Sunday, August 12, 2007

The Mysteries of New FISA Section 105B

Marty Lederman

Just a quick head's up to note that there's a very interesting discussion ongoing in our comments section [yes, it can be done!] over at my last post, concerning the possible impact of section 105B.

I've been arguing that the main problem with 105B is that it is touted as a check on the broad scope of section 105A, but in fact its safeguards are toothless. The Washington Post and others have, by contrast, suggested that 105B actually authorizes new forms of NSA searches, over and above those authorized by the very broad new section 105A. I've been doubtful about this because, as I've written previously, in order for a surveillance program to fall within section 105B in the first place, it has to consist purely of acquisitions that are not "electronic surveillance" -- i.e., something that either wasn't covered by FISA already, or that is now not covered by virtue of the new carve-out in section 105A -- in which case section 105B does not provide any additional authorization authority for stuff that would otherwise be FISA-covered.

But as commenters such as Just an Observer, Anonymous Liberal and others have noted, even if I am right, there are at least two other possible important questions raised by 105B:

First, does 105B authorize types of "acquisitions" of information other than "contents" of communications, in a way that would previously have been prohibited by statutes other than FISA, such as the Communications Act, the Stored Communications Act, and the Pen Register Act? Its broad language ("Notwithstanding any other law . . . ") might be read to suggest as much. In other words, perhaps the evisceration of FISA is the least of it.

Second, even if 105B doesn't authorize any activity that was previously prohibited, the compulsory service provisions do appear to vastly increase the ability of the NSA to enlist the service of many private parties to obtain information that once would have been difficult or impossible for NSA to acquire, thereby dramatically increasing the breadth of NSA surveillance as a practical matter. As JaO writes, section 105B might be construed (and probably will be construed by the Executive branch) as "a powerful new government tool for warrantless surveillance directed at a whole new set of platforms where compulsory process is needed. It is vastly easier to issue an administrative 'directive' under 105B than to get a court order. The recipients of such directives can be the legion of ISPs, e-mail hosts such as employers, colleges, etc., and the fragmented Internet hosting industry in general. [On this reading], 105B all by itself is national security letters on steroids, the equivalent of administrative warrants allowing direct-connect, real-time acquisition of a universe of data that never was part of the TSP." If so, then "105A and 105B are independently important and sweeping, although for different reasons. They are not mutually exclusive, and the government can employ either, for distinct sectors and technological platforms, to the government's best advantage."

These are questions well worth careful consideration. More fundamentally, the fact that no one, not even those on the relevant congressional committees, appears to truly understand whether and to what extent 105B does either of these things, and what the exact relationship is between 105A and 105B, is perhaps the biggest problem of all with this legislation: Congress quite simply does not have any sense of what it has authorized -- and therefore, of course, neither does the public. It therefore is incumbent on Congress, at a bare minimum, to make sure that any further statutory amendments are understandable and not subject to creative, unforseen and secret Executive branch interpretations.



I appreciate your quotation of my own remarks in a more visible setting. I hope the lack of other comments here does not reflect an utter lack of public interest, but my sense is that there is not much mobilized opinion or intellectual energy focused on this matter. Many comments elsewhere tend to be about the politics rather than the substance of the policy coup.

At least Anonymous Liberal has a good post.


I wish I could be more optimistic, but those interested in politics is a small subset, and the subset of those interested in policy outside of partisan political concern is smaller still.

To really bring an upbeat mood of optimism to the party, I would like to posit that since pervasive encryption is the logical answer to widespread surveillance we should not be surprised by the return of the Clipper chip or some derivative to the legislative landscape. I fully expect this to be addressed before Bush leaves office with no opposition from congress.

shhhhhhh.... You'll give them ideas for legal justifications....

No, JaO, I'm still on the case. I just didn't have anything to say.

I helped A.L. out with a skeptic at Crooks & Liars last night and again tonight. (Comments starting at #120.)
I added two points not made here. First, I quoted from a FISA provision whose first half is formally the same as 105B but that adds a second clause to close the loop it opens, in contrast to 105B. This brings out the optionality.

