Perhaps it would be of interest to the ongoing discussion about NSA accumulation of phone use metadata to see how some math gets mixed with metadata in a relatively simple universe. Professor Kieran Healy does just that (found via memeorandum) using a database of 256 names of socially active Bostonians from 1775. Based only on identification of individuals with seven suspected subversive groups,
Prof. Healy is able to use three mathematical tools to identify a particular “person of interest” … Paul Revere … as someone that the government might want to monitor very closely. As it turns out, this would have been correct; Revere was not an independent ringleader of the various revolutionary groups but he was well-connected with nearly everyone from Samuel Adams to Joseph Warren to Robert Newman.
I suppose I had some inkling that modern mathematical tools are this powerful. And to some extent, I have to take Prof. Healy’s word for it both that the math has been properly applied and that the data selected is fair (if its subject matter is provocative) because the math itself is frankly beyond me. But given that demonstration, it takes little imagination to consider what immensely more powerful computers, and immensely richer data, might cook up just from seeing the fact that A communicated with B, and then B communicated with C, and so on.
Many people, myself included, are gravely worried about the privacy implications of all that we have learned in the past few days as a result of Edward Snowden‘s leak. But let me add something new to the mix: the First Amendment’s right of freedom of association. In 1958, the Supreme Court held that the state of Alabama could not compel disclosure of NAACP’s membership list, because to do so would intolerably chill the right of people to associate with one another to advocate for political change.
If Alabama’s Attorney General had access to metadata such as is already held by the Federal government, it wouldn’t have needed to have been given a copy of the list. It would have only needed to have known the names of a few suspected leaders within the group, leaders who could have been readily identified by their own public activities. From there, seeing who they contacted and who contacted those people, it wouldn’t have taken long to come up with a statistically reliable extrapolation of everyone important in the group.
In a way, I’m less uncomfortable with the NSA using search algorithms that actually probe the content of communications than I am with them conducting all this complex analysis of communications metadata, particularly after seeing how powerful that metadata is when coupled with sophisticated math and powerful computers. An algorithm that looks for content has some kind of targeting to it, some sort of goal that it’s looking for. That sort of information “destination” is something that can be evaluated and understood. And communications metadata from someone who has already been identified from other sorts of evidence — a pen register — is useful intelligence and law enforcement information that we should all want the government to have when it goes after actual bad guys.
As a matter of law, pen registers are not protected from warrantless search by the Fourth Amendment. But I don’t think we’re talking about a pen register anymore with this NSA thing. A pen register is related to a specific telephone number. The data that’s been gathered by the government since 2008 has not been limited in any way at all, as set forth on the second page of a judicial opinion which I’m frankly more upset by than the fact that the government asked for the information in the first place. It’s every phone call that’s been made or received. Every text message that’s been sent or received. Every e-mail, every tweet, every post, every communication of any kind sent over the various carriers, who are basically all of them that matter.
Remember that the government has already shown a propensity to use much more clumsy means to identify much less dangerous political adversaries. And while they’re really sorry about that (in the wake of its becoming public knowledge), it demonstrates why we ought to think carefully before we trust the government to have data like this in the first place. Maybe no law has been broken by the NSA; it did, after all, get a warrant in the FISA court.
But there ought to be a law to rein this sort of thing in, and there ought to be more critical oversight than has apparently taken place, because so much information under such tight control in the hands of so few people with so much power is a situation well-fertilized to grow into profound abuse.