant warehouse hubs, fleets of freight planes and trucks, and armadas of delivery vans.Footnote ⁵⁰ Both also have invested directly in cargo ships and related physical infrastructures for container shipping.Footnote ⁵¹

Data-driven rearrangements of physical infrastructures for commercial exchange have produced corresponding rearrangements of human labor, refiguring workers as datafied factors of production called up and deployed on demand. Across regional and sometimes national borders, armies of freelance and seasonal workers move in rhythms dictated by datafied flows of commerce through warehouse hubs and the circuits that connect them. Inside warehouses and freight and delivery vehicles, arrays of sensors enroll human workers within surveillant assemblages that monitor their movements and attention patterns and continually nudge them to maximize task productivity in aid of seamlessness and speed.Footnote ⁵²

A final set of effects is more speculative and flows from platform-based dis- and re-intermediation of information about what is produced and consumed. As the rearrangements of global commerce facilitated by Amazon and Alibaba encounter one another, two formerly distinct models for production and distribution of consumable goods – the industrial model and the petty capitalist model – now appear to be converging. Global supply chains formerly controlled by multinational firms have disaggregated. With relative ease, one can establish a storefront on Amazon and source products through Alibaba or AliExpress. Meanwhile, smaller US and European producers operating via petty capitalist modes of production can access national and global markets via their Amazon storefronts.

In sum, as the enormous and uniquely detailed datasets that Amazon and Alibaba Group have developed about patterns of production and consumption have come to function as infrastructures in their own right, they have catalyzed profound transformations of commercial activity. Those transformations produce inversions in which data-driven processes drive rearrangements of physical installations and flows, including warehouses and the land, air, and sea facilities and routes that connect them. Additionally, as data infrastructures have scaled laterally, merging formerly distinct infrastructures and formerly noninfrastructuralized functions, adjacent service-related infrastructures and activities have become the subject of active data-driven disintermediation and capture strategies.

2.5 Challenges of Governing Data Infrastructures

In the space of a few decades, the dominant digital platform enterprises have come to seem nearly unregulable. The exception, the emerging Chinese platform state, has achieved that status by infusing platform operations with party-state control and deputizing them for authoritarian rule, an approach that is neither viable nor desirable as a way forward for Western legal systems. Viewing platform operations through the infrastructure lens helps to explain some of the challenges that efforts at post hoc regulatory supervision of platform activities in the US and Europe have confronted. The functions performed by data infrastructures for mapping, mobility, and commerce sit at the intersection of several different kinds of legacy regulatory regimes. None, however, speaks to the distinctive considerations raised by infrastructuralization, which involve reorganization around shared standards, embedding, habituation, and transparency to critique. This exercise also suggests some questions that those tasked with implementing Europe’s new experimental regime for platform governance should consider.

First and most obviously, dominant platform control and operation of data infrastructures implicate the kinds of market access concerns that are the traditional subject of antitrust and competition laws. Dominant online platforms, however, have also strained antitrust and competition frameworks to the breaking point. Competition thinking about horizontal and vertical integration of preexisting markets tends to take the contours of existing markets as given. As competition scholars and regulators now widely acknowledge, that approach is not especially useful for thinking about the types of data-driven reconfigurations that platforms effectuate.Footnote ⁵³ Less widely acknowledged is that remedial options such as merger restrictions and breakups are mismatched to the type of infrastructural power that dominant platforms now represent, which inheres in data, protocols, and modular architectures for data collection and exchange. The infrastructure lens highlights these issues, helping to explain how dominant platforms have become ever more deeply entrenched notwithstanding a series of well-intentioned merger reviews and enforcement actions in multiple jurisdictions. Infrastructuring work is difficult to prevent one merger at a time and difficult to unwind one market at a time.

Data infrastructures for mapping, mobility, and commerce also both implicate and frustrate legacy regimes of consumer protection. As I have explained elsewhere, consumer protection regulation traditionally has emphasized information-forcing and information-blocking rules intended, together, to correct market failures, but contemporary platform-based economies comprehensively disable such strategies.Footnote ⁵⁴ This is so in part because they create conditions of infoglut that frustrate regulatory efforts to ensure accurate information about goods and services.Footnote ⁵⁵ To take just one example suggested by Section 2.3, pop-up storefronts governed by boilerplate agreements with Amazon or Alibaba Group and their accompanying data protocols cannot be governed in the same ways that more centralized arrangements might be. More fundamentally, however, platform-based data infrastructures replace market mechanisms with more complex tuning and allocation mechanisms that require different regulatory approaches.Footnote ⁵⁶

The European Union’s new digital economy regulations attempt to create a new framework for structural oversight of platforms that incorporates elements from each of these regulatory traditions while also introducing some new requirements. The Digital Markets Act (DMA) defines obligations for large “gatekeepers” that provide one or more designated “core platform services” in the digital economy.Footnote ⁵⁷ Designated gatekeepers (i.e., the very largest and most dominant platforms) may not combine or cross-use information collected through a “core platform service” with any other service.Footnote ⁵⁸ So, for example, once the DMA takes effect (and following official issuance of gatekeeper designations for the relevant services), Google would be barred from integrating the datasets collected and used by Google Maps and Search and barred from integrating either of those datasets with data collected and used in its digital advertising business. Gatekeepers also may not engage in self-preferencing, whether by imposing tying requirements on individual or business users or by using data generated by their business users to compete with those users.Footnote ⁵⁹ They must afford interoperability to third-party content, apps, and app stores and data portability to individual and business users.Footnote ⁶⁰ Additionally, designated gatekeeper providers of digital advertising services must disclose, on request, daily information about ad prices and publisher remuneration to both advertisers and publishers.Footnote ⁶¹

The Digital Services Act (DSA), meanwhile, defines obligations for “online platforms” that make hosted content available to the public and additional, more stringent obligations for “very large online platforms” that function as central nodes in the informational economy.Footnote ⁶² The DSA’s requirements mostly concern user-facing transparency with regard to removal or recommendation of content (including advertising) and fair process in cases of content removal.Footnote ⁶³ Most notably for purposes of this chapter, however, online platforms that offer online commerce services must adhere to rules that permit identification and tracing of all merchants operating on their platforms.Footnote ⁶⁴ Very large online platforms additionally must conduct risk assessments relating to, among others, fundamental rights and physical or mental health and safety, and they must implement mitigation measures when risks are identified.Footnote ⁶⁵

Taken together, these requirements express considerable ambivalence about the significance and durability of dominant platform power. On the one hand, the “gatekeeper” terminology seems intended to signal a new regulatory paradigm with a particular focus on structural domination of multiple, interlocking activities. On the other, many of the specific restrictions gesture backward toward traditional conceptions of competition- and transparency-related governance. They attempt to achieve the sort of functional disaggregation that competition scholars have recommended as means for reintroducing and protecting competition in digital markets, along with the sorts of transparency and process that scholars who study consumer protection, data protection, and content governance have long urged as means of instilling greater market responsiveness to individual choices. Neither set of restrictions acknowledges the extent to which digital platforms now function as fundamentally nonmarket mechanisms for tuning access to information and allocating goods and services.Footnote ⁶⁶ Both regulations also signal tacit acceptance that certain digital platforms are dominant and likely to remain so. In particular, because some of the interoperability requirements for third-party content and apps may exacerbate data security risks to users, they may have the unintended effect of steering users toward content, apps, and other services offered or certified by gatekeepers and therefore perceived as safer.

