Earlier this year, a man ran into a Paris police station wearing a suicide belt, clutching a meat cleaver, and yelling “God is great” is Arabic. It was Jan. 7, at 11:30 in the morning — the first anniversary of the Charlie Hebdo attacks, down to the minute — and President Francois Hollande was speaking at a memorial ceremony nearby. The attacker was immediately shot dead. Bomb technicians later determined that the explosives attached to his belt were fake, but a paper proclaiming the man’s allegiance to ISIS was found in his pocket.
“Who are these people, ruining all of our lives?” said a Parisian woman to a news reporter later that day. It was — and is — a legitimate question.
For a decade, the American oracle for such security questions was Michael Hayden, the former director of the CIA and NSA. In a new memoir, and recent New York Times op-ed, he lays out a vigorous defense of his tenure, focusing on drone warfare. “To Keep America Safe,” the headline implores, we must embrace this secret program. But this over-promise can only under-deliver; the threat to the West has moved past one so easily solved by drones.
Hayden emphases deliberative leadership and precise weapons, quickly glossing over the specific intelligence upon which drone operations are based. And for good reason — it is all classified. Hayden’s op-ed and new book both had to be approved by the CIA’s Publication Review Board, so he is unable to provide any details that we can compare to Wikileaks or documents published by the Intercept. So Hayden’s insists we take his word for it. “The data was near encyclopedic,” he writes.
But in researching a book about the search for the bomb-makers of Iraq and Afghanistan and Syria, I discovered that our intelligence generally is more notable for its holes than its comprehensiveness. The analysts I interviewed found themselves afloat in a sea of data, yet always lacking the key bit they needed most.
Hayden’s argument tackles the intelligence problem backwards. Possessing enough evidence to justify specific strikes on terrorists already in your drone’s crosshairs is a dated and Beltway-centric problem. Identifying the specific terrorist who poses the greatest or most imminent threat, and understanding their place, if any, in a network, is the challenge of average analysts today. Hayden defends the picking of fruit within reach, but we still struggle to describe the tree as a whole. Meanwhile, apples are falling on our heads.
Who are we even fighting 15 years after 9/11? To a surprising degree, we still don’t know.
Consider the highly coordinated attacks in Paris last November. One of the three suicide bombers at the Stade de France is still unidentified. Another of the gunmen, who shot up the cafes in the 10th Arrondissement and detonated himself during a police raid five days later, was not named until Jan. 15th. Before the incident, less than half of the attackers were known to French authorities or on watch-lists.
Intelligence agencies are drowning in data. Billions of phone records, thousands of security cameras per major city. An FBI warehouse for bomb evidence that looks like the final scene of “Raiders of the Lost Ark.” Big Data advocates always want to add more. Keith Alexander, Hayden’s successor at the NSA, described this approach as “collect it all.” The solution to incomplete data is more data, and the solution to data overload is faster computers. Maybe they have a point. Yet if we still haven’t identified one of the Paris attackers, does that mean we have too much information or too little? As James Cole, former deputy attorney general, told Congress in 2013: “If you’re looking for the needle in the haystack, you have to have the entire haystack to look through.”
Databases are also only as useful as the quality of information they contain. In the case of well-known Al Qaeda bomb maker Abu Khabab al-Masri, the FBI put the wrong photo on their wanted poster for years. The two Paris attackers who crossed into Europe with refugees gave aliases and incorrect passports, slowing the identification process, and making all poor fleeing Syrians suspect.
Drones strikes add to the problem of voluminous data, they sweep up up to 400 terabytes of video per flight, yet they destroy, fingerprint, DNA, and bomb fragments. As one contractor told me, if he couldn’t positively identify the target, “I’d like to get a hand on him rather than a drone strike on him, to find out.” NBC News found that in 2010 and 2011, one quarter of those killed by such strikes were anonymously listed as “other militants.”
“We made no excuses about killing lower-ranking terrorists,” writes Hayden, completely at odds with the lessons of the surges in Iraq and Afghanistan. There, we avoided killing as many of those young men as possible because it was counter-productive, created more terrorists, not fewer. The drone program has not stopped the lone wolf attacks in San Bernardino, Fort Hood, Paris, or at the Boston Marathon, but it does inspire next one.
Drones and databases are a classic Western technical solution to a fundamentally cultural problem. Referencing that the Charlie Hebdo attackers were known needles lost in the haystack, Edward Snowden told the Guardian “When you monitor everyone, you understand nothing.”
This view comes not just from outside critics, but from within the intelligence apparatus itself. As one intelligence analyst quipped to me, “While the plural of anecdote is not data, neither is the plural of data intelligence. And we get a lot of data.” We didn’t need a drone strike to kill Abu Bakr al-Baghdadi when he was imprisoned in Camp Bucca in Iraq in 2004. We needed cultural understanding, to anticipate that concentrating so many radicalized and sympathetic men in the same compound could create ISIS and its leader.
Brian Castner is the author of “All the Ways We Kill and Die.”