In order to justify its striking venture into mass domestic surveillance, the executive branch had come up with a radically new—and entirely secret—interpretation of law. Back in 2001, Section 215 of the USA Patriot Act gave the FBI simplified access to business records that the FISA Court found “relevant to” an authorized investigation of terrorism. Relevance is a legal standard that typically limits the scope of the government’s power. It covers this, not that. It distinguishes what investigators may take from what they may not. Under Presidents Bush and Obama, the Justice Department turned that concept on its head. Government lawyers secretly persuaded the FISA Court that every record of every call met the relevance test because a terrorist plot might involve a party or parties unknown. The NSA proposed to find those ghosts by way of “contact chaining,” a mathematical analysis of links among friends, friends of friends, and so on. The computational methods had implications beyond the competence of the court to assess. (FISA judges employed no technical advisers.) What the court knew, and chose to authorize, was that the NSA wanted access to the whole universe of domestic telephone calls.
By design, the intelligence court met in classified session. It heard no opposing views. Not only the substance but the existence of its ruling remained unknown to the public and to most members of Congress, including a principal drafter of the Patriot Act. The court’s order was classified TOP SECRET//SI//NOFORN, a level that left the great majority of legislators without a staff member cleared to read the complex material and advise them. Not all these points made it into the Guardian’s account, but it was a world-reverberating story and Greenwald told it well.
Now, knowing the race was on, I told Shawn Turner that the Post was speeding up its timetable on the PRISM story. When my telephone rang next, I heard the voice of Robert Litt for the first time in two years. As general counsel to Director of National Intelligence James R. Clapper Jr., Litt was the government’s senior intelligence lawyer. He sounded tense, but he began with a pleasantry. We had met before, he said, at the Wye River Conference Center in Maryland. He was not sure I remembered. I could hardly have forgotten. We had spent two intense days in a conference room in the spring of 2011, when the Aspen Institute convened a small group of journalists and intelligence officials. It was supposed to be a dialogue, as it happened, about news coverage of national security secrets. There were twenty-eight men and women at the table, including a future FBI director and a former deputy director of national intelligence. No two of us, I think, disagreed more profoundly than Litt and I.
Litt told me he had the PRISM presentation in hand. There were others around his speakerphone, but he left them unnamed. It became clear that they meant to hold this conversation right here on an open phone line, which astounded me. Litt and Clapper, I later learned, had been summoned to Capitol Hill. There was no time to arrange a meeting face-to-face.
Litt and his unnamed colleagues had to be worried, I knew. I am no worshipper of secret stamps. I have seen them on reprints of stories published under my byline. But there were portions of the PRISM presentation that just about anyone would agree had earned their Top Secret markings. Some of the slides included jaw-dropping excerpts of intercepted communications and files. There were foreign nationals caught unawares as they laid plans to kill Americans, construct weapons forbidden by treaty, and break into classified U.S. contractor files. PRISM collection had discovered a dangerous cover-up by one unfriendly government and secret equipment that another had positioned to avoid the gaze of U.S. spy satellites. No one who believed that espionage had a valid role in national defense would be likely to dispute the importance of those finds. If I mentioned them, the targets would no doubt disappear.
“I’m not generally given to hyperbole, but this one really raises the hair on the back of my neck,” Litt said.
“I have some idea of what may be on your mind,” I told him. “I think I can save us some time.”
There were things in the document that the Post had no intention to print. I did not name them on the phone, and I will not name them here. Instead I walked through the details by page number and position. Our story would publish nothing from slides 14, 19, 21, or 22. Nothing, for sure, from the map on slide 23 or the transcript in the speaker’s notes or the hyperlinked video file. Nothing from the list on slide 28, because one line would imply another and the whole stack would be compromised. These were not decisions we made for fear of lost favor or retribution from the government. We made our own judgments, and these were not hard ones. Speaking for myself, I am not agnostic about my loyalties. I am not—in this context—a global citizen, indifferent to the outcomes of national conflict. Disclosure of the details I allude to here would have caused self-evident harm to my country and some of its allies. We saw no countervailing interest in public policy debate. These would have been valid intelligence targets by any legal standard.
I could hear the mute button engage on the other end. After a while Litt came back onto the line and said, “We’re very glad to hear that.”
The conversation went south pretty fast after that. The Post did not want to blow operational details, but we strongly believed there was a story here. The sheer volume of collection, the impact of incidental collection, and the secret expansion of legal boundaries were important news. I told Litt I intended to name the nine companies and cite Rick’s description of direct access to their servers.
Litt is a vigorous advocate. He pushed back on everything. The U.S. government, he said, did not condone public disclosure of any classified information. It reserved all legal options in response. I have no doubt he meant it, but this part was pro forma. He knew that ship had sailed. I was looking for confirmation and context, to be sure I understood what I had read. His whole reason for speaking in detail was to distinguish which of the classified facts the government believed most important to protect. Litt did not dispute Rick’s description of data pulled “directly from the servers” of Silicon Valley, but it later emerged that he had not focused on that. “Obviously we’d much prefer that nothing be written about this at all,” he said. What he really cared about was outing the company names.
