Effective immediately, we will only pursue phone calls that are two steps removed from a number associated with a terrorist organization instead of three. And I have directed the Attorney General to work with the Foreign Intelligence Surveillance Court so that during this transition period, the database can be queried only after a judicial finding, or in a true emergency.
These promises have been taken to limit all queries to two hops (which was NSA”s practice in any case) and, except in an emergency, to require FISC to approve the Reasonable Articulable Suspicion determination an identifier before it is used to query the database.
That’s not exactly how the modification implements the change. Rather, it lays out 3 ways to access the database:
- With prior FISC review, by motion, of the RAS determination
- With an assertion of emergency from the Acting Director of NSA or DIRNSA, in which case FISC reviews it after the fact
- Using an identifier for which FISC has already found probable cause under traditional FISA
Access under the terms of the last bullet, which has actually been part of dragnet orders since the second order, is accomplished in the supplement with this language:
For any selection term that is subject to ongoing Court-authorized electronic surveillance, pursuant to 50 U.S.C. § 1805, based on this Court’s finding of probable cause to believe that the selection term is being used or is about to be used by [redacted–describes a tie to a foreign terrorist organization], including those used by U.S. persons, the government may use such selection terms as “seeds” during any period of ongoing Court-authorized electronic surveillance without first seeking authorization from this Court as described herein. Except in the case of emergency, NSA will first notify the Department of Justice, National Security Division of its proposed use as a seed any selection term subject to Court-authorized electronic surveillance.
Now, with one minor caveat, I actually have no problem with this. As I said in this post, it makes sense that NSA should have access to the metadata of calls it already has access to content of. And this third access still complies with the language of Obama’s promise: rather than a judicial finding regarding RAS, such queries would have been justified by a judicial finding regarding probable cause, a much higher standard.
I’m mostly interested in this detail for what it might suggest about the way the NSA is currently using the dragnet. I have repeatedly focused on Theresa Shea’s description of how NSA uses the dragnet to prioritize which content they read.
Section 215 bulk telephony metadata complements other counterterrorist-related collection sources by serving as a significant enabler for NSA intelligence analysis. It assists the NSA in applying limited linguistic resources available to the counterterrorism mission against links that have the highest probability of connection to terrorist targets. Put another way, while Section 215 does not contain content, analysis of the Section 215 metadata can help the NSA prioritize for content analysis communications of non-U.S. persons which it acquires under other authorities. Such persons are of heightened interest if they are in a communication network with persons located in the U.S. Thus, Section 215 metadata can provide the means for steering and applying content analysis so that the U.S. Government gains the best possible understanding of terrorist target actions and intentions.
If this is primarily how the dragnet is currently being used — to tell NSA which call content that it has collected it should listen to or translate first — then it would explain why the FISC didn’t complain about having to approve a bunch of new query identifiers: because it wouldn’t really have to do much pre-approval beyond the traditional FISC warrant review it has already done.
And given that NSA ran queries on 288 identifiers in 2012, a year when FISC approved 1,788 FISA warrants (though some were for physical searches), it is feasible that many or even most of the dragnet queries were tied to FISC warrant targets.
If that’s right, it suggests the dragnet no longer serves primarily as the alert function it has been sold as, but instead serves an indexing function (which is, after all, what James Clapper said months ago).
So here’s my one caveat to my assertion that I have no problem with this.
In making this modification, DOJ actually changed the way they refer to what FISC-approved targets automatically qualify as RAS-approved. In the order itself, it describes it this way:
Selection terms that are currently the subject of electronic surveillance authorized by the Foreign Intelligence Surveillance Court (FISC) based on the FISC’s finding of probable cause to believe that they are used by [redacted description of tie to terrorism] including those used by U.S. persons, may be deemed approved for querying for the period of FISC-authorized electronic surveillance without review and approval by a designated approving official. The preceding sentence shall not apply to selection terms under surveillance pursuant to any certification of the Director of National Intelligence and the Attorney General pursuant to Section 702 of FISA, as added by the FISA Amendments Act of 2008, or pursuant to an Order of the FISC issued under Section 703 or Section 704 of FISA, as added by the FISA Amendments Act of 2008.
I think this works out to be a distinction without a difference, or even an improvement. The language of the order says targets of FISA orders — except those targeted under Section 702 (bulk collection targeted at foreigners outside the US), Sections 703 and 704 (US person target outside the US) — are pre-approved as dragnet identifiers. The language of the modification says targets only of traditional FISA orders (authorizing electronic surveillance of either US persons or foreign individuals in the US) are pre-approved for dragnet identifiers. If anything, the modification language is more narrow, as it would also exclude those against whom FISC has approved physical search warrants from automatic RAS approval. If this reading is correct, it would seem to support my supposition that the dragnet is increasingly serving primarily as an index to already-collected content.
But given the way they’ve expanded the intent of traditional FISA in the past, I do wonder whether something else is going on.
All that said, I mostly intend with this post to point to yet more evidence suggesting that the dragnet increasingly serves as an index rather than the early warning system it gets billed as.