To Protect Private Data, Collect It
June, 2015
The recent passage of the USA Freedom Act and the second anniversary of Snowden’s release of information calls for serious reflection on the nature of the security state, as Snowden himself has done in a recent New York Times editorial. Unfortunately, while the NSA’s program of data collection has been heavily critiqued, we have little discussed what the potential consequences of the project actually are. The traditional line taken by Snowden and his various supporters (or supporters of his line of thinking if not the man himself, including individuals such as Senator Paul), is that data collection by itself is a breach of privacy and results in tyranny. Many people vaunt this thinking as powerful, democratic wisdom. On further examination, however, this idea is more problematic than it appears. While it is tempting to assume that any collection of private information by the government directly threatens individual liberty, as the experience of the twentieth century indicates, this argument lacks a clear mechanism connecting the collection of private information with tyrannical government. Without this connection, it relies on the assumption that where private data collection goes, tyranny must follow ipso facto. Unfortunately for those who pose this argument, but fortunately for those of us who must live under systems where the collection of private data for a variety of purposes is increasingly necessary, this connection is neither absolute nor automatic. Instead, the connection between data collection and oppression must be implemented through organizational and bureaucratic channels to realize a tyranny which is otherwise only potential at best. As such when examining the threat data collection poses to our liberty, we should not primarily examine the collection itself, but the mechanisms that surround and enable it.
Before I examine this point, I will explore the logic behind the assumption that data collection directly involves tyranny. It emerges from the relationship we have developed with the data that informs and surrounds our lives. For us, data is ubiquitous. Not only is it a constant subject of conversation, such as citing a study/survey, it is also something we constantly engage, such as using it to search the internet or order food through an app. It is also something that is promised to us as a panacea to alleviate social and physical ills. Data and its products, we believe, will generate technology that will ease global warming, create self-driven cars, cure cancer, etc. I am not citing these examples to critique our use of data. There would be little point in doing so as it will not change our relationship with it, nor am I convinced that that relationship should change. Instead, I intend to indicate that data is not merely omnipresent in society, but we perceive it to have a kind of omnipotent quality as well.
The idea of a social phenomenon that is both omnipresent and omnipotent, at least in appearance, is not new. In the nineteenth century, the emergence of capitalism seemed to create a similarly all-encompassing, agent-laden impersonal phenomenon in economic transactions manifested through the commodity. Marx, the great critic of capitalism, gave this emergent trend a name: commodity fetishism. In Capital, Marx explains how commodity fetishism works: through the labor for the creation of an item, man invests himself in it. After, he exchanges that item for another, something he needs more than the first. This exchange creates social relationships which take on lives of their own, becoming embedded in the objects of exchange themselves and forming the basis of many if not all aspects of society. As a result, people confuse the act of exchange for the power of social relations, removing social affect from the hands of human actors and transferring it to the commodity and the realm of material exchange. As a result, rather than acting and interacting with each other directly, people become bound by their material circumstances and relate to each other through objects, giving these objects outsized levels of agency. One of Marx’s students, Georg Lukács developed this idea into the concept of “reification,” arguing that people can assign agency and power to inanimate phenomena generally, which then determine social relations.
Without commenting on the accuracy of Marx’s description of capitalistic society, we observe that the assumption he and Lukácz indicate, i.e. that people assign action and agency to material relations rather than social actors who enact them, is certainly true as a human tendency. The simplest (and perhaps overly simple) examples we can cite from our own society are the idea of “luck” being embedded in an athlete’s old socks or the sacral power of a relic to heal. We reify data similarly. Our reification of data is in fact analogous to how Marx argued that society reified commodities, i.e. we often assume that there is some inherent power that resides in data eo ipso and that the possession of data is somehow enough, by itself, to affect action and influence in social relations. In short, we create “data fetishism.”
It is this data fetishism that led Snowden to his crusade. While he has discussed at great length how the US government collects data, who has access to it and what is collected, he has never explained why this is a threat or how the government uses it to oppress either the American population or the world at large. Instead, he and his supporters rely on the assumption that simply by collecting data there is an existential threat to American freedom, but this argument is only coherent in so far as the data itself is fetishized, i.e. is assumed to have some power by itself such that its possession will affect oppression and tyranny in a similar way that old socks magically (the fetish works through magic) improve one’s ability to play ball. What this position overlooks is that data collection alone is neither a sufficient nor a necessary condition of tyranny. Instead, what must be in place is a political and legal structure (a constitution) that enables a government to use data tyrannically. Tyranny, in other words, is defined not by possession, but by action.
