Sam Lucente, Diagram of a Logic Chip, 1986.
Since the proliferation of the World Wide Web in the 1990s, critics of widely used internet communications services have warned of the misuse of personal data. Alongside familiar concerns regarding user privacy and state surveillance, a now-decades-long thread connects a group of theorists who view data—and in particular data about people—as central to what they have termed informational capitalism.1 Critics locate in datafication—the transformation of information into commodity—a particular economic process of value creation that demarcates informational capitalism from its predecessors. Whether these critics take “information” or “capitalism” as the modifier warranting primary concern, datafication, in their analysis, serves a dual role: both a process of production and a form of injustice.
In arguments levied against informational capitalism, the creation, collection, and use of data feature prominently as an unjust way to order productive activity. For instance, in her 2019 blockbuster The Age of Surveillance Capitalism, Shoshanna Zuboff likens our inner lives to a pre-Colonial continent, invaded and strip-mined of data by technology companies seeking profits.2 Elsewhere, Jathan Sadowski identifies data as a distinct form of capital, and accordingly links the imperative to collect data to the perpetual cycle of capital accumulation.3 Julie Cohen, in the Polanyian tradition, traces the “quasi-ownership through enclosure” of data and identifies the processing of personal information in “data refineries” as a fourth factor of production under informational capitalism.4
Critiques breed proposals for reform. Thus, data governance emerges as key terrain on which to discipline firms engaged in datafication and to respond to the injustices of informational capitalism. Scholars, activists, technologists and even presidential candidates have all proposed data governance reforms to address the social ills generated by the technology industry.
These reforms generally come in two varieties. Propertarian reforms diagnose the source of datafication’s injustice in the absence of formal property (or alternatively, labor) rights regulating the process of production. In 2016, inventor of the world wide web Sir Tim Berners-Lee founded Solid, a web decentralization platform, out of his concern over how data extraction fuels the growing power imbalance of the web which, he notes, “has evolved into an engine of inequity and division; swayed by powerful forces who use it for their own agendas.” In response, Solid “aims to radically change the way Web applications work today, resulting in true data ownership as well as improved privacy.” Solid is one popular project within the blockchain community’s #ownyourdata movement; another is Radical Markets, a suite of proposals from Glen Weyl (an economist and researcher at Microsoft) that includes developing a labor market for data. Like Solid, Weyl’s project is in part a response to inequality: it aims to disrupt the digital economy’s “technofeudalism,” where the unremunerated fruits of data laborers’ toil help drive the inequality of the technology economy writ large.5 Progressive politicians from Andrew Yang to Alexandria Ocasio-Cortez have similarly advanced proposals to reform the information economy, proposing variations on the theme of user-ownership over their personal data.
The second type of reforms, which I call dignitarian, take a further step beyond asserting rights to data-as-property, and resist data’s commodification altogether, drawing on a framework of civil and human rights to advocate for increased protections. Proposed reforms along these lines grant individuals meaningful capacity to say no to forms of data collection they disagree with, to determine the fate of data collected about them, and to grant them rights against data about them being used in ways that violate their interests.
Both the propertarian and dignitarian conceptions of data governance improve upon the status quo of data governance. But each is out of step with how data production, as a population-wide practice, results in both benefits and risks for individuals, and their approach—grounded in legal rights to individual freedom—consequently fails to resolve the most urgent questions around data governance. Propertarians rightly identify that data is a social relation—one that is deeply unequal, and thus one that must be reformed. But in an effort to address that inequality, they end up casting data relations as relations of private property or wages that are themselves exploitative. Dignitarians rightly diagnose current forms of data production as extractive. But in an effort to address this extraction, end up foreclosing the forms of data production that may be vital to achieving robust social welfare provision, while failing to consider the political economic conditions required for individuals to meaningfully exercise the freedoms they do grant.
Rather than proposing individual rights of payment or exit, data governance should be envisioned as a project of collective democratic obligation that seeks to secure those of representation instead.
Treating data as property or labor
Propertarian entitlement may occur in one of two ways. Most reforms propose a property right over data about the subject, in which the data subject may sell usage or full ownership rights. Alternatively, data may be conceived of as a form of the subject’s labor that entitles the data subject to command a wage in a data-labor market. The general notion behind such proposals is that formalizing the market for data about subjects provides a solution to the problems of data extraction. This solution transforms data about the subject into an asset that generates wealth for the subject.
Legally speaking, this occurs by “capitalizing” a data-asset—coding it via law with protections and features amenable to wealth creation.6 Propertarian data reform posits a solution to the problems of data extraction by fitting data into the existing concepts and legal apparatuses we have around "property" or "work." The governance of data then becomes the governance (via contract law, property law, and/or employment law and labor law) of property relations or labor relations.
Propertarian reforms respond to the quasi-enclosure and de facto ownership of data resources by technology companies. By formalizing the current informal propertarian status of data, such reforms directly counteract the collection of data by large firms like Google, Facebook, and Amazon, and formally invalidate the current practice of capturing value from subjects without compensation. Propertarian reforms advance a claim of enhanced individual control or self-determination in the data economy, as well as a redistributive claim: that data subjects are owed some share of the wealth that data about them helps to create.
There are several reasons to be skeptical of propertarian data reforms and how they conceptualize the problem of data extraction. The first is a pragmatic one: operationalizing the kind of complex and comprehensive micro-payments system suggested by these proposals at the scale they require may simply not be feasible or cost-effective.7
Second, it is empirically unclear if propertarian solutions can lead to material redistribution or increased control for data subjects. The mere granting of a labor or property right does not guarantee that the conditions underlying the sale of that labor/property will be non-extractive and uncoerced. And the conditions of the current data market are not conducive to small producers or laborers securing good bargains and exerting market power: while large data collectors are highly concentrated, data subjects are hundreds of millions of individuals around the world. Moreover, personal data from any one data subject is essentially value-less by itself, reducing the capacity for individual data subjects to meaningfully exert bargaining power. Under current conditions of collection, data subjects may not be able to overcome individual powerlessness to organize and exert collective bargaining power of the kind that necessitated the creation of traditional labor and employment law. For instance, datafication does not result in the kinds of visceral oppression that spurs organizing and builds counter-power. In contrast with domination in the workplace, data extraction is designed to occur as seamlessly and painlessly as possible, transmitting flows of data in parallel with data subjects living their online and offline lives. As a result, data subjects may lack the conditions necessary to identify as a common social group, uniting to demand better conditions of exchange.
Third, propertarian data reforms may exacerbate certain civil rights and privacy violations that occur in the digital economy. Consider the recent example of ICE purchasing access to a database that tracks the location of millions of phones—information collected via weather and game apps—to track and arrest suspected undocumented immigrants.8 Paying data subjects at the point of collection does nothing to address uses of such data that may violate the civil rights of others and amplify existing social oppression. In fact, by legitimating the marketplace for data, payment may legitimate downstream practices that result.9 Payment may also incentivize people to share data about themselves and thus further degrade privacy. Because those least able to forego income face added pressure to sell their data, propertarian approaches risk transforming privacy into an even greater privilege than it is today. A data subject may decide the risk of privacy loss is worth the payment provided, but privacy risk is notoriously easy to under-value at the point of exchange. Privacy risk associated with data isn’t static, nor is privacy loss linear—it accumulates and grows over time based on composition effects from multiple sources, varied downstream uses, and new applications. (People who tagged online photos of themselves and their friends in 2009, for example, could not have anticipated that companies contracting with law enforcement in 2018 would use such information for facial recognition.) Finally, data exchanges generate considerable privacy externalities: information about one person may well be used to make inferences applied to another.
Propertarian reforms concede existing processes of data commodification and mass data extraction in the digital economy. The horse having fled the barn, what data subjects can and should secure is a piece of the action.