“Why?” I asked.
Public exposure, Litt replied, could cause them embarrassment and leave them reluctant to cooperate in the future.
“If the harm that you’re asserting consists of damage to a company brand because the public does not like what it’s doing, then I can’t accept that as a reason to hold back,” I said. “The same goes for damage to intelligence collection if voters decide they want to scale it back. That’s actually why we publish news, to let readers decide what they do and don’t support.”
I expected Litt to take the fight over my head. I gave him Marty Baron’s private telephone number, with the obligatory joke that he must have it already. Litt did not call Baron, and neither did his boss. We moved the PRISM story that evening, June 6. The Guardian’s version, which included none of the slides or security details we left out, followed less than an hour later. Soon after that, fierce objections began to pour in from Silicon Valley. Companies that had declined to comment in advance, or had said nothing of substance, now issued categorical denials that any U.S. agency had “direct access” to their servers. I scrambled to reconcile those statements with the NSA program manager’s explicit words—repeated twice—in the authoritative PRISM overview. Later that night I found a clue in another document from the Snowden archive. There, in a description of a precursor to PRISM, I found a variation on Rick’s formula. “For Internet content selectors, collection managers sent content tasking instructions directly to equipment installed at company-controlled locations,” it said. That sounded as though the U.S. government black box was on company property but might not touch the servers themselves. I updated my story to disclose the conflicting information and the new evidence.
Direct access, which seemed to be a clear enough term, turned out to be unexpectedly imprecise. It meant v
ery different things in Silicon Valley and Fort Meade. The technology companies understood the term to signify that government spy equipment connected physically into their core hardware. That would represent a devastating loss of control, and in the end they were right to deny it. The NSA’s internal account spoke of collection “directly from the servers” in counterpoint to what the agency calls indirect collection, which is done by passive interception of data on the fly across the internet. As Rick explained to the analysts, PRISM did not have to chase photons across fiber optic cables. It acquired data straight from the source in the nine companies’ data stores. From the point of view of an NSA analyst, the connection was quite direct: she tasked a selector and buckets of data came back. From the points of view of Facebook or Google, the firebreak between government equipment and their own was of cosmic import. “The actual architecture is very different, but the material effect, with some latency because of review, is as if it were direct,” Chris Inglis, the NSA deputy director, explained to me later.
In retrospect, I do not love the way I wrote the story. I knew a lot less then than I learned later, with more time in the documents and many more interviews. The “direct access” question became a big distraction, rightly essential to the companies but not so much to the core questions of public policy. And I completely missed a story hiding in plain sight in the PRISM briefing. The FISA amendments had authorized two new forms of lawful collection, which operated in parallel. PRISM was one. The second, which I neglected, was called Upstream. That one became more important in the months to come.
Three days after the PRISM story went live, Snowden announced his identity to the world in a twelve-minute video that Laura Poitras shot for the Guardian. Danielle Massarini was killing time in a Newark Liberty International Airport bar on her way to Germany. “I’m scrolling through my Twitter feed and keep seeing ‘Snowden’ ‘Snowden’ ‘Snowden,’” she recalled. Why was that name so familiar? Then Snowden’s face appeared on a screen above the bar and Massarini screamed aloud, “Oh, shit!” The operations officer of her new unit, the Army Foreign Counterintelligence Activity, looked at her quizzically. “I hired that guy to teach cyber security” at the Defense Department’s counterintelligence academy, she said. “Better call the the boss,” the ops officer responded. The FBI would soon come calling on everyone who had ever crossed paths with Snowden.
FIVE
BACKLASH
We’re not going to open this program up to the sort of scrutiny that the zealots want.
—Admiral Dennis Blair
I scanned for a promising table, steering clear of sources who would not welcome me nearby. A long line of defense and intelligence leaders wound through the buffet. When a seat opened up near Admiral William McRaven, I set my tray down two places to his right. This kind of casual contact, at mealtimes and downtimes, brought me back every year to the Aspen Security Forum. If there was such a thing as a national security establishment, it summered here under Rocky Mountain skies for four days in July. Deputy secretaries and undersecretaries, intelligence chiefs, combatant commanders, defense contractors, presidential advisers, and members of Congress gathered to float ideas and chew on the state of the world. A target-rich environment for a reporter like me. It had been, anyway. When we sat down for lunch on July 18, 2013, the Snowden fires had been burning for six weeks.
McRaven and I were acquainted. We had spoken at some length two years before, not long after he took the helm of U.S. Special Operations Command.
“Good to see you,” I said.
Silence. His face grew taut.
“Admiral?”
He would not look at me. McRaven cut a formidable figure, an angular six foot two, with the eagle and trident of a Navy SEAL on his dress blue uniform coat. Twelve years before, he had clawed his way back to fighting shape after shattering his pelvis in a disastrous training jump. A midair collision with a fellow SEAL, after two miles of free fall, had left him barely conscious as time ran out to deploy his parachute. When he popped the canopy, the risers tangled around his legs and “split me like a nutcracker,” he told me in 2011, the year he commanded the operation that killed Osama bin Laden in Abbottabad, Pakistan.