If we examine the historical examples that form the bedrock of our concern, this argument will become clearer. The twentieth century saw the rise of regimes that used the possession of private information to further the iron-fisted domination of a state over its people. Some of the most prominent include Stalinist Russia, Hitler’s Germany, Maoist China and the Kim regime in North Korea. It is important to note that the emergent tyrannies in these states did not gain their power through exploiting control over the private lives of citizens, but in each case the totalitarian grip on power emerged before any government agent could breach the inner world of its population.
Instead, the manipulation of private lives for the gain of the dictator and his party developed from the removal of political power from the public, governmental apparatus that at least officially supported individual rights and that authority’s transfer to the bureaucratic organization of the party, which had previous experience policing the inner lives of its members, but little access to the private lives of the public until it came to power. It was only after the totalitarian party transferred control of the government from the state to itself, which happened in each case with public acquiescence, that government tyranny over citizens’ private lives became effective. Even then, it was often not governmental departments that operated intrusive oppression. Such oppression remained largely within the party organs themselves. In each case, a constitutional structure that could use private information for tyrannical purposes had to be developed before the exploitation of private lives could be used as a tool of oppression and its development happened outside of the official government of the state (for a discussion of this process, see Arendt’s The Origins of Totalitarianism, Totalitarianism).
The thrust of this point, and what the history shows, is that the collection of private data does not per se catalyze tyranny. Instead, the functioning bureaucratic apparatus capable of using that information must be implemented before such data can be used. Of course, one can argue that having a large collection of private data makes oppression easier if such a party were to come to power, but the same argument holds true about a standing military. In the hands of a tyrant, a military becomes not a tool of support for the state, but one of oppression for her people as the experience of twentieth century Latin America and Africa shows. Many Founding Fathers, including most famously Thomas Jefferson, used this argument to prevent the formation of a standing army. Despite such fears, when the army was institutionalized the tyranny predicted did not materialize. The reason for this was not the impossibility of its tyranny, but because the Constitution was written in such a way that the powers and interests of the military are divided between its civilian commander-in-chief and Congress who holds its budget. As a result of this division, and the fact that the military does not control itself but reports to civilian authority, the potential abuse of power it represented was headed off and a martial coup d’etat rendered an institutional impossibility.
Unfortunately, few if any discussions of the NSA data collection project have considered this point or its nuances. Most have paralleled the farce conducted by the comedian-cum-critic John Oliver who, when interviewing Snowden, focused on whether the government had access to pictures a hypothetical person might take of their genitalia. Nowhere in this discussion was the question raised either of how the government’s possession of a genital picture would impact liberty or how likely it was that that picture might actually ever be seen by state authorities.
This latter point, namely how practical it is for the government to use private information it collects to affect tyranny, is another question that while linked to the concern for the institutional capabilities of oppression is both distinct and similarly under-discussed. Most Americans agree that some collection of information is necessary and even expected to preserve their security. I argue that if we must collect some information, it is far better to collect information on the large scale, for this, counter-intuitively, reduces the ability of the government actually to oppress the citizenry as a whole because analyzing such large quantities of data is extremely difficult.
If we consider the hypothetical person’s picture above, it is such a small quantum that if considered against the entire set of data collected by the US government, isolating and analyzing it is impossible. This is because the government cannot actively screen all the pieces of metadata that it collects either in real time or in delay. The computational effort is too vast and the energy required too large. It stands to reason that if the government cannot analyze something, it cannot use it for nefarious purposes. Far more dangerous would be limiting the collection of information to a sub-set of individuals, for that information would fall within the realm of analytical possibility.
Of course, this in and of itself is difficult because of constitutional restrictions on data abuse. In the case of the data collected by the NSA, we remember that the NSA has no policing power. It cannot arrest or interrogate an individual. If it were to desire such an outcome, it would have to cooperate with another agency. The FBI is the most representative example. The FBI would face a problem, however, for such information cannot be used to arrest a subject or, assuming that subject is arrested under another pretext, to prosecute that subject unless the data collected by the NSA resulted from a court-approved wiretap.
In other words, in order for political oppression to emerge from the data collection program, the NSA would have to collude with the FBI who would have to sway a judge who ultimately would have to convince a jury to convict a person. Even this scenario is too simplistic, however, for it ignores the counsel the accused would have for his defense and the indirect mechanisms of protection such as the media they could alert for support. Ultimately, all of these actors would have to be co-opted for oppression to operate, a point support by each historical precedent.
Even if this came to pass, why would this chain of events occur in the first place? Who would be the nefarious instigator of this chain? Could the President do so without attacks from Congress, and any who would seek to take his place in the next election, undermining his efforts? Could Congress do so with the tumult that would result within that body? To what end could this chain of oppression move with a myriad standing to gain from the stumbles of a would-be oppressor?