Dignitarian critiques and egalitarian shortcomings
Where propertarian reforms conceive of data as the subject of individual ownership, dignitarian data governance conceives of data as an expression (or extension) of individual selfhood. Such accounts advance legal rights over personal data akin to natural rights: they obtain in the data subject as an extension of their right to self-determination. The most prominent dignitarian data regime is part of the European data governance regime, which alongside other affirmative data processing obligations, affords universal and inalienable rights over personal information and enshrines data protection and privacy as fundamental rights.10 Other dignitarians argue for extending the human rights framework to data governance, or champion a strengthened consumer-protection regime based in a suite of inalienable consumer rights.
Many dignitarian reformers claim that data extraction involves not only individual stakes, but also societal ones. Zuboff says the world’s digital information is a public good; the EU Data Protection Supervisor notes that privacy is not “only an individual right but a social value.”11 Yet in practice, the legal solutions advanced under dignitarian conceptions of data governance still primarily subject data to individual ordering. Dignitarian reforms typically grant individuals a set of inalienable rights over their personal information—freedoms secured for data subjects against certain future uses (e.g., use without consent, use that goes beyond the purposes originally given, or use once consent has been withdrawn), and that obtain with respect to data collected about them.12 These negative freedoms secure personal data’s quasi-personhood status in law governed by civil, not contractual, rights.
Dignitarians argue that by enshrining the alienability of data and regulating personal data production via markets, propertarian data reforms further incentivize unjust self-commodification. This is a form of personal violation and a kind of “legibility harm”—a degree of representation that violates the sanctity of our inner lives. These critiques effectively expose normative concerns one may have over the propertarian embrace of datafication, but they leave other aspects of propertarian reforms unaddressed.
First, dignitarian critiques do not account for the enduring egalitarian appeal of propertarian reforms. By promising data subjects payment, propertarian reforms articulate a redistributive agenda for the political economy of data. Meager though it might be, such an agenda promises material redistribution for data subjects that dignitarian reforms, on their own normative commitments, cannot provide.13 The redistributive agenda of propertarian reforms respond to claims of injustice based on the inegalitarian social effects of data extraction, not only the personal violation datafication entails.
Second, dignitarian critiques may chafe against positive accounts of data infrastructures that require large-scale mandatory data collection. Like private technology companies now, systems for social welfare provision—be it healthcare, housing, or a basic income—would almost certainly require rendering individuals legible, in some instances against their will. But they would do so in service of vital democratic welfare state endeavors rather than private gain.
Dignitarian critiques thus suffer from being simultaneously overly narrow and overly broad. By focusing on datafication’s violation of self, they do not address the economic imperatives that drive such harm nor do they provide a compelling agenda for addressing inequality in the data political economy. At the same time, the focus on datafication paints over meaningful distinctions regarding the purposes of data production and the conditions under which such purposes are determined.
Data egalitarianism: democratic alternatives
What these shortcomings suggest is that alternative conceptions of the data political economy are needed. Such alternatives must be resistant to private market governance of the data political economy, attentive to the structural incentives at the root of data extraction, and responsive to the wealth accumulation, privacy erosion, and reproduction of social oppression it facilitates.
One path forward reconceives data about people as a democratic resource. Such proposals view data not as an expression of an inner self subject to private ordering and the individual will, but as a collective resource subject to democratic ordering. Democratic data governance schemes consider the relational nature of data: information about one individual is useful (or harmful) precisely because it can be used to infer features about—and thus make decisions affecting—others. Data production places individuals in population-based relations with one another; the social effects that result cannot be adequately reduced to individualistic concerns nor can they be addressed via individual-centric institutions. Instead, data production requires legal regimes that can comprehend and govern via population-scale, democratic governance. Scenarios where the exercise of freedom of one necessarily constrains the exercise of freedom of others require public, rather than personal, governance.
This conception transforms the project of achieving freedom in the data economy from one of reclaiming individual liberty, to one of achieving positive conditions of freedom through collective obligation. Democratic data governance allows us to operationalize the normative force of the impact our actions in the digital economy have on others.14 It also recasts the egalitarian response to unjust wealth accumulation from one of personal repayment to one of collective determination of the purposes for which data infrastructures are developed and the quality of relations under which data collection occurs. Indeed, the ubiquitous and highly attuned feedback structures built for data extraction offer new possibilities for non-market social coordination that revitalize old socialist calculation debates.15 Far from offering terrain on which to re-impose private market ordering, data governance may expand the spheres of contemporary life plausibly retrieved from private governance. Conceiving of data as a democratic resource may thus better achieve relational and distributive justice and point the way towards the positive and essential role data infrastructures will play in any effective state welfare provision.