Most notably for purposes of this chapter, the DMA and DSA do not fully engage the question of platforms’ infrastructural power. As I explain elsewhere, dominant platforms have attained infrastructural scale and transparency to critique by leveraging a particular modular structure for data collection and exchange. Formally separate elements of platformized communication systems that face different users, including the interfaces and displays provided to individual and business users, the SDKs provided to app developers and information partners, and the dashboards used to transact with advertisers, are interconnected at the infrastructural level, and the linkages between those elements give data flows in the platformized economy their distinctive attributes.Footnote ⁶⁷ Those flows, in turn, give rise to the enormous datasets now under the control of dominant platform providers of mapping and mobility services, and they generate the revenues to support the very large concentrations of processing power that dominant platform providers deploy.

The DMA and DSA permit most of these elements and the linkages between them to remain proprietary and undisclosed. Although dominant platforms must allow interconnection with their services, the DMA’s unbundling and interoperability mandates do not appear to encompass full access to all of the modular protocols for data collection and exchange that make up a given platform’s distributed “ecosystem,” nor do they require disclosures about the specific infrastructural arrangements through and across which data flows are directed. With the exception of a provision in the DSA requiring very large online platforms to create public searchable databases of all digital advertisements placed through the platform, neither the DMA nor the DSA mandates access to the data infrastructures amassed by dominant platforms or the processing power that dominant platforms control.Footnote ⁶⁸ And, although each regulation addresses certain types of data processing, neither purports to govern data collection at all. Questions about the legitimacy of data collection and about the legitimacy of mechanisms purportedly designed to ensure genuine consent are left for data protection regulators to consider.

These issues present complex tradeoffs. As scholars of both technology law and regulation and governance have long recognized, tailoring regulatory obligations to particular technological configurations invites simply designing around them. So, for example, functional disaggregation in the digital advertising system has impeded the search for a “controller” to whom legal obligations under the General Data Protection Regulation (GDPR) might attach.Footnote ⁶⁹ Relatedly, an extensive and growing literature casts doubt on the ability of the GDPR’s consent provisions to achieve their intended purpose, explaining why privacy-protective preferences do not translate into privacy-protective behaviors and why individualized, notice-and-consent-based mechanisms are ineffective tools for disciplining predesigned, networked assemblages for data-driven patterning that operate at scale.Footnote ⁷⁰ Simply mandating open access to dominant platform datasets, meanwhile, would exacerbate the problems the GDPR attempts to address and would risk naturalizing the predicate processes of data collection and processing as dominant platform providers have designed, implemented, and infrastructured them. Last but hardly least, it also would risk naturalizing and perpetuating the environmental and sustainability costs that those infrastructures create.

Here it is worth noting that data infrastructures for mapping, mobility, and commerce also implicate a very different set of regulatory traditions relating to public works and public utility infrastructures.Footnote ⁷¹ Such regimes vary a great deal – among other things, some such infrastructures (such as roads and water and sewer systems) traditionally have been publicly funded and operated, while others (such as energy systems) have relied on private provision – but all typically include public access requirements, nondiscrimination mandates, reasonable rate requirements, and subsidies for lower income users.Footnote ⁷² To be clear, legacy public utility regimes also do not speak to the full range of questions surrounding governance of infrastructural configuration. For example, so-called “smart” home systems that connect to electricity grids and municipal water systems raise difficult questions about privacy for which the public utility paradigm has no ready answers. And legacy public utility regimes have themselves been under sustained challenge by advocates of neoliberalized deregulation for many years. So, for example, even as systems for energy provision now confront pressing challenges related to sustainability, programs designed to govern from the edges by encouraging self-restraint in energy consumption and/or reorienting energy markets in ways that reward sustainability have achieved only limited successes. Those experiences counsel skepticism, if more were needed, toward efforts to govern data-driven infrastructures from the edges by (re)introducing more effective transparency and choice or more effective market-based competition. But older traditions of public utility regulation also contemplate more direct roles for regulators acting in the public interest.

At a more abstract level, regimes of public utility regulation are concerned with the regulation and governance of flows of human activity and exchange across scales. They are, therefore, suggestive of the needed path toward a public-regarding regime of digital infrastructure regulation.Footnote ⁷³ If, in the information era, digital platforms have come to function as uniquely powerful and (re)configurable mobility infrastructures – as indispensable mechanisms for both facilitating and governing the movement of people and goods – then the public’s interest should be understood as extending to all of the mechanisms through which the data-driven infrastructuring of mapping, mobility, and commerce is performed. In particular, it should be understood as encompassing both matters of configuration – that is, the modular protocols for data collection and exchange now so thoroughly embedded in our economic and social lives – and matters of knowledge production – that is, the large datasets that both reflect and enable mobilities and the scaled-up reconfigurations of social and economic activity that they produce.

2.6 Conclusion

Data infrastructures for mapping, mobility, and commerce are remaking economic, social, and even physical geographies. Relative to physical infrastructures, their capacity to scale vertically and laterally renders them uniquely nimble, flexible, and adaptable to new uses. By the same token, however, they also facilitate assertions of both private and state power in ways that physical infrastructures alone cannot. Current approaches to public governance of the data-driven operations conducted by giant digital platforms have, largely by default, privileged post hoc interventions that do not engage platform operations on the infrastructural level. Effective governance institutions require mechanisms for asserting public interests more directly within infrastructural logics and their corresponding sociotechnical formations.

3.1 Introduction

Much technological governance scholarship, including that which analyses digital borders, focuses on how digital technologies differentially constitute individuals as members of broader publics. For example, ethnographies evaluate how migrants are sensed and scrutinised by the tools regulating European Union borders and how this treatment is distinct from the experience of EU citizens.Footnote ¹ Social histories illustrate how data-centric tools create ‘the norm’ as a category of people and, in doing so, distinguish others as deviant, or apart, from that norm.Footnote ² While this work evaluates how subjects are individually treated, it also suggests how this treatment constructs subjects as broader publics and political communities.

