"Reported Drop in Surveillance Spurred a Law":
Today's New York Times has this very interesting story about why the Democratic leadership in Congress recently agreed to amend FISA.
Features
Stuff from us
Academic Legal Writing: personalized bookplates
Sources on the Second Amendment
But I am dumbfounded that today's backgrounder and sidebar never explain what the bill Congress passed actually does. In fact, reading them together, one would have the false impression that all the legislation does is legalize warrantless surveillance of foreign-to-foreign traffic passing through a U.S. switch. Actually, it also legalizes such surveillance of foreign-to-domestic traffic, which is what the controversy is all about.
Risen and Licthblau, winners of the Pultizer for originally disclosing the NSA warrantless surveillance program in 2005, have not performed well at all covering the legal and legislative aftermath, IMHO.
I, for one, despite being a supporter of the government here, would not have opposed some more Congressional oversight. There was some more oversight, but not as much as even I would have liked, as Congress had painted themselves into a corner and had to act before recessing.
I agree that the "directed" language is significant, and perhaps I should have mentioned it. If I had been writing more than one paragraph to summarize the issues for the cognicenti, here. I certainly would have mentioned it.
My criticism of the NYT reporters is that they wrote a long story and a sidebar purporting to explain the legislation in context, and they ignored the heart of the controversy completely and mischaracterized the effect of the act. The intelligence briefings apparently centered on foreign-to-foreign communications, and so did the political arguments for the bill. But the actual legislation is much broader.
Now, as for the meaning of the "directed" language, that is at best unknown. The key new section 105A certainly does not make clear that it requires an identified "target." The section easily can be read not to require that at all, but only a reasonable belief that some person abroad is the likely sender or receiver of a surveilled communication.
.
some links to administration pronouncements
.
Don't expect the NYT (or anybody else) to do your homework for you.
.
"Targeted" implies a degree of specific knowledge, usually the connection between a communication and a particular person. E.g., we want to listen in on Osama bin Laden conversations.
.
"Targeted" could be less specific than that, such as when we know a voice, or an anonymous handle on an e-mail, etc. -- but just the same, "targeted" implies a degree of tight focus.
.
"Directed at," on the other hand, can be a location or an entity larger than an individual person.
The point of all that is that I don't assign legal equivalence between "surveillance of a target person reasonably believed to be outside the US" and "surveillance directed at a person reasonably believed to be located outside the US."
.
In the formula passed into law, the individual person outside the US whose communications have been acquired could be a target, but isn't necessarily one.
I have posed the same hypothetical about the dedicated international switch, which is not fanciful because it resembles what is speculated to have been involved in the so-called "TSP" surveillance.
I think the "reasonably believed to be located outside of the United States" part is easily satisfied in that scenario, and I think an NSA officer in that situation could swear to that. (Maybe he also would filter to ensure that the phone number or IP address is in a foreign domain, which of course could also be done at a non-dedicated switch.)
To play devil's advocate, that same NSA officer would be making an assumption that the foreign sender/receiver was a "person," rather than a farm animal, potted plant or space alien.
.
When you are in the tree stand, your surveillance is directed at deer.
.
When a deer appears, that deer may (or may not) become a target.
I'd think there would be some amount of uncertainty involved in every parameter used to select traffic for the tap; that said, I'd also think that a significant factor would be quickly cutting down the amount of clearly irrelevant traffic they have to put in front of human analysts.
But if the traffic selected for human inspection has to pass through a particular physical location, have an ip addresses from specific networks in the ip packet header, and mention specific email addresses in the smtp envelope or message headers, I'd think that would be specific enough to be reasonably certain that the communications were to or from a target of surveillance.
But I'd have nearly the same degree of certainty about the accuracy of the filter if the tap were placed between any two routers in an ISP backbone.
.
There is a physical layer. There are fibers that cross borders, and some of those border crossing transit oceans.
.
Just give me all of that. It's surveillance directed at a person reasonably believed to be located outside of the United States.
.
-- I'd also think that a significant factor would be quickly cutting down the amount of clearly irrelevant traffic they have to put in front of human analysts. --
.
Hence the secrecy in sorting and decoding capability. Just pass over the raw communications flow, if you please.
I think the real problem would be creating good algorithms to find what you are looking for, because of the quantity of data. Eventually a human would have to review what was captured too.
Amazing.
I despair that it will take another 9/11 (or maybe a nuke) to wake these people up.
.
When the administration submitted its TSP to FISA, it knew, in advance, that such submission could impinge on foreign-to-foreign interception.
.
Perhaps the administration submitted TSP to FISA as a means to deliberately lose intelligence, so as to create a problem where the solution is to modify FISA.
.
So, yes, it might be technologically almost feasible to tap some traffic on the bottom of the ocean, but the only realistic way to tap enough of it to make it worthwhile is to do so at the switches in the U.S. There, the signals have been turned from optical back into electrical.
Another problem with at least fiber optics, is that a lot of the signaling information is out-of-band, meaning that it is not sent with the conversation, but either in another time slice of the TDM slices at a certain frequency on a certain fiber, or at a different frequency (or color), or indeed, on another fiber entirely. This is the information about who is calling whom, etc. And it is only tied together with the corresponding conversations at the switches. So, tapping a single fiber, or likely even a fiber cable, out in the middle of the ocean, might be an interesting academic exercise, but is unlikely to get the NSA the information they need.
As noted above, the Internet is even more problematic. Email and pretty much everything else, is packetized, and the packets are sent over different routes, depending on any number of factors. Only the email header has the real source and destination information (and that may be in more than one packet). And it is fairly easy to route stuff through relay sites that can be all over the world. That is something that most of us don't bother with, but knowing that the NSA is looking for your email is likely an incentive to utilize that sort of thing. With email, it is often difficult, if not impossible, to know whether either end is really in the U.S.
The way that they described the operation of programs like the TSP, there are foreign targets out there, and the NSA is looking for traffic to or from those targets, and when communications from or to those targets is detected, the interception is automatic. In other words, a list of targeted phone numbers, and any calls from or to them are recorded, regardless of where in the world the other end is. Almost all of the traffic being recorded is purely foreign, since that is who the targets mostly deal with. But some does involved destinations in the U.S., and that is part of the problem. This was made much worse because that one FISC judge was requiring certitude that neither end was in the U.S., and often the NSA couldn't provide that. They thought it unlikely, just not impossible (and thus new language about reasonable belief).
You may also have read the same thing in this blog, in Orin's recent background interview with a senior White House official. No doubt Orin reported that interview honestly in paraphase. But even if such administration officials said all that on the record from the White House podium, it would not be anything like statutory language binding the government.
There is simply nothing in the statute equating "surveillance directed at" with "acquisition ... of the contents of any wire or radio communication sent by or intended to be received by a particular, known ... person ... if the contents are acquired by intentionally targeting that ... person."
Compared to the New York Times report at the top of this blog post, Post reporters Joby Warrick and Walter Pincus
produced a more informative report of both the political events and the real controversies about the legislation.
.
But as the scope of the surveillance widens, it's less sensible to claim the object is a "target." E.g., we likely have surveillance directed at Iran.