www.slate.com

"Subject: Kaplan crosses into propaganda
From: Publius
Date: May 15 2006 3:45PM

There are a number of significant questions that might be raised legitimately about the NSA's collecting a massive database of which phone numbers were connected with which other phone numbers, but the legality of creating such a database is actually beyond dispute. This has nothing whatsoever to do with FISA or any of the court cases pertaining to FISA or the Church Committee or even the reported NSA program of intercepting certain international phone calls involving persons inside the United States (which Kaplan naturally refers to as "domestic surveillance").

Indeed, Kaplan is clearly trying his damndest to make no distinction between intercepting communications -- i.e., listening to them -- and acquiring from phone companies records of calls that have been made (not incidentally, records these companies keep endlessly for their own commercial purposes and without any consent from their customers). This is not merely dishonest but comes awfully close to being sheer fakery -- political propaganda of the worst sort.

The Supreme Court held in 1979 in Maryland v. Smith that even in the case of a government agency asking a phone company to place a "pen register," which gathered information about phone numbers called and calling into the targeted phone, there was no need for a warrant since (a) a pen register does not constitute a search within then meaning of the Fourth Amendment, and (b) the phone customer has no expectation of privacy about numbers reached (as opposed to the content of conversations). In large part, this was because, even in the 1970s before total computerization, that information was routinely known to the phone companies of necessity do they could bill for their services and phone customers took that for granted. As a result of this SCOTUS decision, the fruits of that "pen register" became admissible as evidence in a criminal trial.

Today, law enforcement agencies investigating crimes routinely "dump" phones -- that is, find out from the service provider all the incoming and outgoing calls of targeted phones and their times and durations (for various investigate reasons but usually in search of suspects) all without needing to seek any warrants. Due to automation, this no longer requires placing traps on specific lines; the phone company can disgorge all the details in seconds.

What NSA apparently did, if the reports are accurate, is to get the same sort of "dump" on a large scale, not specific to targeted numbers. Presumably to gain the phone companies ongoing cooperation, the information supplied was de-personalizied. Thus, NSA got info not on Publius's phone number but only on the number without my name or other personal data. Thus, the phone companies would feel more secure about disclosing information their customers might consider inappropriate, however legal it might be.

But now we come to the issue of what use NSA made or might make of the data? Oddly, Kaplan touches on this only to the extent that generally endorses what he supposes the NSA process to be as probably a "good idea" -- just one that runs afoul of FISA, etc.

But this is the real issue: with all its computers, experience with this kind of "traffic analysis" of communications and 35,000 employees, what exactly can NSA get that is of any intelligence value out of billions of phone call records? Kaplan speculates -- and it is wild speculation -- that if we're suddenly onto suspect X, we'd want to know with whom he talked in the past and having the data bank would somehow make that happen faster or better. But any one of the phone companies can summon up such data linked to a particular phone within seconds, while on the line with their NSA liaison -- and then, if need be, retrieve calling information on any or all of the numbers X called. So, it makes no sense to spend resources to assemble such a databank for this purpose.

More likely, NSA's plan was and is to perform a sophisticated type of link or traffic analysis on this data. With what precise goal, I don't know and am not going to speculate. I think it's fair to say, though, that whatever it is, it probably takes a lot of analysts' time and other resources, given the mind-bogglingly huge amounts of data involved. Again, I don't know but I suspect that there are more targeted ways to use those time and resources that would yield more and better results.

This is one of the endemic problems of intelligence agencies: they tend to go after big, budget-building programs when they can, and it's essentially impossible for anyone -- not the President, Congressional oversight or budget committees or even the new Director of National Intelligence -- to get a handle on the effectiveness of any particular program. Yet, as in all organizations, one choice precludes another. Might it not have been more productive for NSA to have hired more linguists and analysts? Devoted more traffic analysis to the myriad of jihadist websites? Established more senior-level liaisons on the spot with foreign intelligence services?

I don't know the answer to my own questions of course, since I know basically nothing about the NSA data mining program and what it may have yielded. I have a hunch, however, that bigger is probably not better in this case."