The digital technologies regulating state borders construct another equally important, but less often explored, subject: the state agency. State agencies are crucial legal subjects. They create geopolitical borders by delineating spaces where sovereign power is exercised. They are also the targets of conventional accountability measures. In administrative law, for example, someone who is mistreated at the border might challenge that treatment in court and seek remedies against the agency or ministry responsible for that treatment. Yet, when digital tools contribute to the governance of borders, state agencies – and the agency or power of the state itself – are reconfigured. This reconfiguration complicates legal accountability mechanisms even as it concentrates the effects of state power on some (often already marginalised) people.

Where and how might we locate digital border agencies? Through which traces, sites, and actors (human and otherwise)?Footnote ³ This chapter takes up the challenge of answering these questions. It investigates how the digital border’s socio-legal-technical infrastructure enables new relations and reshapes institutional practices and dynamics. In doing so, it highlights points where distinctions between human, non-human, and institutional processes are redrawn to suggest how dispersals of authority and reconfigurations of agency through infrastructure play out. Throughout, it emphasises not just the social and technical but also the legal intricacies of border infrastructures, as legal technicalities crucially impact how borders operate.

Drawing on a recent Canadian example, this chapter argues that digital border infrastructures focus and diffuse state agency (or agencies) in ways that fundamentally challenge administrative law and the legal subjectivity of state agencies. In developing this argument, the chapter intentionally plays with the double meaning of the term ‘agency’. Agency evokes the ability of state actors to generate effects; it also encompasses institutions (ministries, departments, authorities) that are agencies in an administrative law sense. Administrative law is the chapter’s focus because, in Canada, it remains the tool that is most often used to hold border authorities accountable for their actions. Yet, as this chapter will suggest, contemporary administrative law may be ill-equipped for this task if we take the infra-legalities of digital borders seriously.Footnote ⁴

This chapter first explores state agencies as legal subjects, positioning them among other popular subjects in digital or dataist state scholarship.Footnote ⁵ It then critically assesses how state agency (i.e., the state’s capacity to effect action through its practices) contracts and expands, using a recent digital border conflict to illustrate the argument. Finally, the chapter urges scholars and advocates to take digital infrastructures and algorithmic violence seriously. To that end, it encourages the pursuit of robust horizontal accountability strategies that are as networked and diffuse as the digital governance phenomena that they seek to regulate.

The digital border is a useful example of the simultaneous focus and dispersal of state agency because it contains many features common to digitalised administrative operations. Like initiatives explored elsewhere in this volume, the digitalisation of border administration has been underway for at least two decades.Footnote ⁶ Digital border infrastructures combine older list-and-algorithm tools and newer anticipatory machine learning tools (or are piloting these tools for near-future use).Footnote ⁷ The Canadian example fits somewhere between old(er) and new(er) digital border infrastructures. Sophisticated machine learning may operate at some nodes in the infrastructural web, but many of the tools described below rely upon rudimentary risk models and decision tree architectures. Ultimately, when tracing how digital bordering infrastructure fosters new relations and shifts institutional practices, the presence or absence of a specific algorithmic tool may matter less than how multiple tools, together with officials, agencies, and others, interact to focus and disperse authority.

This chapter thus conceptualises digital bordering infrastructure expansively and focuses on infrastructural relations. The infrastructure described below includes ‘technical’ things (databases, software, and interoperability mechanisms, for instance), but extends beyond them to encompass practices (such as the practice of assessing risk), state officials (immigration and border officials, even judges), institutions (ministries, departments), and ‘non-state’ actors (such as airlines). This arrangement of tools, practices, institutions, and people often fades into the background of conventional legal analyses, which privilege some elements (state officials, for instance) while overlooking others. A conventional legal approach, in other words, simply cannot robustly envision or evaluate the simultaneous concentration and diffusion of state agency that is my concern here.

3.2 State Agencies as Legal Subjects

Literature on data, technology, and law tends to address a unique subject: the people who are the subjects of data collection and regulation. This focus may be linked to the recent introduction of the ‘data subject’, a legal category with roots in EU data protection law. This legal device encapsulates and protects the person whose data is captured and stored by a variety of technologies (websites, automated decision-making systems, etc.). While the data subject concept has become so pervasive that it is sometimes shorthand for any (natural) person who is the target of data collection practices,Footnote ⁸ some scholars question its universality. Fleur Johns, for instance, argues that data-driven tools shift subjectivities in humanitarianism initiatives in ways that are distinct from the European concept of a data subject. She identifies a new ‘tinkering’ subject generated through the novel sensorial capabilities of contemporary institutions like Pulse Lab Jakarta.Footnote ⁹ Others examine how being the subject of data collection and processing impacts subjectivity itself. To constrain these effects, some scholars propose that the ethical frameworks governing ‘human subjects’ research should extend to protect those impacted by data science initiatives.Footnote ¹⁰

At the same time, a growing body of scholarship across the humanities and social sciences analyses the links between subjects, objects, and power with (re)new(ed) vigour. New materialism literature, for example, sharpens how we might distinguish between (legal) subjects and objects.Footnote ¹¹ This work challenges conventional notions of who, or what, might be responsible for outcomes and who has (or ought to have) agential power. Critical new materialism scholarship, in particular, traces how subjects and objects are enmeshed. Its aims are both illustrative and political: the goal is to show how power functions in the world to make critical political economy analyses (and presumably their transformational outcomes) possible.Footnote ¹² Similar observations are foundational to the arguments of critical data scholars, which show how data is always constituted in relation to humans, institutions, and technological infrastructures.Footnote ¹³ Within socio-technical scholarship more broadly, interactionist and infrastructural scholars such as Lucy Suchman and Susan Leigh Star have studied how subjects and objects (from photocopiers to scientific categories) are enmeshed, raising questions about agential relationships and power.Footnote ¹⁴ Far from esoteric, such lines of inquiry are vital to reconceptualise how technologies govern, how objects act, and how legal accountability might operate effectively.Footnote ¹⁵

Traceable throughout these literatures is a distinct, often understated, subject: the state agency, an equally important subject for digital government and for legal accountability. In digital government initiatives, state agencies work closely with algorithmic tools. They may remain the face of the state even when decision-making authority becomes distributed across a network of actors. Yet, state agencies are less often directly scrutinised. Data subject literature largely assumes that state agencies exist and are (or ought to be) responsible, at least in part, for regulating how data subjects’ data is collected, stored, and used. In new materialism literature, meanwhile, the peculiarities of state agencies may recede behind a flattened topography and network of actors.Footnote ¹⁶ State agencies are vital actors in many digital government projects, however.Footnote ¹⁷ Their status as such is especially evident in border contexts, where administrative agencies plus other actors exercise sovereign power on behalf of states with often devastating consequences.

