One thing struck me immediately about the recent al Qaeda threat warning, when U.S. officials claimed that the Hoover-like NSA surveillance program was what enabled us to detect the threat. Senator Saxby Chambliss (R-GA), ranking member on the Senate Intelligence Committee, was emphatic in making this claim:
Those [NSA] programs “allow us to have the ability to gather this chatter,” Chambliss said. “If we did not have these programs then we simply wouldn’t be able to listen in on the bad guys.”
What struck me was that this simply isn’t true – at least, not in terms of the comprehensive implication Chambliss seems to suggest. We’ve been able to listen in on the bad guys for as long as there have been cell phones. Identifying them specifically when they’re in chatter can range from easy to hard, but when they’re in Yemen and we have a list of the terror suspects we’re especially interested in, and a growing record of their associations, it doesn’t take a metadata-analysis program to cue us that there may be significance in their communications.
It certainly doesn’t take a program that vacuums in the communications metadata of millions of Americans, and stores it for years. That enormous, constitutionally problematic data set is irrelevant to listening in on al Qaeda terrorists in Yemen.
The theory of an omnidirectional metadata-crunching approach is that looking for patterns in the metadata will give us potentially relevant events – phone calls, texts, emails – to focus more attention on. Basically, it’s a pure “profiling” process: create suspicion, which will focus our efforts, through the observation of patterns throughout the universe of metadata.
There is nothing invalid about the data-crunching idea, per se; what’s invalid is suggesting that we needed it to know that al Qaeda terrorists may be talking to each other via phone at any given time in Yemen. That’s pretty much like saying we need to record every license plate on the 405 freeway in Los Angeles to know that illegals may be driving on it.
Of course al Qaeda operatives are chattering in Yemen. That’s a known, and extensively developed, enemy pattern. If we needed a metadata-crunching program, and in particular, one vacuuming in data on billions of American communications events, to tell us it might be happening, then we’re too dumb to survive.
The point is that we didn’t need uncued, omnidirectional data collection to vector us in on the communications of a demographic we have had under intensive surveillance for more than a decade.
I consider it quite possible that we were vectored onto specific communications, at least in the first instance, by a cued, and undoubtedly filtered, use of the crunching algorithms employed in the NSA programs. Their utility for picking out relationships in a universe of “connection” events among individuals is unquestioned.
But that’s not the same thing as suggesting, based on the recent al Qaeda threat, that we have to buy wholesale into the uncued – blind – data-vacuuming approach, because without it, we “simply wouldn’t be able to listen in on the bad guys.” Horse feathers.
With al Qaeda in the Arabian Peninsula (AQIP), we already have the biggest hurdle – identifying the collection target – out of the way. With a well-developed target like AQIP, moreover, all of the sources of intelligence are regularly brought to bear. There should be no doubt that we are getting “HUMINT” – intelligence direct from individual humans – on AQIP, which serves to improve the focus of other-source collection efforts. U.S. intelligence officials who spoke to The Washington Times last week voiced concerns about human intelligence sources, given the Obama administration’s unusually detailed explanation of how we knew about the threat:
Intelligence officials are dismayed that the administration provided so much detail on what prompted the closings, and that the disclosures could work against obtaining new information. Militants are now likely searching for the sources of the information to both the U.S. and Yemeni officials, and almost certainly will kill anyone they suspect of working with Western intelligence. …
Other sources are also likely to reconsider their relationship with the United States over the disclosures. “These guys know their lives are in danger. As soon as the U.S. shows we can’t be trusted, they will go under ground and we won’t hear from them again,” says a current intelligence officer.
In the normal course of intelligence-gathering on a target like AQIP, HUMINT would figure as significant; the more we have of it – and of other intelligence sources on a target – the less significant the blind crunching of metadata is. It is reasonable to suppose that, in this case of AQIP communications, blind metadata-crunching was not significant at all to detecting the particular threat over which our embassies have been closed.
It is precisely the blind metadata collection and crunching that Americans object to; specifically, the collection of our metadata. (You want to collect metadata on Azerbaijanis, Germans, or Brazilians, knock yourself out. They don’t have any expectation of protection under the U.S. Constitution. Their telecom carriers may not be very cooperative, of course.)
Detection of the recent al Qaeda threat certainly does not justify this collection. And, in general, there is zero likelihood that we will fail to detect a terror plot if the U.S. government doesn’t keep a record of Joe Smith’s call to Pizza Hut, or Jameela Jones’s email to her financial planner. The implication that gathering and keeping such records is necessary is an unreasonable – absurd – proposition.
Moreover, we can have the benefits of pattern-recognition algorithms without buying into the blind-collection proposition. The two are separable – and that’s a decisive point. The likelihood that the omnidirectional data set we don’t collect will be the only thing cueing us to the bad guys is vanishingly low. That said, the pattern-recognition algorithms are indispensable – if we have at least one factor to filter on, such as the identity of one individual suspected to be an associate of terrorists, or even something more general, like where in the United States the concentrations of Islamists with established connections to terrorists are.
The metadata fight is a fight over our government’s fundamental relationship to us. It is a dodge to point out that there is no expectation of individual privacy related to the metadata, which is owned by the telecom and IT companies anyway. That’s not the point. The point is what the assumed posture of our government is to be: universal prior suspicion of the American public? Or an institutionalized and constitutionally restrained posture of non-suspicion, in the absence of positive indications from an individual’s activities?
What are government’s powers to be, if the former is our assumption? This issue is another of the tipping points we are encountering, right at this very moment, in the trend of man-and-the-state relations in America and the West. What is government assumed to “own” and have authority over – everything that doesn’t belong to the individual?
Is there no such thing as the social or commercial interaction that is not assumed to be under government supervision in all its aspects?
If not, why are there not vast tracts of human-interaction space over which no centralized authority has supervisory “rights”? How has our view of human life, authority, and moral meaning been transformed, that we no longer have the vision of genuine liberty and truly limited government?
For the purposes of law, in the context of our answers to these questions, we may ask this: even if a commercial company owns the metadata of our communications, should not we expect, a priori, that there will be no resort to it, by government agencies, without specific and documented suspicions registered about individuals in an auditable process? That is what we have long considered our Fourth Amendment expectation to be.
The premise of the blind-metadata program’s advocates is precisely the opposite: that we should expect, a priori, that government may have constant access to commercial companies’ data on our transactions, independent of any suspicious manifestations on our part. This premise is actually more like the premise of Stalinism than like any other model of government. What does it imply about the scope of government’s rights and authority over us? Why isn’t the bias in the other direction?
We must come to grips with these questions, in this age of advanced information technology. In the absence of a positive determination on them by the body politic, the American people are currently, in essence, the subjects of the same spying regime our government uses against foreign targets. No government, in any time or place, can be trusted with this level of surveillance against its own people.
This is an urgent and very big issue. It doesn’t help keep the issue clear in our minds, for public officials to make heroic but unreasonable claims about the intelligence collection program when a threat has been announced. The point here is not to accuse anyone of cynical manipulation of the news; I’ll let others wage that war, and note that it’s mostly a fruitless endeavor anyway. The point is to get this man-and-the-state business sorted out. It will kill the republic if we don’t.