Up onstage, the NSA general counsel, Rajesh De, skirmished with Anthony Romero, executive director of the American Civil Liberties Union. Snowden loomed large in their conversation about “the tension between a free press and national security.” McRaven glared at his plate. He pumped a leg under the linen tablecloth. From time to time he expelled a breath, more growl than sigh. I reached for one of those little notepads they hand out at conferences, wrote a few words, and slid the pad to my left.
“You clearly have something on your mind. Can we talk about it?” the note said.
McRaven glanced down, shook his head, and pushed back his chair with a little more force than required. He took a couple of long steps toward the exit. Then he stopped, rounded back, and thrust a forefinger at me.
“We didn’t have another 9/11,” he said, cords of muscle bunching on his neck. “Until you’ve got to pull the trigger, until you’ve had to bury your people, you don’t have a clue.” He flicked a backhand toward Romero onstage. “None of you. None of these people. None of you has a clue.” A low murmur from nearby tables trailed him out the door.
As a literal matter McRaven was surely right. My encounters with danger in the field—looking down muzzles at checkpoints, watching a firefight from behind a wall—had not exposed me to anything like combat. (I speak for myself. Some of my colleagues encountered far worse, and not all survived.) I had never held another person’s life in my hands or death on my conscience. Even so, the heat of McRaven’s anger caught me off guard. Warriors and civilians are tribes apart. There was nothing new in that. Nor was McRaven known for losing his cool. A witness told me after the bin Laden raid that McRaven “was the voice of Walter Cronkite” as he live-briefed President Obama, reporting with equal sangfroid when a helicopter crashed and when the al Qaeda leader’s death was confirmed.
Once upon a time McRaven had majored in journalism at the University of Texas. That emphatically did not mean he was prepared to look at national security secrets through a reporter’s eyes. Years before our encounter, as he neared midcareer, McRaven had written a treatise on special warfare that gave paramount place to stealth: concealment, surprise, clandestine surveillance of a foe. When he spoke of the raid in Abbottabad, he deflected credit to the CIA, NSA, and National Geospatial-Intelligence Agency. “This will go down in history as one of the greatest intelligence operations ever,” McRaven told me then. He meant it.
McRaven did not linger long enough at lunch to explain his reference to another 9/11. I took him to mean that electronic surveillance, the kind that Snowden exposed, had fueled the hunt for bin Laden and his lieutenants. NSA’s expansive reach had forced al Qaeda leaders underground, depriving them of the command and control to organize another complex attack. Intelligence had led the SEALs to bin Laden’s gate, and intelligence gave an edge to commando teams in thousands of lightning assaults against superior numbers. If sources and methods had been lost at Snowden’s hands, McRaven must regard that as sabotage. From that point of view, I would qualify as a saboteur myself.
I had never lived in McRaven’s world, where secrets sometimes carried mortal stakes. That did not exhaust the conversation I wish we could have had. McRaven had not lived in my world either. In my world, secrets sometimes gave cover to lies. Secrets sometimes shed darkness on conduct that was hard to defend in the light. Secrets sometimes built a mirage of consent to choices the public never knew it had. The dilemmas of self-government and self-defense had absorbed me since graduate school, where my thesis on wartime secrecy covered some of the same terrain as McRaven’s. We served distinct and equally vital interests, he and I, in a democracy at war.
Years later, I reached out to McRaven again. He had retired from active service by then and taken the post o
f chancellor of the University of Texas, his alma mater. I thought I knew what had set him off at Aspen, I wrote, “but I would much rather hear it from you than offer surmise” in my book. Sometimes long shots pay off. “I’m happy to chat with you,” McRaven replied by email an hour later. The following day he telephoned from his fourth-floor office in downtown Austin, where a John Wayne coffee mug and a Minuteman soldier’s figurine stand sentry by his desk. He could not recall exactly “what precipitated my outburst,” but he apologized graciously. “My issue has always been with the safety and security of Americans in harm’s way,” he said. “I’m sure it’s a good reporter’s concern too. . . . How do you balance what you think the public needs to know with the potential to put lives at risk?”
He wanted to find some common ground. He did not come up with much. “I’m a big believer in transparency, so please do quote me on that,” McRaven said. “And the processes are out there that allow the transparency to occur at the right level.” The right level, as he meant it, did not fall within the public domain. Quite the reverse. McRaven believed in transparency inside the walled precincts of the FISA Court and the House and Senate intelligence committees. The public had no need to know or contribute outside views on policy or law. Classified transparency, in other words. McRaven saw no contradiction in that. This model, the prevailing one among McRaven’s peers, extended far beyond surveillance policy. How many noncombatants died in special operations raids? Did the rules of engagement conform to American values or international law? Should U.S. drones be allowed to make autonomous decisions of life and death? All those things were classified, exempt from debate.
Dark Mirror Page 16