State agencies are crucial legal subjects. In administrative law, their subjectivity has two intertwined elements: first, state agencies subject others to the force of law, and, second, state agencies are the subjects of legal mechanisms to constrain the worst excesses of state power. The power or authority of state agencies is tempered by a mix of statutes, legal principles, and constitutional or human rights, each of which may impose conditions that state agencies must meet for their actions to be lawful.Footnote ¹⁸ State agencies remain the subjects of these accountability mechanisms even though their activities have evolved over decades of ‘digital-era government’ initiatives.Footnote ¹⁹ For example, anyone who contests a digital border decision through court proceedings must name the correct state agency (or ministry responsible for that agency) in their pleadings documents. This naming process is essential for vertical legal accountability. It ensures that any remedial orders issued by a court are legally enforceable against high-level state actors (government ministers, for example). It also assumes that these actors have the authority necessary to ensure that front-line officials within a particular administrative agency abide by the terms of a court order.Footnote ²⁰

The notion of an identifiable, accountable state agency is deeply embedded in domestic legal systems and in theories of public law. For this reason, legal scholars may assume that a (legally) responsible state agency exists and can be found somewhere within a digital border infrastructure. Recent literature on accountable algorithmic administration remains focused on vertical accountability mechanisms that require top-down implementation to ensure fair processes and appropriate outcomes.Footnote ²¹ Vertical accountability is also a foundational assumption within digital constitutionalism literature,Footnote ²² as well as in mainstream international law literature on algorithmic governance.Footnote ²³ When we examine digital border infrastructures more closely, however, the ‘state agency’ quickly becomes a problematic accountability device, not least because many actors contributing to digital border decisions fall outside of its purview.

When state administration is operationalised through socio-legal-technical infrastructures like those that constitute the digital border, a legal accountability paradox arises. This governing arrangement intensifies the visceral effects of state power, while simultaneously diffusing state agency among many agencies, making it harder to ‘name, blame, and claim’ them.Footnote ²⁴ People feel the undeniable sting of state authority facilitated through digital tools, but if they try to identify a responsible entity (in a legal sense), they often find themselves punching a shape-shifting cloud of wasps. Here, the swarm is not nested in the sovereign (the state agency responsible for the border decision).Footnote ²⁵ Rather, the swarm (officials, software, hardware, and other actors channelling state authority onto someone) momentarily holds the form of the sovereign before dispersing when anyone seeks to identify its components as legally accountable state agents.Footnote ²⁶ The ‘sharp edges’ of digital borders are created by no one in particular.Footnote ²⁷ Responsibility and agency are dispersed (or deferred) across a socio-legal-technical infrastructure.Footnote ²⁸

This paradox is likely unsurprising to socio-legal scholars, who have convincingly unsettled the assumption that state power is unidirectional or that conflicting modalities of governance cannot coexist.Footnote ²⁹ Nonetheless, their insights have yet to substantially influence legal scholarship more broadly. Consequently, calls for accountable digital government continue to emphasise traceable vertical delegation and responsibility, which assumes sequential decision-making, distinct human agency and discretion, and clearly identifiable agents or agencies.Footnote ³⁰ These proposals rely on a ‘liberal proceduralism’ that is drifting far from the distributed arrangements they seek to regulate.Footnote ³¹ Administrative law can evolve. The key challenge is how existing legal accountability frameworks designed to constrain administrative action might become responsive to the emergent realities of digital government infrastructures.

Inspired by Gavin Sullivan’s infra-legalities approach and its invitation to consider how socio-technical infrastructures reconfigure actors and reshape legal practices and power,Footnote ³² the next section illustrates how digital border infrastructures concentrate and scatter agency across the web of entities governing the interaction between a traveller seeking to visit Canada and state authorities. In the process, it emphasises the socio-legal-technical nature of digital border infrastructures.

3.3 The Concentration and Diffusion of State Agencies in Digital Border Encounters

Encounters involving pre-approved temporary resident visas demonstrate how state agency becomes simultaneously concentrated and diffused. This section first sets out important technical details about Canada’s digital border infrastructure, then describes the case of Smail Khaniche who was forcefully denied entry to Canada in 2018 despite his pre-approved visa.Footnote ³³ This experience is then critically analysed to show how digital bordering infrastructures enable new relations and reshape institutional practices. This critical assessment shows how digital border infrastructures intensify and disperse state power in ways that frustrate external review. In the process, responsibility and agency, and relationships between state agencies and other institutional entities, are also reconfigured.

3.4 Tackling Challenges for Legal Subjects and Legal Accountability

Building on the insights into how the digital border’s socio-legal-technical infrastructure operates, how should scholars and advocates respond? Specifically, how might we address the challenges that a network of agencies and actors with few responsible – or willingly responsible – principals raise for conventional accountability mechanisms? To return to the metaphor of the swarm nesting in the sovereign, what is a swarm without a ‘queen’? This concluding section proposes three main points of action. It urges scholars and advocates to take decision-making infrastructures seriously, to take algorithmic violence seriously, and to develop robust horizontal accountability mechanisms that better address both decision-making infrastructures and algorithmic violence.

First, decision-making infrastructures, including those governing borders, must be taken seriously. This proposal is deceptively simple. It challenges accountability strategies that centre on human action and those that focus on regulating a single node within a decision-making process. Human-centred and tech-centred approaches are standard in the growing literature on how to regulate algorithmic tools, but they underappreciate the interwoven nature of infrastructural relations. For example, a human-centred response to Mr Khaniche’s experience might propose that CBSA officers’ discretion ought to be better guided. It might also suggest that human officers ought to act as a constraining force ‘in’ or ‘on’ the loop of risk assessments and automated capabilities incorporated into GCMS.Footnote ⁸⁵ A tech-centred approach, meanwhile, might identify GCMS as the key ‘actant’ in the network that forcibly removed Mr Khaniche from Canada and recommend a wholesale re-evaluation of GCMS’s functionality to regulate data flows between GCMS and other systems.Footnote ⁸⁶ On their own, however, both approaches fail to address how humans, technologies, and agencies co-produce the conclusion that Mr Khaniche is inadmissible to Canada. The authority of state agencies, from the CBSA to the intelligence units constructing no-fly lists, and of other agencies whose actions support the outcome here (airlines, banks, tech firms) grounds the functionality of digital border infrastructure. Taking infrastructure seriously arguably requires us to analyse the proliferation of state agencies to ensure that state power remains governable particularly in dataist state initiatives.Footnote ⁸⁷

Taking infrastructure seriously also enables us to use infrastructural knowledge to identify new sites where strategic action might impact outcomes and to trace how such action affects an infrastructural web.Footnote ⁸⁸ This approach would require that we commit to long-term horizontal accountability strategies and to follow-up studies of their effects. Both features disrupt contemporary academic and regulatory conventions, which are governed by relatively tight three-to-four-year cycles based on the temporal constraints of external funding and election cycles. To address this challenge, we must creatively envision, scope, and support both research and regulation to achieve long-term goals despite the pressure for short-term deliverables. For instance, researchers and advocates alike might undertake projects to expansively trace infrastructural elements in a format that is easily accessible and intelligible. These initiatives could mirror ‘open government’ practices and could be regularly archived and updated. Government agencies relying upon digital infrastructures could revise their consultation practices to adopt ongoing, robust, agonistic processes that continue past the initial adoption of any given algorithmic tool.

