By Gordon Hull
In the New York Times last week, Woodrow Hartzog and Evan Selinger underscore the importance of obscurity to privacy. They begin with an easy example: most of us do not remember the faces or names of those who stood in line with us the last time we purchased medicine at the drug store. Presumably, they don’t remember our faces either. Yet purchasing the drugs was done in full visibility (and invariably recorded on surveillance devices). Hartzog and Selinger note:
“Obscurity bridges this privacy gap with the idea that the parts of our lives that are hard or unlikely to be found or understood are relatively safe. It is a combination of the privacy you have in public and the privacy you have in groups. Obscurity is a barrier that can shield you from government, corporate and social snoops. And until lawmakers, corporate leaders and citizens embrace obscurity and move to protect it, your freedom and opportunities to flourish will be in jeopardy.”
They then adduce several advantages to obscurity: it gives us breathing room to live without fear of intrusion by advertising and the like; it makes intimate relations possible by allowing us to choose with whom to share information; it gives us the space to grow as individuals without surveillance; it protects us from pressures to be conventional; and it fosters civic participation by removing the fear of being on government watch lists.
The attentive reader will notice – and I assume this Hartzog and Selinger’s point – that obscurity is shown to do a lot of the work that privacy is often said to do. For example, Julie Cohen argued in an early copyright paper that the “right to read anonymously” is important to civil society and democratic participation because it enables the space to explore different opinions and views without fear of governmental repression. More generally, she has argued more recently that privacy is necessary as an aspect of subjectivity; as she puts it “privacy is shorthand for breathing room to engage in the processes of boundary management that enable and constitute self-development” (1907). Accordingly:
“But here we must come back to privacy, for the development of critical subjectivity is a realistic goal only to the extent that privacy comes into play. Subjectivity is a function of the interplay between emergent selfhood and social shaping; privacy, which inheres in the interstices of social shaping, is what permits that interplay to occur. Privacy is not a fixed condition that can be distilled to an essential core …. It enables situated subjects to navigate within preexisting cultural and social matrices, creating spaces for the play and the work of self-making.” (1911)
If this is what privacy is for, then why might obscurity help attain that? A central problem is that privacy is poorly understood by policymakers, being generally defined in terms of whether someone has chosen to reveal information publicly. This construction of privacy as an individual right is unfortunate at best. Companies want our data, and so privacy doctrine construes data as something of value that we then trade for someething else we want – say access to a web service. The model is congruent with market norms that push toward ever more disclosure and tends to lead to a market failure to meet consumer preferences. Consumers are completely unable to effectuate their preferences, and the model subtly teaches us that this is the right way to view privacy.
What I want to emphasize here is a point behind this: the construal of privacy as a market transaction pushes toward contracts as a method of enforcement because, as we have known since Hobbes, enforceable contracts are how you overcome the irrationality of trusting those with whom you may not have future interactions.* Contracts successfully protect rights when the terms are fully disclosed and freely agreed to, and when there is recourse to the state for enforcement when private parties violate those terms. Viewing privacy through the lens of contract inevitably pushes toward openness: as Judge Posner concluded in an influential economically-based reading, a claim to privacy is likely an effort to withhold information that would make a transaction more efficient.
More importantly, none of these conditions in the background of a contracts view actually hold when the issue is privacy. Even if we assume (against all evidence) that privacy policies are disclosed in a way that consumers can understand and digest them, it is hardly optional to opt-out of all websites and credit-card payments and telephone use to exercise one’s privacy right. Each of these activities involves a substantial data footprint and they are all-but mandatory in modern society. In addition, recourse to the state to enforce these supposeldy contractual terms is also shaky: for one, we are highly unlikely to notice when companies violate their own privacy policies or inadequtely protect data (this is part of why they fight rules that require disclosre of privacy breaches).
