Analysis: Are 50 Foiled Terrorist Plots Worth Your Privacy?

In the most candid explanation of the National Security Agency's surveillance program to date, agency head Gen. Keith Alexander said Tuesday that his organization's listening activity has helped foil more than 50 terrorist plots against the United States and its allies. One of those involved Najibullah Zazi's attempt to blow up the New York City subway; another concerned an early-stage plan, news of which was previously withheld from the public, to blow up the New York Stock Exchange.

Alexander and other witnesses before the House Intelligence Committee made sure to highlight key details of these foiled attacks. Understandably so: The more we focus on the program's successes, the less harshly we might be inclined to judge its alleged excesses. But what exactly is the tradeoff being made here, and how do these revelations address concerns about the potential for NSA over-spying?

We now know at least this much: Of the millions of phone numbers that the NSA could summon for intelligence purposes under Section 215 of the Patriot Act, about 300 specific numbers were actually "queried" from the database in the course of federal investigations. (This database is made up of the same telephone metadata Verizon has been handing over to the government on a daily basis since 2006, the subject of NSA leaker Edward Snowden's initial bombshell.) Twenty-two officials are responsible for approving these queries within the NSA, and the agency doesn't require special court authorization to inspect the phone numbers in question.

The NSA says it takes steps to rule out accidental snooping on U.S. citizens. Among the most important? Inspecting the area code, said John Inglis, the NSA's deputy director. If the number begins with "301," for example, you'd know it was a Maryland number.

"That would be your only insight as to whether this would be attributable to a U.S. person," Inglis said, adding that if the person was simply exercising their First Amendment right to free speech, "that is not a reason to approve a query." 

Inglis didn't elaborate on how the NSA would go about determining whether the subject was in fact exercising free-speech rights (Would they send an agent to interview the suspect? Tap his lines?), but he assured lawmakers that the phone-records program had "a very narrow purpose."

(It's not entirely clear, but these minimization procedures seem to be separate from those applied to PRISM, the NSA surveillance program that summons data from tech companies under Section 702 of the Foreign Intelligence Surveillance Act.)

Yet despite Inglis's insistence that from this database the NSA can only retrieve phone numbers—and that "queries" approved by the 22 agency officials would return only those phone numbers that the first number dialed—statements by other officials at the hearing seemed to suggest that connecting metadata to specific individuals was a trivial matter.

Take one of the four cases the officials highlighted as plots disrupted by either the 215 program (the phone records) or the 702 program (the business records). Using the 215 program, said FBI Deputy Director Sean Joyce, law-enforcement officials were able to positively identify an individual who was providing financial assistance to a listed terrorist group in Somalia. Not only that, but the FBI was also able to identify that suspect's coconspirators.

We already know that metadata acts as a kind of digital fingerprint, and it takes little more than a couple of data points to connect anonymized metadata to a specific person.

Even President Obama hinted at how flimsy the "it's only metadata" argument is in his interview with Charlie Rose last night:

"[Critics will] say, you know, 'You can—when you start looking at metadata, even if you don’t know the names, you can match it up, if there’s a call to an oncologist, and there’s a call to a lawyer, and—you can pair that up and figure out maybe this person’s dying, and they’re writing their will, and you can yield all this information.' All of that is true. Except for the fact that for the government, under the program right now, to do that, it would be illegal."

That any abuse of the system would be treated after the fact as a crime doesn't do anything to assuage Americans worrying that the crime is possible in the first place. It's also not outrageous to say, as my colleague Conor Friedersdorf does, that the tradeoff we've made between liberty and security is out of balance, and that maybe we've let our fear of terrorism get the better of us.