Second, we must better recognise and respond to how data-driven infrastructures enable and shield algorithmic violence. While this term may lack a clear definition,Footnote ⁸⁹ what is clear is that socio-legal-technical infrastructures tend to target the same marginalised people and communities for unfair, even violent, treatment.Footnote ⁹⁰ For these reasons alone, we should heed Rocco Bellanova and others’ call to critically assess ‘how algorithmic systems feed (into) specific forms of violence, and how they justify violent actions or redefine which type of violence is considered legitimate’.Footnote ⁹¹ Doing so necessitates reflection on what counts as ‘violence’. The forced removal of Mr Khaniche fits, as does the officer’s attempts to coerce Mr Khaniche’s signature on the voluntary departure form. But Mr Khaniche also spent one month’s salary on his plane ticket believing his multiple-entry visa would ensure his admission, only to be denied entry to Canada. To what degree does algorithmic violence extend to family in Algeria, or his brother in Canada? How might the digital traces of ‘riskiness’ resulting from this border encounter affect his life or the movements of those ‘associated with’ him in the future?

Taking algorithmic violence seriously requires that we interrogate how infrastructural relations extend the state monopoly on violence to other actors, and with what effects. In addition to examining computational force (the force required to enable computation and the force leveraged by computation),Footnote ⁹² we must also investigate how infrastructural arrangements distribute opportunities for violence through computational and other digital border components. To what end are airlines, municipal police, financial institutions, and even voluntary departure forms sources of algorithmic violence? To what degree do (or ought) legal accountability mechanisms extend to these actors? Put differently, how does algorithmic violence become shielded from recognition and scrutiny when performed by an infrastructural network that seems infinitely expansive? How might we recognise and critically assess other components of a socio-legal-technical infrastructure, including the violence of infrastructure maintenance (i.e., biometric data collection, or forced access to Mr Khaniche’s phone).Footnote ⁹³

Finally, to take both infrastructure and algorithmic violence seriously, we must develop accountability strategies that are as horizontal and networked as the phenomena they seek to regulate. Public administration and digital governance scholars highlight the urgency of this task.Footnote ⁹⁴ It is arguably a necessary companion to infra-legalities research, as infra-legal studies suggest that only a more robust set of tools will effectively govern the web of actors that constitutes digital border infrastructures. Specifically, socio-technical solutions that ensure responsibility flows across infrastructural nodes must be paired with legal mechanisms. Administrative law updates, for instance, might allow multiple actors to be named in legal proceedings, expand available public law remedies, or even impose joint and several responsibility for implementing court orders. We might also imagine how legal standards themselves might be encoded into the networks and nodes described above to alter their affordances.

Horizontal accountability theory is one place to start. As a theoretical approach, horizontal or shared accountability has two main benefits. First, it encompasses the diverse actors (front-line officials, managers, ‘non-human’ software, databases, and state agencies) that co-create outcomes.Footnote ⁹⁵ Through its attention to these actors, this theory may suggest how responsibility can be shared between many agencies of the state. Second, as a theory of public administration, horizontal accountability generates potential regulatory responses that extend beyond law and policy amendments or technological tweaks. For example, it may reveal other responsibility pathways and accountability mechanisms that would be highly effective components of a larger toolkit to govern state agency and agencies.Footnote ⁹⁶

I am not an expert on either digital flows or infrastructures: I am a social-political theorist, turned socio-legal theorist, who like many of us is attempting to deal, intellectually and politically, with the information practices of the present. In my recent work I have paid particular attention to information practices and representational strategies in the context of major infrastructure projects, but without specific attention to software or platform mediated practices.

Therefore in this comment I will address, somewhat tangentially, some of the points made in the two initial chapters of this book by Julie Cohen and Jen Raso; but I will mainly focus on questions that pertain to the larger enterprise, that is the research project that has given rise to this book.

Scanning much of the literature on ‘big data’, as well as the almost completely separate literature on infrastructures,Footnote ¹ it has long struck me that many key terms are taken for granted even in critical analyses. Thus, while in my youth I studied what was called ‘continental’ philosophy,Footnote ² not analytic (English) philosophy, I believe that the kinds of questions most often asked by analytic philosophers (such as: ‘what do you mean by ‘data’ anyway?’) can be very useful to clear the ground to prepare it for discussions such as the ones represented in this volume. This comment is thus meant as a contribution to the collective discussion rather than as the kind of ‘comment’ on individual papers usually presented in academic conference panels.

In highlighting what is unique about our present, from a progressive perspective – which in my view means that we avoid presupposing that capitalist market mechanisms are usually better than alternatives – most people turn to words such as ‘the information society’ or ‘big data-society’. These terms have given rise to derivative terms – perhaps most influentially, the term ‘data-driven’, a term used as it usually is in Julie Cohen’s contribution. I find the very term ‘data-driven’ problematic, in that the recourse to data, big or otherwise, is mainly a means to an end – not the kind of Aristotelian end that drives processes in a teleological manner. Data infrastructures can be means to a public-purpose end, but the same tools can also serve the goal of increasing market share and/or profits for a capitalist corporation. Of course it is necessary for public institutions to pay attention to the often unintended effects of ‘datafication’. This scrutiny is not easy since digital tools most often come as a blackboxed package (as when Canadian universities, in the early twenty-first century, chose either Microsoft or Google for software services that become compulsory for students and staff, with little scrutiny of the effects of these choices).

Perhaps because I am a newcomer to the field of ‘big data’ studies, I hold that people and institutions are driven by certain pre-digital, old-fashioned motivations: to increase profits and/or market share, in the case of for-profit corporations, to promote democracy and accessibility in the case of NGOs, or to increase or modify social control in the case of authoritarian states.

This raises a broader point about the way in which we study data flows. While as a Foucaultian of sorts I do pay particular attention to ‘how’ questions, which in scholarly practice means mainly analysing the effects of particular techniques of governance,Footnote ³ I have come to believe that it may be risky, both politically and intellectually, to do research as if the forms of power and the character of the networks made possible by modern computing are wholly new. I hold that while we may live in data-saturated fields of governance, the moves made by authorities and would-be authorities that implement digital systems are not necessarily wholly new. There may be more continuity between the pre-digital past and the present than is often assumed by writers who focus on the innovative features of the ‘information’ or ‘big data’ society.