Of course, one could argue that an entity might circumvent the legal process for some extra-judiciary action, but again, this is where the importance of the anonymity of the herd resides. While it is easy to protect individuals as a generic group formally within the legal system, that same system might be unaware of or ignore small-fry that agencies could abuse outside of the popular eye or to popular apathy, but if that individual or group of individuals is nearly impossible to winnow down from the massive data collected, if the effort to bring the individual in is sufficiently great because of the amount of and restrictions upon data’s use, then that volume stands as protection for liberty.
These points do not mean that the system cannot be reformed, but it casts doubt on the necessity of Snowden’s treason and the cogency of his suggestions for reform. Cessation of data collection will not solve our problems, for government will restrict the field, leaving it open to the selective abuse described above. This is true because total cessation is impossible in a world where we lead lives, whether licit or illicit, in digital worlds that are by necessity traced, whether that necessity emerges from the police or from our need to have groceries delivered, pay bills online or apply for medical procedures at a hospital’s website.
What we need to reform, therefore, is not the collection itself, but how that collection is controlled and institutionalized. Having a secret court be the ultimate venue for oversight is probably unwise and unsound. Greater disclosure of the program’s benefits, such as can be made while preserving its integrity, are also likely in order. Involving both the Congress and the President in its operations is also necessary as the balance of their interests (and their potential mutual suspicion driven by differing ideologies and self-centered political opportunism) will prevent practical abuse from emerging as one will always have popular support to gain from outing the other’s abuses.
Another point which is sometimes but not often enough addressed is the selectivity with which we approach our concerns over privacy. We protest the government’s possession of our data, but are comfortable with Google, Apple, AT&T and others having huge amounts of it with little question. This is quixotic, for while the possessors of state power are beholden to us as a citizenry, executives at Google are not. While a proponent of free corporate action might lodge a reasonable objection that shareholders can exercise control over company policy either by voting directly or by exercising authority indirectly through appointing board-members, as so-called activist shareholders do, and thereby restrict possible abuse of private information, this objection is not applicable to the major and most threatening data companies like Google and Facebook. These companies possess a specific two-tier share structure that deprives any but the inner circle from voting rights and ensures that the executives of the company can exercise sole discretion over their use of our information.
Given the choice, if I had one option I would suggest the government and its accountability to be safer, but fortunately we do not have to make that decision. One of the advantages to the USA Freedom Act as passed is its creating a balance of powers between providers of data technology and the state, for their interests are not in-line and mutual supervision will prevent either from exceeding their mandate.
How this operates is simple: the federal government is beholden to the people both legally and through their control over elections. As a result, it is responsive to their demands when they feel their freedom is unduly restricted (as the current debate evinces). This responsiveness can extend to regulating companies. Consequently, if private companies use any of our data for nefarious purposes, such as commercial targeting, blackmail, etc., the state can intervene. The companies, however, have their own interests when it comes to preserving business. In this case, American technology companies have lost patronage as a result of privacy concerns. Consequently, they have an interest to police the government’s use of their data and to seek the necessary court injunctions to prevent unauthorized disclosures. The result is a system that balances the interests of each party, creating two watchdogs, the one of the government observing the companies and the other of the companies observing the government.
Of course, we should remember that both corporation and state have the potential to abuse the people, which is why transparency and due-process is required in the legal architecture both surrounding the collection of data itself and the political control of those elected officials who are ultimately responsible for approving intelligence agencies’ budgets, objectives and operating structures.
Considering this discussion, it is not the cessation of data collection itself that is important, but the ability of the people to supervise and control the access to and use of that data that is important. Ultimately, it is the operation of the political system itself and not the state’s collection of data that matters.
At the end of the smug and self-congratulatory letter sent from his hide-out, Snowden writes of current legislation that “…we see the outline of a politics that turns away from reaction and fear in favor of resilience and reason.” He is correct in the sense that the bill as passed is more reasonable than the program as it had been executed, but he is wrong in assuming that the previous collection project was an unreasonable, unguided tyranny just as he is wrong in assuming that his stance is a completely rational or resilient one. It is a popular one in many circles, that is true, but it is based on a knee-jerk opposition to authority, the extraordinary fetishization of data and a simple lack of consideration for the deeper historical, political, social and legal factors that build tyranny as well as a fault in understanding what tyranny itself actually is. Were he to consider these facts in a sharper and clearer light, he might be enlightened by actual reason instead of the fallacious thinking he claims to support, but instead he prefers the image of the self-satisfied radical and would-be martyr.