Wednesday, May 17, 2006

David Ignatius attempts to argue that the NSA monitoring of all US calls is useful.

To do so he attempts to escape into a fantasy world. The absurdity of his argument is made clear from the 4th word in his column "hypothetical". It is possible to imagine a world in which the program were useful as it is possible to imagine a world in which pigs fly. In this world, it is clear that Ignatius's argument is nonsensical.

Let's take a hypothetical problem: An al-Qaeda operative decides to switch cellphones to prevent the National Security Agency from monitoring his calls. How does the NSA identify his new cellphone number? How does it winnow down a haystack with several hundred million pieces of straw so that it can find the deadly needle?

The problem may seem hopelessly complex, but if you use common sense, you can see how the NSA has tried to solve it. Suppose you lost your own cellphone and bought a new one, and people really needed to find out that new number. If they could search all calling records, they would soon find a number with the same pattern of traffic as your old one -- calls to your spouse, your kids, your office, your golf buddies. They wouldn't have to listen to the calls themselves to know it was your phone. Simple pattern analysis would be adequate -- so long as they had access to all the records.

This is almost too obvious to need arguing. I only do so because Kevin Drum seems to be semi convinced by Ignatius. The hypothetical example assumes something we know not to be the case and is irrelevant to the real world. It assumes that there, on September 11, there was a known al Qaeda operative in the USA and that the problem was to find out which cell phone he was using. We know there was no such operative. He would undoubtably have been arrested if his location was known. If it wasn't his photo would be in every post office in the country. Sometimes known criminals are allowed to remain free in an effort to catch conspirators. 9/12/01 was not one of those times.

Any effort to argue that the program is useful requires the definitely false assumption that al Qaeda operative calling patterns are known. They can only be learned by observing the calling behavior of known al Qaeda operatives. It is obvious that there were no known al Qaeda sleepers in the USA when the program began. It is almost as clear that none have been found since then, because the event would have been trumpeted for political reasons. Thus the hypotheical example of how the program might be useful is clearly irrelevant to the real world. The worlds largest data base and the worlds most powerful computers can not identify a pattern typical of something without examples of that thing. This is obvious.

Uh oh I am having trouble respecting fair use limits. I only want to highlight the clearly absurd passages in the column but I find myself quoting the whole thing.

This, in simple terms, is what I suspect the NSA has done in tracking potential sleeper cells in the United States. The agency can sift through the haystack, if (and probably only if) it can search all the phone and e-mail records for links to numbers on a terrorist watch list. The computers do the work: They can examine hundreds of millions of calls to find the few red-hot links --

Here he claims that having all phone and e-mail records, the NSA can "sift" through them. The hypothetical example of how it could sift is clearly purely hypothetical as it is based on a counterfactual premise. Yet Ignatius asserts definitely that the NSA can sift through phone records not that, for all we know, they might have the necessary sample of al Qaeda operatives but that they can. Having a lot of something does not imply the ability to sift it. You need a sifter to do that. The computers can do a lot of work, but if there is no data on typical al Qaeda operatives, their output is only as good as the human guess as to what al Qaeda operatives do. The assertion that computers have magical powers and can tell which links are red hot is techno mysticism. If you can't tell a plausible story about how to find the al Qaeda operative in a sample of three people using pen and paper, you can't convince me that a computer can find the al Qaeda operative out of 280 million people. Computers can solve problems which require too many calculations for a person. Computers can not solve problems which aren't scaled up versions of problems which we can solve.
The proof that the explanation of what is really happening is hypothetical is that Ignatius assumes the existence of a valid "terrorist watch list". I ask how was that obtained ?

Ignatius does not even attempt to defend his really crucial claim "(and probably only if)." The claim is simply stated. Since Ignatius feels free to assert that his conclusion (the NSA can sift) is true, his use of the word probably can probably be interpreted to mean "certainly not". To justify the program it is necessary to assert that the NSA needs all phone and e-mail records to "search all the phone and e-mail records for links to numbers on a terrorist watch list." Clearly Ignatius knows perfectly well that this is nonsense. If there is a terrorist watch list, the NSA would need the records of calls to and from numbers on that list. The idea that it would need to have all records on in ft Meade in order to find calls too and from those numbers is totally absurd.

That hack (sorry I can't name him, since I don't have the DC phone book with so I can't search for people named David Ignatius so I don't have the name) must know that if you want records of the calling activity of a number on a hot list you do not need to take records of calls too and from all numbers to then search for calls to and from that number. That's crazy. Ignatius is trying to convince us that the NSA but not the phone companies can generate a reverse phone bill with calls to a number (notice I live in Italy and can conceive of the idea of local calls appearing on a phone bill).

If there is an arguably valid "terrorist watch list" the NSA could easily obtain all calls to and from numbers on that list with a FISA warrent. Recall that by )/11/01 the FISA court had never ever denied a requested warrent.

Ignatius tries another example (which he got straight from Gen Hayden)

To explain the basic concept of pattern analysis, Hayden has told audiences that if you could monitor, say, the timing and pattern of calls on Super Bowl Sunday, you would know which teams were playing, how the game progressed and perhaps even who won
. OK see we have a sample of super bowls. The NSA has a large sample of calls made during super bowls involving a known team and victor. The super bowl is about the most public thing in the USA. Al Qaeda sleeper cells are quite different. The example illustrates again that with tons of data including zero data on the group of interest you have, in effect, zero data.

It is amazing really that people argue about the NSA program given that it is impossible to present a plausible hypothetical example in which the program could be useful and legal means would not work just as well. This is not at all a usual phenonmenon in policy debates. The illegal NSA program might be the most obviously worthless waste of public money on record.

No comments: