06 June 2013

Preventing Pearl Harbor

I am outraged — but not surprised — by the revelations of sweeping, not-otherwise-authorized collection of cell phone message-unit data disclosed in today's Grauniad. What follows is not a defense of this conduct; it is at most an explanation, slightly sanitized due to my own nondisclosure agreement, of "what the hell they were thinking?" That I understand the mindset behind it does not mean I share it.

In intelligence-gathing terms, what has been revealed thus far is an effort to collect data for traffic analysis. The intelligence/counterintelligence meme of traffic analysis is a refinement of "where's there's smoke, there's fire": Knowing who talks to whom, when, and how much/long is extremely valuable information in and of itself. A change in traffic characteristics — typically, but not always, a spike or valley — can in a military context be a highly accurate indication of a forthcoming change in posture or operations. Too, the ability to draw a communications net helps focus on the actual communications at issue... and helps determine which members of a conspiracy or operation are worthy of further attention.

And that, rather precisely, is the problem: The presumption of existing hostilities. In the military-intelligence context, one is required to always assume that all communications are potentially related to hostile action. Further, proper intelligence-related traffic analysis begins from the equivalent of "reasonable suspicion": One does not start looking at the call records of a sendai in Honolulu until one has at least some reason to believe that there's a connection between that particular sendai and hostile activity.1 The gathering of this information in the first place represents a kind of subtle constructive paranoia: That once there's a specific connection to a staff officer at the Japanese Consulate in Honolulu (the Naval Attache trying to determine what vessels are in Pearl Harbor in early December 1941, to be specific), nobody wants to take the time or trouble of getting even a FISA warrant for the communications of that sendai with people other than that staff officer. The second level of paranoia is that nobody wants to reveal that the intelligence-gatherers think that information might be important by disclosing to a third party (the telephone company) what is being sought; therefore, blanket collection of data so that it can be analyzed behind whatever firewalls there are is "best."

But note that this is a contextless justification for broad gathering of data that is extraordinarily subject to misinterpretation and misuse. Traffic analysis is, above all, an art; the problem is that it is cloaked in relatively hard data and uses statistical methods, making it seem to the unsophisticated (such as every member of Congress who voter in favor of the TRAITOR Act,2 and especially its domestic-intelligence-gathering provisions) more scientific than it is. Constructive paranoia can be useful when confined to individual analysts scrutinizing a particular, well-defined problem; it's extremely dangerous to everyone when it becomes an institutional mindset. One of the downsides of traffic analysis is that it is, in reality, searching for needles in haystacks... and not all needles are properly the subject of FISA warrants and the TRAITOR Act in the first place.3 More to the point, traffic analysis has a hellaciously bad proportion of false-positive indications, no matter how sophisticated its analysis, without full consideration of human-intelligence sources... and the legal term for that consideration is "probable cause."

There's a helluva lot of blame to go around here. Remember what is paved with good intentions, even intentions like "prevent mass terrorist attacks." Although there are already no doubt rumblings and denials of the authenticity of this information (indeed, it's entirely probable that the program as implemented varies somewhat from what the documentation says — just like every other automated electronic system in the world, including the operating system of the computer on which or through which you're reading this), it's too consistent with what "the authorities" have been doing since 2001 — and have wanted to do since the 1940s — to be dismissed. This seems like yet another prelude to yet another bad episode of 24... and very, very few people have ever considered the graveyard full of innocent reputations and bodies left behind by Jack Bauer in his minimally successful efforts. If you stop and think about it (and it's readily apparent that almost nobody "inside" has done so since the late 1990s), they're victims, too... and a strictly Benthamist analysis ("the greatest good for the greatest number") at least calls into question the balance of victimhoods.

I'm uncomfortable enough with Verizon having all of this data in the first place; sharing it with someone else with whom I did not consent to "do business" is unacceptable. And, by the way, the same goes in reverse; this isn't about the Gummint, but about anyone who seeks to fill an abuse-of-power vacuum, including Verizon itself.


  1. Cf. Korematsu v. U.S., 323 U.S. 214 (1944); Hirabayashi v. U.S., 320 U.S. 81 (1943). These Second Thirty Years' War cases are every bit as damaging — and wrong — as was Plessy v. Ferguson, 163 U.S. 537 (1896); it's past time to overturn them. Their echoes in the kinds of people targeted in the current "War on Terror" are rather disquieting.
  2. The USA Totalitarian Regime Activity Incitement To Obscure Reality Act, Pub. L. No. 107–56. Some people, including its proponents, mislabelled this the USA PATRIOT Act, thereby demonstrating that they were not themselves patriots: They had to destroy civil liberties and representative democracy to save them. Riiiiiiight. What part of "protect and defend the Constitution of the United States against all enemies, foreign and domestic" did they not understand... aside from "all of it"?
  3. That's why using this to go after "organized crime" is so wrong. That's why using this to obtain political blackmail material is so wrong. And it will be so used, because to the kind of people who would use these mechanisms for such ends, the ends justify the means... instead of being shaped by them.