In the furor over revelations that the U.S. National Security Agency has been tracking phone calls, social media, and other Internet metadata, the question remained whether this monitoring actually keeps Americans safer.
Today, the head of the NSA told the Senate Appropriations Committee that data from the program broke up “dozens” of plots aimed at the U.S. and its interests abroad. “It’s dozens of terrorist events that these have helped prevent, both here and abroad, in disrupting, or contributing to the disruption of terrorist attacks,” said Gen. Keith Alexander, who promised to get a more exact figure out later this week.
He did cite two specific instances, that of an Afghan-American planning a suicide attack in New York and of the American who helped scope out Mumbai before the 2008 attacks there, in which the data contributed to unraveling the cases.
The Guardian newspaper, citing an NSA memo, reported that as of March some 97 billion pieces of data had been collected by Prism, a mighty haystack indeed.
That role for the intelligence—useful, but after it’s been directed where to probe—fits in with the findings of terrorism expert Martha Crenshaw. At the behest of the Department of Homeland Security, the senior fellow at Stanford University’s Center for International Security and Cooperation has been plowing through data on the last two decades of terrorism tries in the U.S. She’s found “not that many”—if any—of those attempts were discovered through blanket electronic surveillance. Crenshaw talked about her work over coffee this morning in California even as Alexander was testifying in Washington.
However, Crenshaw said, having a ready-made treasure trove of Internet-derived data—terrorism masterminds don’t generally recruit via the Internet, she says, but they can be chatty—to delve through does help in cases where authorities already have a lead or a tip. It creates the sort of “Google for police” that Torin Monahan described yesterday for our Marc Herman.
Crenshaw wears the Homeric of “terrorism expert” reservedly. For the record, the political scientist has been elbows deep in terrorism her whole academic career, starting in 1973 with her Ph.D. dissertation on the National Liberation Front, or FLN, in Algeria’s war of independence. Her lonely specialty focusing on terrorism and its psychology—and her belief, alongside folks like Ashton Carter and William Perry, that terrorism might be a genuine threat to the U.S.—welcomed a host of compatriots after Osama bin Laden’s attacks in New York City. “I’m not someone who jumped in as a terrorism expert after 9/11,” she now says with a smile.
In an unofficial sequel to her heralded Mapping Militants project graphing terrorism’s “family trees,” she’s in the midst of looking at “jihadist plots” against the U.S., Canada, Australia, and Western Europe since Ramzi Yousef’s 1993 attempt to blow up the World Trade Center. Unlike most efforts to collect and categorize terrorist actions, the Failed and Foiled project collects information on terrorist plots that did not come off as well as those that did. This catalog of flops includes plans foiled by authorities as well as those that failed because of chance or ineptitude (like the Times Square car bomb) or that were called off by their putative perpetrators (and there are a few of those, too). Coding for failure, as other researchers have noted, “avoids systematically omitting behavior on the basis of its success” which might otherwise color counterterrorism efforts.
The results will complement the exemplary Global Terrorism Database hosted by the National Consortium for the Study of Terrorism and Responses to Terrorism at the University of Maryland.
Noting that even with federal funding her Failed and Foiled examination doesn’t get to troll through top secret stuff—and until contractor Edward Snowden leaked its existence NSA’s Prism was definitely hush-hush—Crenshaw said that public sources covering terrorist attacks and attempts do not show that electronic monitoring, a la Prism, has initiated any successful intercessions in the West. But she added that her understanding suggests that’s not automatically the main selling point of data mining anyway.
“They’re just there to scrape data and then store it. When you get a name, then you can go through the data,” she said. “You’ve got to have some sort of idea of what you’re looking for as a practical means of dealing with this amount of material.” The Guardian newspaper, citing an NSA memo, reported that as of March some 97 billion pieces of data had been collected by Prism, a mighty haystack indeed.
So the hope that a serendipitous needle might pop out is unlikely; “we don’t know which is the signal and which is the noise.” But given some direction, the haystack by its very volume becomes the sleuth’s friend. Crenshaw notes the experience of MI-5 after the 7/7 London train bombings in 2005. Investigators with Britain’s MI-5 was familiar with some of the suspects before the attack, but given the number of potential troublemakers they had to follow, decided these were people on the periphery of any threat and didn’t pursue them. But after the attacks the investigation was aided by having these same people on their “surveillance radar.”
Reviewing some of Failed and Foiled’s preliminary determinations, Crenshaw said it’s identified roughly 100 jihadist-type terrorist plots within the borders of the United States since 1993. Three—the World Trade Center bombing, the 9/11 attacks, and the Boston Marathon bombing—drew blood, and she estimated roughly 10 percent of the hundred reached a threat level she considered “really serious.”
In a large number, an informer had been part of the cabal from the start, so much so that in many cases, “There was no plot until there was an informer,” she said. In others, the hopeful terrorists made stupid missteps like disclosing their intent on Facebook—one for Prism’s ‘win’ column—or were “sad sacks” whose ambitions outstripped their intelligence.
Might this NSA surveillance program alone prevent a future 9/11? “I don’t know that that’s its use at all,” Crenshaw said. But that by no means suggests it’s of no use at all.