Jen Raso’s brilliant chapter is a case in point, in my view. The dialectical dynamic she attributes to digital processes utilised in many state agencies (a dynamic that features simultaneous processes of centralisation and dispersal) is, I would argue, far older than the modern computer. My research on the history of special-purpose districts and of arms-length semi-public agencies governed not through ministerial accountability but by means of relatively autonomous corporate-like boards has found that the dispersal of state ‘agency’ through special-purpose bodies and what British administrative lawyers call ‘the boardisation of the state’ coexists with a kind of rebirth of sovereign power: the arms-length agencies, Quangos and so on, operate as if they were private for-profit corporations, subjecting people and entities to top-down undemocratic rule.Footnote ⁴ Under the populist banner of ‘depolitisation’, starting well over a hundred years ago, many public and quasi-public agencies have been turned into what one might call fake private corporations. These have minimal transparency requirements and even less by way of accountability mechanisms. This whole complex ‘ecosystem’ of governance tends to be neglected by political science, and its unique features are not necessarily captured by administrative law’s category of ‘agencies, boards and commissions’.

A significant literature relied upon in this volume concerns the processes by which corporations (especially oligopolistic ones based in either Silicon Valley or, as Julie Cohen helpfully reminds us, China) appropriate the ‘data exhaust’ emitted by us and our ‘devices’ as we work, seek health care, shop for goods, engage in leisure activities, educate our children or travel. That the process by which private corporations (be they for-profit US corporations or corporations controlled by the Chinese party-state) appropriate the vast reams of what we have learned to call ‘data’ generated both by individuals and by infrastructures in use as well as entities such as Jen Raso’s ‘state agencies’ has come to be known as ‘data extractivism’ – through a mostly implicit parallel with the better known older practice of the state licensing for-profit companies so they can legally appropriate arguably common natural resources such as timber, water, and minerals. Data extractivism is a useful term, in highlighting the monetary value of data that has no value for the individual (I cannot bring to any market I know the location data generated by my phone) but has much value for platform businesses such as Google, Airbnb, Amazon and so on, and also for commercial data brokers.

Data extractivism is a very useful term in discussions about the nature and the likely future of what Shoshana Zuboff has influentially called ‘surveillance capitalism’, in part because it underlines the commonalities between the management of ‘physical’ resources whose enclosure and appropriation has long been studied and today’s digital or electronic resources. Fortunately, this book focuses not so much on the kind of data extractivism that privacy activists fight against, but rather on broader patterns of data extractivism through which both states and corporations build networks based on the extraction of intangible resources (including but not limited to datasets, or aggregate data not yet made into a dataset).

Personal privacy gets much public attention (e.g. the European Union’s ‘right to be forgotten’); but as many have pointed out, that battle was arguably lost when Google was founded, and in any case impersonal aggregate data has been found to be an excellent resource for many corporations. The ubiquitous use of assemblages of digital information, software and hardware such as the microchips embedded in ‘ordinary’ objects can be said to have infrastructural effects, as Julie Cohen points out in her chapter. This is very true, but what does that claim mean, exactly? It seems to me that we collectively need to have a little bit of clarity about what the highly mobile and currently ambiguous term ‘infrastructure’ does. There is much talk in the social sciences (not so much in law) about the ‘infrastructural turn’, but little clarity about what that might mean beyond trying to redefine and resignify many entities and processes under the banner of ‘infrastructure’.

The two key terms that need reflection and commentary, for this volume as a whole, are ‘data’ and ‘infrastructure’. I will begin with the less familiar (in the law and technology field) term of ‘infrastructure’, since I have carried out empirical research on the governance of infrastructure projects.Footnote ⁵ I will then jump into the pool – or swamp? – of digital-world discussions by asking a few questions about that largely taken for granted term, ‘data’.

What Do We Mean by ‘Infrastructure’ and ‘Infrastructural Effects’?

Infrastructure is traditionally defined as networks that provide public services, not necessarily through state ownership and control but to meet public as distinct from individual needs: communication, transportation, the energy to power things and activities, the management of both waste water and drinking water and so on. The term has no fixed boundaries, since in the global South or majority world what might look like ‘jerry-rigging’ might work reasonably well as an infrastructure.Footnote ⁶

But in this day and age, infrastructure is seen as more than physical. That is not a result, it needs to be highlighted, of the intangible character of digital data. Well before the internet, telephone communication was enabled by cables and phones but the importance of the content of the communications could not be reduced to gadgets. And conversely, despite the tendency in legal circles for digital data to be regarded as purely ‘intangible’ (perhaps because of the prominence and dissemination of intellectual property (IP) terms of art), the datasets cannot exist in thin air: they too require cables (the submarine cables that make the internet global being often literally attached to existing phone cables) and of course physical computers, many of them quietly gathered by big corporations in anonymous-looking hangar-like server farms.Footnote ⁷ That is: what are often dismissively called ‘legacy’ infrastructures include both physical and intangible components, and so too does the world of digital data. Legal scholars tend to underestimate the materiality of law,Footnote ⁸ as many critical socio-legal scholars have noted, so it is not surprising that they have also tended to underestimate the materiality of digital data – especially since IP law insists on classifying property as either ‘tangible’ or ‘intangible’, an oversimplification of complex realities that is parallel to and perhaps based on the assumed binary distinction between ‘real’ and ‘moveable’ property.

Like Foucault pondering what came before ‘the deployment of sexuality’, if we want to understand the proliferation of the term ‘infrastructure’ we need to consider what it replaced, what came before. The answer is easy: in countries around the world ministries of infrastructure replaced what were until recently ministries of public works. Municipalities tend to still have a ‘public works department’; but at higher levels of government what we have are infrastructure ministries and, even more important, infrastructure ‘agencies’ (to re-use Jen Raso’s useful term, recovered from the boring public administration literature on bureaucracies). Why the change?

I argue that the varied phenomena that huddle under the ambiguous megaterm ‘neoliberalism’ have combined to create a devil-like negative halo around public expenditures – and hence public works. But of course roads need to be built and maintained today as in the past, as do water systems, electricity grids, railway networks and so on. So what was the solution to this problem of justifying expenditures on public works within neoliberal regimes? The solution that was palatable to neoliberal politicians was a semantic one: to rename the public expenditures previously listed under ministries of public works as ‘investments’. The operative phrase became ‘investing in infrastructure’. Try googling that phrase for the jurisdiction you live in and you’ll be persuaded by my argument.

Just as public expenditures could be saved for neoliberal purposes by being renamed ‘investments’,Footnote ⁹ so too the term infrastructure, which as many scholars have pointed out does not have a single clear meaning, was highly useful. Even the most right-wing tax-cutting politicians can boast about ‘investing in infrastructure’ – and they do, regularly. Even Boris Johnson, in his day, attempted to show that he too was ‘investing in infrastructure’ by financing with huge amounts of public money the ill-fated British high-speed rail project known as HS2. In the meantime, social democrats, and even the Chinese communist party, are also incentivised to adopt the new language of investing in infrastructure to justify pursuing New Deal-style dreams of major public works. And, relevant to the expansiveness of the term ‘infrastructure’ in today’s world, social democrats usually add that by infrastructure they don’t just mean roads and bridges and pipelines; they are also in favour of investing in ‘social infrastructure’ (formerly known as the welfare state).