Second, I noted that "directed at," when used everywhere else in FISA, always contemplates more than one directee and captures them all under its limitation, again in contrast to 105A.

A.L. made an interesting Youngstown Steel point, also new.

But our words were wasted on the skeptic, who thinks we're imagining things.

There are just a few possibilities to pursue, but they all assume a pooling of our scattered observations into a single list. I'd start if asked, so long as people agree to add to it, since I can guarantee it will need work.

To be clear, my own position on the policy questions raised by 105A is not that, across-the-board, I oppose allowing surveillance interpreted to be outside of FISA. To the contrary, I favor reconstructing the same kind of balance that was struck in 1978, and I recognize that the general intent was to allow foreign surveillance abroad without regulation.

But I think we need much greater transparency in the law governing these matters. As we consider this policy issue, we ought to know what the law actually does. (In the case of the PAA, it would have been better if Congress knew what it was doing before it passed the new law. And it would still be good to know if the administration will sometimes use inherent authority implicit under section 105A as an alternative to the optional 105B procedures.)

In addition, reconstructing FISA's balance in response to new technology means reconsidering both sides of that balance.

On the one hand, Congress in 1978 effectively did exclude from FISA's scope most international communications where a U.S. person was not explicitly targeted. But that was not done very openly. Rather, it was understood by insiders to be the case only because most international traffic then was carried by satellite, and satellite interception reportedly is deemed not to occur within the United States. Technology now has shifted most international traffic to fiber-optic cables, and the best way to intercept it is at digital switches here.

On the other hand, technology itself -- not just the law -- used to make it feasible to initiate surveillance only by starting with some individual target (such as a phone line or analog channel, etc.). It now is feasible for the government effectively to capture everything digitally and then crunch it in massive databases. That is why ambiguous terms such as "directed at" are so important today. And there is a huge, public gray area of what can be done legally, and the principal safeguard is in "minimization" procedures that are not explained to us.

When terms are defined in statute, often they turn out to depend on other terms that are interpreted in secret. For example, what is the meaning of "acquisition?" On that definition depends much of what FISA means by targeting and minimization. If "acquisition" does not occur until a human analyst actually listens to a call or reads a message -- as I have read in unofficial speculation -- that can have a profound effect on where the balance is. In some ways that might protect privacy during minimization; but it would raise other questions about what can be done by automated means before "acquisition" even occurs.

Overall, when Congress considers a permanent "modernization" of FISA without a sunset, it needs to legislate with more sunshine.


I'm not for using a change in the physics of international communication as an excuse to shut down needed intelligence gathering. But as you've said, the Democrats made a proposal to strike a balance and were turned away. And if we were in a fool's paradise since 1978, it's nothing compared to what's in store.

I don't understand the technology and can't conjure a picture for words like "directed at." But I can see that FISA knows how to throw in "solely" when it wants to and that it didn't here. More importantly, I can see indicia of accountability and note their absence, and the House proposal makes it plain that the PAA is designed to turn off lights one by one, and to keep itself in effect for 18 months, through this administration, which I don't trust.

Another question to layer into the debate we won't have: as invaluable as data gathering is, there's the question of how much good it does without more. Tim Weiner's Legacy of Ashes (not read but on my list) is a record of CIA failures to change the world for want of understanding. To take just one point of his, the languages needed for vital foreign intelligence are the ones that pull GPAs down. (Even NYC's first public school to focus on Arabic and Arab culture is meeting with resistance and can't find a home.) He thinks it will take a generation just to make up for the resulting ignorance because being an intelligence professional is really a life's work though it's now treated as a rotation.

I am really surprised about the excellent job that you have made in this blog, which contains very important and amazing content about "The Mysteries of New FISA Section 105B", it allows everyone to express a variety of opinions, this web site is pretty cool like generic viagra.

Post a Comment