Several proto-democratic data governance projects show us how the democratic governance over information flows may be realized. German Social Democrat Leader Andrea Nahler has argued for a national data trust, likening big tech companies to pharmaceutical companies that enjoy a limited right to their products in service of the public good.16 Barcelona implemented a civic data trust to manage its data commons, democratizing data governance so that citizens have a greater say over what data is collected and for which purposes, securing stronger rights over citizen data in its contracts with technology companies, and using its data infrastructures to deepen democratic engagement.17 Labor activists are developing worker data collectives to counter growing workplace surveillance by employers, track and monitor forms of workplace oppression, and document OSHA violations and wage theft.18 Finally, old models of public data collection and management like that of the US Census and its statistical agencies, which adhere to strong purpose limitations and confidentiality rules, may be expanded into more general bodies for the governance of data.
Depending on the legal instruments involved, such solutions may well treat data as a form of property to be placed under common management or a form of labor to be governed via worker collective. However, all of the examples begin from a commitment to governing data via collective methods, and thus cast data as democratized data-capital or democratically governed data-labor. From the vantage-point of legal governance, then, the question of data’s actual classification—property, labor, or something else entirely—may be less important than the practical governance opportunity its current unsettled state affords. Regardless of what data is, the democratic shortcomings of the law governing labor and the law governing capital should leave one skeptical about slotting data neatly into these regulatory regimes. Why subject data to disempowered and marginalized labor protections, or deeply anti-democratic financial regulation? While egalitarian advocates in each area seek to gain emancipation and re-politicize critical questions of democratic governance, data governance presents an opportunity to leapfrog these struggles altogether to develop true alternatives.
Once coded as an asset in law, the conceptual distinction between capital (K) and labor (L) vanishes. Both render data as subject to relationships mediated by contracts, and with a bit of legal engineering, L is easily turned into K. Take for example partners in an LLP. They contribute their labor to the corporate entity as in-kind services and take out dividends as a shareholder in lieu of a salary, thus benefiting from the better legal protections and a lower tax rate afforded K for the same exact work that would be performed where it coded as L instead. Beyond law, the concept of “human capital” also attempts to collapse this distinction. See Katharina Pistor, The Code of Capital, Princeton University Press, 2019. ↩
But critics should be wary of relying on arguments of implementation. While such a payments system may look byzantine and Rube Goldberg-esque from a consumer perspective, as a technical matter these challenges may not be all that different from those of algorithmic auction platform systems and exchanges through which advertisers purchase impressions and clicks from users, and which similarly requires billions of instantaneous transactions at scale. ↩
It is the fact that data is commoditized at all—that ICE was able to purchase access to the database from its provider, Venntel, as opposed to gathering this data itself—that provides ICE such strong legal protection for using this data. Under Carpenter v. United States (2018), ICE may need a warrant to obtain this data from carriers or app companies directly. But because ICE simply purchased access to the database from a data broker, as would any other private entity, a possible Constitutional challenge is skirted entirely. ↩
Under Europe’s General Data Protection Regulation, these dignitarian rights are accompanied by a series of affirmative obligations imposed on data processors regarding the storage, transmission, and processing of data. Whether such affirmative obligations supersede the GDPR’s dignitarian rights framework such that they overcome its shortcomings is a matter of scholarly debate. ↩
Even a robust human rights framework applied to the data political economy may not provide meaningful redistributive effects. While human rights serve as a high-water mark of progressive universal legal rights, they remain compatible with significant and rising levels of inequality, as their embrace in corporate social responsibility efforts alongside sharply rising inequality demonstrates. See: Samuel Moyn’s Not Enough and Jessica Whyte’s The Morals of the Market. ↩