It seems that virtually everything can be refigured as infrastructure in the present day. If Raymond Williams were to rise from his grave to analyse some ‘keywords’ of our own present, he’d have to include ‘infrastructure’. Re-naming welfare state projects such as homes for the aged or daycares for toddlers as ‘social infrastructure’ serves to justify spending public money on them at a time when ‘public expenditure’ is a no-no word.

The American pragmatists turn out to have been right: ‘infrastructure’ is whatever the term does in practice.Footnote ¹⁰ One can even redefine parts of the environment as infrastructure, if one stresses the public or social value of entities such as lakes, forests and beaches, whether or not those are in fact publicly owned. As has been pointed out by many infrastructure experts, the question of who owns a resource is not always determinative. A railway can serve the public at large, as a legally enforceable obligation, even if it is privately owned (as in the US legal doctrine of the responsibilities of ‘common carriers’).

Brett Frischmann is a noted US legal scholar who has argued that the term ‘infrastructure’ has an inherently public or at least shared logic. Like most American legal scholars he is keener on empowering civil society than in expanding the reach of the state; but in arguing that infrastructural goods are those that are not ‘enclosed’ but rather can be used by an indefinite number of people and groups for diverse purposes, he seems to line up, at least in some respects, with the social democratic notion first mooted (in American legal context) by noted property scholar Carol Rose. In 1986 Carol Rose published an influential article, in the Chicago Law Review of all places, on ‘the comedy of the commons’ – meaning the opposite of the better known ‘tragedy of the commons’.Footnote ¹¹ In that article she presented the idea that there are some entities that are ‘inherently public’ property. That claim has certainly been challenged, especially by those that would make a distinction between ‘state owned and controlled’ on the one hand and ‘public’ on the other. But for present purposes we need not take a stand on the difficult question of whether state property is or is not coterminous with public property – or whether the category of ‘the commons’ is a useful one (‘the digital commons’ being a term of highly uncertain content).

Here let me briefly try to address Jen Raso’s and Julie Cohen’s chapters together by focusing on Raso’s claim that datafication (or maybe just some forms of software use?) is a process studied by critical scholars only as it negatively impacts individuals (through breaches of privacy) or on groups (through the discrimination embedded in what AI people call ‘training data’).Footnote ¹² Raso shows, through secondary literatures and through the detailed study of the case of the hapless Algerian denied entry into Canada to visit his brother, that the systematic and use of networked data has a constitutive effect on ‘the state’. Instead of adding to the by now stale discussions about the character and extent of ‘the state’, however, Raso usefully highlights the dual role of state agencies (known in other literatures as bureaucracies, a term that is also intrinsically ambiguous).

On the one hand, sometimes collaborating and sometimes competing state agencies utilise digital datasets not only to do their own work but to communicate with other agencies, across ministries and even across national boundaries. Thus, on the other hand, the supposed unified agenda and ‘agency’ of the state becomes fragmented and dispersed. This dispersal means that legal mechanisms designed to address complaints and provide institutional accountability become moot or useless: it is very difficult, even impossible, for citizens to clearly identify which agency treated them inappropriately, or for journalists and scholars to find a single culprit to denounce.

Raso’s focus on specific state agencies is a breath of fresh air, after decades of largely unproductive theoretical debates about ‘the state’. (These debates, younger scholars may not know, petered out around the fall of the Berlin wall, as progressives rushed to embrace ‘civil society’ and sidelined the inconclusive intra-Marxist debates about ‘the state’.)

However, one point I would make, not so much by way of critique as by way of suggestions for further research, is that the strange dialectic whereby state power seems to concentrate and disperse at the same time is hardly new. The use of interconnected datasets – adopting here Rob Kitchin’s definition of ‘big data’, which stresses mobility and connectivity of datasets rather than simply sizeFootnote ¹³ – can certainly facilitate the dispersal of accountability across agencies, while simultaneously bringing back the oldest mechanisms of sovereign state power. But this dialectic of centralization and dispersion is visible in pre-internet, pre-digital state systems, if one sets aside the Marxist spectacles through which ‘the state’ was presumed to be unified by its being based on the interests of the ruling class.

Administrative law’s imaginary features particular state agencies acting autonomously, which as Raso points out can be problematic for mistreated citizens who may be the victims of assemblages of data exchanges and data flows. But state agencies have for a long time now had both the ability and the incentive to hive off subsidiaries and otherwise foster wholly new entities. Agencies as well as legislatures and ministries have long engaged in the co-creation, either with other parts of government or with the private sector, of new agencies – especially ‘arms-length’ special-purpose agencies.

Special-purpose agencies proliferated in the US from the 1920s onward, with the Port Authority of New York regarded by US historians as the ‘mother’ of all such ‘special-purpose’ or ‘special district’ governments. Sometimes the state (as distinct from the federal government) created agencies, such as urban development corporations, that used the vagaries of jurisdiction to muddy the waters of accountability, and often also of mere transparency. This development, which greatly undermined accountability in local governance, also took place in the UK, though flourishing only under the Thatcher regime.

When in 2017–19 I closely studied the failed ‘smart city’ project proposed by Google for a waterfront parcel of land in Toronto (together with my colleague and administrative law guru Alex Flynn), we found that the fact that the governance of most of the waterfront had long been downloaded by the three levels of government to the non-transparent agency Waterfront TorontoFootnote ¹⁴ really mattered. In that context I saw that ‘the state’ is not only fragmented through the workings of federalism and unclear jurisdictional dynamics but also has the opportunity and even the incentive of abdicating jurisdiction altogether, letting somewhat randomly chosen board members make decisions affecting a very large public.

Hence, it would be useful if scholarship were to investigate other periods and other fields of state action to see if the dynamic that Jen Raso has documented is visible elsewhere and in previous eras. In any case, it is healthy to have a professor of administrative law remind law and technology scholars that there is no such entity as ‘the state’, and that the proliferation of agencies is not just Weberian bureaucratic expansion gone mad but may act – very conveniently – to shield ‘the state’ from lawsuits enabled by the parallel proliferation of ombuds-type bodies and constitutional rights jurisprudence in the late twentieth century.

What Are Digital Data, Exactly?