In addition, the remedy for contract breach is private enforcement. I can go to court to force the other party to compensate me for whatever it did in breach of the contract. But the costs to individuals for engaging in this strategy are likely to be prohibitive, and in any case, it is notoriously hard to monetize in a judicially legible way the damage associated with a privacy breach. In that context, the state would need to empower an enforcement agency, but doing so would require overcoming significant legislative obstacles. These obstacles go beyond questions of lobbying money or political capture; as Lior Strahilevitz notes, they are structural:
“From a median voter perspective, privacy protections limiting firms’ ability to collect or use data probably benefit unsophisticated consumers (who do not engage in strategic disclosure) and consumers who have worrisome profiles. They likely harm sophisticated consumers and those with less worrisome profiles. This dynamic may in turn help explain why the U.S. legal system has been so tolerant of data mining” (2031)
In such a situation, we have a substantial disconnect between a legal regime that is structured according to the assumption that we engage in a market norm and a de facto situation that is governed by other norms. As Samuel Bowles argued some time ago, contract-based regmies work to replace reputationally-based ones. Given the security of contracts, individuals don’t have to worry so much about the other party’s reputation; state enforcement reduces the need for that. Thus, “where markets govern the exchange of well defined (meaning third party enforceable) property rights, reputations of any kind will tend to be both costly for people to acquire and of little benefit to those who do, and for these reasons unlikely to be favored by differential replication” (92). But here we have a theoretical model that assumes complete contracting, but a lack of complete contracting on the ground. Bowles: “markets may undermine the reproduction of traits necessary for eficient market transactions in the absence of complete transacting” (92). In other words, the assumption that privacy is based on efficient markets undermines the sorts of norms that would enable markets in privacy to function.
Obscurity takes a short cut around the economization of privacy. It recognizes that information about individuals is likely to be “out there.” It thus builds upon at least two earlier strains in prvacy theory. One is Helen Nissenbaum’s well-known exploration of privacy in terms of “contextual integrity” – roughly, that information is only ever revealed in a context, and that privacy concerns may change as the information switches from one context to another. For example, a given disclosure that’s apporpriate in a doctor-patient relationship becomes inappropriate when removed from that context. The move to obscurity also recalls Lior Strahilevtiz’s social network theory of privacy, which argues (to oversmplify) that one has a reasonable expectation that much information will not circulate widely, because it will generally stay confined to close-knit social networks. In other words, plaintiffs in a privacy case can argue that information about them would not likely have had a wide audience, but for defendant’s actions. Strahilevitz also explicitly makes a connection to obscurity of the sort that Hartzog and Selinger make: “noise has long been an important method of protecting privacy” (969), he suggests, and notes that one way to have a private conversation is to go to a crowded pub where there’s a lot of noise and nobody cares about you.
The reference to noise also recalls Donna Haraway, who noted in her 1985 “Cyborg Manifesto:”
“Cyborg politics is the struggle for language and the struggle against perfect communication, against the one code that translates all meaning perfectly, the central dogma of phallogocentrism. That is why cyborg politics insist on noise and advocate pollution, rejoicing in the illegitimate fusions of animal and machine” (55 in the linked edition; 176 in the 1991 book)
Critical subjectivity, to return to Cohen’s formulation, requires enough noise to derail the process of algorithms. It is perhaps no wonder that data scientists have been experimenting with “differential privacy” – protecting against re-identification in data sets by introducing random noise to obscure the contribution that an individual record makes. Given the importance to the twentieth-century imaginary of the ability to disappear into large urban areas, it is perhaps also not surprising that smart city initiatives appear deeply suspicious to privacy advocates.
In sum, it seems to me that obscurity thus makes terminologically explicit something that is important under the surface of some contemporary privacy theory as well as theories of resistance to late capitalism: not just that the individual rights model of privacy fails, but that the economization that either accompanies or undergirds it is itself deeply problematic. In this context, obscurity – noise – is an important stand in for the normative value of inefficiency (with Brett Frischmann, Selinger also makes a case for inefficiency in a recent book). Obscurity, in other words, also stands for the proposition that economics needs to stay in its lane.
* This is my reading of Hobbes’s account of the fool in Leviathan L 15.4. I defend it here, pp. 99-117. My reading of the passage on the fool is indebted to Pasquale Pasquino’s and on the state of nature more generally to Yves Charles Zarka. To the extent that games theoretic readings are appropriate, I tend to think (against Jean Hampton’s account) that (in Pasquino’s phrasing) “non-cooperation is the equilibrium state." See also the helpful account by Kinch Hoekstra.
Recent Comments