I would first of all point out that just as the internet does not actually reside in the clouds, so too digital ‘goods’ are not ghosts hovering above the ‘real’ material world. Their financial and/or their use value might be mainly intangible; but they are embedded, inextricably, in vast material ‘supply chains’ that among other things use rare-earth minerals to produce smartphones and other emblems of digitality. These chains or networks may be primarily valued because of the digital data and digital predictions they enable; but they have to use electricity to run and to cool the humming computers in the server farms of Oregon and Washington State, where climes cooler than California’s has meant a reduction in air conditioning expenses for the Silicon Valley giants that quietly buy land and build anonymous-looking server farms.Footnote ¹⁵

There’s a subsequent question often bracketed by law and technology scholarship: the question of what counts as ‘data’. Many media studies and information scholars have argued that information can come in any format, from origin stories to oral histories to census returns: by contrast ‘data’ is a term best reserved for information that is suitable for quantitative and statistical manipulation and is collected and preserved digitally, in databases.Footnote ¹⁶

And yet the emphasis on numbers may be somewhat misplaced. When my children were in elementary school they learned the rudiments of what was called, in their school, ‘data science’ (not math). A typical exercise had them asking classmates whether they preferred apples, oranges, or bananas, then producing simple bar graphs representing the aggregate information. This was a pen-and-paper exercise. Now, were their simple graphs examples of ‘data’? Perhaps. But perhaps their graphs were not ‘data’, but simply information.

It does seem, however, that digital data, or perhaps better, datasets, are not merely internet-enabled versions of older practices of numeration (such as the mortality statistics collected by life insurance companies long before state pensions). As IP lawyers remind us, ‘facts’ are in the common law in the public domain. Facts (say, the French Revolution started in 1789) can of course be worked up into a dataset; and IP lawyers could then argue that the dataset is private property, since creating it involved some human creativity, not just facts in the public domain. What can be privatised is largely a political question (as seen in the historic case of the massive British ‘enclosures’ of land previously classified as ‘commons’).

I suspect that it will be more useful for the kind of work represented in this volume to not worry about ‘facts’ versus ‘data’ but rather to distinguish ‘data’ in general (here including my children’s hand-drawn bar graphs) from ‘big data’ – especially if we adopt Rob Kitchin’s view that it is mobility and networking that identifies ‘big’ data. He gives the example of census returns as an example of a collection of data that is large in size but is not ‘big data’, since in many countries census data are by law kept separate and immobile. The immigration and border-control agencies in Jen Raso’s case study could not have made use of census returns, for legal reasons.

While digital data is not ontologically unique, since contrary to the imaginary of IP law data always has a physical location or substrate, one curious feature of the digital world is that datasets do not exist on their own. Datasets are produced by some combination of humans and machines, and are managed, shared, and made to interact with other datasets in different ways. As is well known, the noted corporate success of the Alphabet/Google companies is largely due to that corporation’s capacity to connect different datasets.

If we want to appreciate the mobility and interactivity of datasets it is useful to highlight those corporations known as ‘platforms’. Platforms (Uber, Airbnb, Amazon) have infrastructural effects, Julie Cohen argues, very cogently. This means that they have effects on the physical world – as when road traffic changes as drivers frantically consult Waze to avoid traffic jams.Footnote ¹⁷

That platforms are unique in their infrastructural effects is a plausible claim, though of course empirical studies are needed to confirm this insight. The infrastructural effects of Airbnb have been documented by many entities and municipalities (mainly under the rubric of the effect, which cannot be called unintended, of diminishing the local stock of long-term rental housing). Similarly, the Uber corporation has caused tens of thousands of additional cars to prowl city streets – and yet no regulatory system that I know includes congestion charges as part of the Uber licensing system.

The infrastructural effects of platform capitalism have not received as much attention as the actual or potential negative effects on privacy and the effects on existing economic relations (such as the fostering of a precarious labour force known as ‘gig workers’.) But the vast literature on ‘smart cities’ contains many examples showing the effects on both municipal and privately owned infrastructure not so much of digital data but of how those data get used and circulate – effects that are not separate from the legal weakness of local governments in much of the common-law world.Footnote ¹⁸ It may be that data gathering is not the real issue, contrary to what civil-liberties organizations believe. It could be that how datasets are accessed, re-used and recycled across institutions and jurisdictions is the real problem, not so much data collection in itself.

A related point Julie Cohen makes is that just as the switch from older methods of governance and information gathering to digital data has infrastructural effects, there is also a parallel ‘datafication’ of infrastructure. The shared bicycles that the city of Toronto owns and that I use on occasion, for instance, are ‘datafied’ to the fullest extent: the app for the system shows me, as soon as I open it on my phone, a real-time map of the immediate neighbourhood that shows how many bikes are available in each of bike station and how many spots are open, if I want to leave a bike rather than take one. The bikes themselves likely have a location-data exhaust, presumably used to prevent theft and enable city workers to find ‘lost’ bikes. The bikes must also have a digital clock, since the first thirty minutes are free to anyone who like me has a yearly subscription.

I imagine every contributor to this book would agree that even if we stop ignoring the materiality of the so-called digital world, it is not a good idea to simply transpose categories and governance regimes developed for such infrastructural networks as roads and bridges to the realm of largely intangible digital goods and resources. Barbara Prainsack, a health data expert, suggested a few years ago that well-meaning arguments to the effect that many datasets should be considered legally as a ‘digital commons’ will founder if one simply uses existing categories of property and governance (state property; individual property; common property) and tries to apply these to digital datasets.Footnote ¹⁹

One reason mentioned by Prainsack for the care needed to develop governance regimes for digital data is that the digital world includes ‘open access’ resources as well as resources (such as commercial secrets) that are not shared, including many that might be ‘common’ but not common to all. Open access texts are qualitatively different from physical commodities in that they can be consumed in a ‘non-rivalrous’ manner, as economists say: many people can simultaneously read an online post, unlike conventional goods, which are subject to the zero-sum game of individual tangible property.

The mention of ‘open access’, a property regime increasingly relevant in the world of scholarly publications, may suggest to some that the jurisprudence and the political theory that has developed in regard to the physical (especially the environmental) ‘commons’ might be useful in elaborating policies for the extraction and the sharing of digital data. However, Prainsack argues forcefully that the physical commons cannot be transposed directly into the digital world. In her view, the much-touted idea of a ‘digital commons’ is not inherently impossible but would require much thought and discussion.

I raise Prainsack’s point about the commons vs. open access for the scholars included in this anthology – and the readers – to consider. I would put the issue thus: given that digital data can be shared by many, even owned by many simultaneously, since contrary to Walter Benjamin in the online world the copies are the same as the ‘original’, scholars and public officials need to think about how the intangible components of digital data can be governed in such as way as to empower civil society in a progressive, equitable manner (instead of only empowering groups of people who all have a subscription to an online service or own goods with a shared element). The term ‘the commons’ is often thrown around by legal scholars with an anti-private enterprise bent, with little attention to the distinctions between state ownership, public ownership not controlled by the state, ‘commons’ and ‘open access’.

To conclude: the two fascinating initial chapters provide an excellent introduction to many of the important issues surrounding ‘data infrastructures’, as well as raising important points about the infrastructural effects of some methods of data collection and use. If this comment serves to further the exciting collective discussion enacted in this book, it will have served its purpose.