The Watchers (25 page)

Hayden knew that FISA warrants weren't handed out freely. They required thick sets of paperwork accompanied by sworn statements from lawyers at the Justice Department. And if said officials thought that a surveillance request didn't meet the muster of the law, they would never argue for it in front of a judge. To do so imperiled their career, maybe even their ability to practice law.

But most of the NSA's activities weren't covered by FISA. The agency's surveillances were directed overseas and usually against foreigners. It wasn't unusual for the NSA to scoop up the communications of some protected Americans in the course of monitoring foreign targets, but when that happened the agency replaced their names in its reports with anonymous markers such as “U.S. Person 1.” This privacy protection, known as minimization, kept the NSA from incriminating presumably innocent people. And as long as the NSA was fighting foreign wars, that worked just fine.

Congress and the intelligence community both understood FISA. But their interpretations differed markedly. Some of the authors of Reagan's executive order saw it as a marker against Congress after it passed the surveillance law. They saw FISA as an unjust limitation on the president's
inherent
authority to collect foreign intelligence. The lawyers at the NSA saw 12333, not FISA, as their agency's marching order. (After all, the NSA didn't routinely spy on Americans, and so the law bore less day-to-day relevance to their operations.)

As Pelosi and her colleagues listened to Hayden talk about the NSA's new counterterrorism activities it was unclear whether what Hayden authorized after 9/11 was permissible, or whether it fell into some murky gap between a law and an order. Pelosi was concerned that Hayden had crossed a line. Whether that was true would not be resolved during his appearance on the Hill.

After the briefing wrapped up, Pelosi thought about what Hayden had said. She still wasn't sure she understood, or agreed, that he had the authorities to expand his agency's mission. She told her staff to do some digging.

What little they could pry out of the administration, which had clamped down on the flow of national security information to Congress, didn't satisfy Pelosi. She wasn't sure if Hayden needed, or had obtained, specific authorization from the current president to conduct these new activities. Hayden had a difficult case to make
,
she thought.

Hayden seemed to agree. Not long after the attacks, George Tenet made the rounds to the various intelligence agency chiefs, and he asked Hayden a question: “Is there anything more you can do?”

“Not within my current authorities,” Hayden replied.

 

 

 

Hayden had in mind a far more aggressive role for his agency, and one that mirrored the plan Poindexter was hatching at precisely the same time. Hayden wanted to build an early-warning system for terrorist attacks.

He knew as well as Poindexter, as well as anyone, that the keys to terrorist planning lay in their electronic transmissions. So it was before 9/11; so it was before the Beirut bombing. And he knew that law, policy, and culture had impeded his efforts in the past. But technology was not one of his big problems. In fact, it could save him.

Well before the attacks, Hayden understood his agency was still collecting intelligence with a cold war mind-set. So he tasked a team of senior managers, including Wertheimer and another NSA lifer named Maureen Baginski, to reshape signals intelligence for the digital age. They put a premium on analysis, and on targeting the right sources of intelligence, instead of on just vacuuming up as much information as the agency's computers could stomach. Baginski liked to call it “hunt, not gather,” and that mantra was reflected in the agency Hayden was trying to build.

By the time terrorists invaded America, the NSA had come a long way. In October 2001, when CIA and military forces deployed to Afghanistan to hunt Al Qaeda leaders, Wertheimer and Baginski had so improved the agency's global tracking systems that the NSA could identify human targets for soldiers and airmen in the theater, telling their precise location within a few yards. The agency had learned to find people through their communications devices, and through the signals they emitted and the traces they left. Hayden understood the power of this technology—conceivably, the NSA could have more awareness about terrorists than they had of themselves.

But to accomplish that feat the NSA would have to collect more—a lot more—in places that had traditionally been out of bounds.

 

 

 

Tenet asked Hayden to brief Bush administration officials, including the president and vice president, on what more he needed, and wanted. Hayden explained that any effective system for spotting terrorists before they struck had to meet three criteria: It must be technologically feasible. It must produce useful intelligence. And it must be lawful. Hayden spoke to the first two elements; but he left the last for others to decide.

Technology was the part Hayden knew best. Since the attacks the NSA had been tracking suspected terrorists using the traditional model, starting with a known target and then branching out through their associates. This was not pattern analysis, like what Poindexter wanted to pursue, at least not in the purest sense. Hayden's electronic sleuths might start to notice suspicious indicators in the way terrorists used phones—maybe they used intermediaries or limited their calls to under a certain number of minutes. But this wasn't the start-from-scratch concept behind TIA, in which a system would ingest every piece of data and look for patterns in it without known targets. At least not yet.

In order to build an early-warning system, the NSA would need access to the customer records of the phone companies, Internet service providers, and other purveyors of electronic communications. In the case of the phone companies, call detail records—or logs that showed who called whom; from where to where; when; and how often—constituted a valuable potential resource.

They also would need access to a source that the NSA had not been collecting systematically in the past: e-mail. Despite the agency's reputation for collecting anything and everything electronic, the agency had shied away from this increasingly common medium. One reason was parochial—the agency's terrorist hunters simply didn't believe their targets were using e-mail. They were holed up in caves and dusty campsites, after all, and had seemed to get by just fine with telephones.

But there was also a legal rationale. E-mails, unlike phone calls, were not transmitted point-to-point between two parties. Instead, an e-mail message was broken up into pieces called “packets,” each of which traveled different routes through the Internet before being reassembled at a destination. Most packets traveled through the United States, home of most of the world's Internet infrastructure. So the NSA could not be sure that even a foreign e-mail wasn't passing through America. Furthermore, the agency couldn't know for sure whether the sender or receiver was a protected U.S. person. In the face of these uncertainties, the NSA left e-mail alone.

In early 2001, the NSA had begun experimenting with e-mail as an intelligence source, but not against Arab terrorists. Wertheimer had led an exercise to find out whether Russian mobsters were supplying weapons of mass destruction to Iran. He told his team to pull in a variety of data sources, including e-mail, from different agencies. Here too the analysts weren't sold on the idea that Russian organized crime figures were logging on to the Internet to send messages. After 9/11 he told his team to operate as if the terrorists were doing just that. “You are now to presume they use e-mail. Find it,” he instructed.

As Hayden set up the new surveillance program, it was increasingly clear to NSA officials that e-mail, not phone calls, would constitute the bulk of their collection. It made sense. Terrorists had to presume that their phone calls were being tapped, or could be easily. But e-mail allowed for a certain amount of stealth. One could open a new account, use it to send a few messages, and then never return to it again. An Al Qaeda operative could log on from different Internet cafés so that he never used the same service provider. He could mask his physical location by using an Internet address that was actually based on a server in another country. As analysts began to study terrorist e-mailing habits more closely, they came to believe that members of a cell actually were sharing a single e-mail account and writing all of their messages to one another as “drafts,” which were actually stored in a folder within the account and never sent over the Internet. It was the electronic version of a drop site.

But no communications system, no matter how complex, offered total anonymity. The NSA hunters understood, as did Poindexter's team, that terrorists were susceptible to detection every time they used a phone, sent an e-mail, or made a credit card purchase. Every move left a trace, and though it might be hard to find, it wasn't impossible. But Hayden's system lacked a key component of Poindexter's brainchild—privacy protection. The NSA had no appliance of the kind Poindexter envisioned. No agency did. In this new and treacherous realm of domestic communications, the best assurance that Hayden could provide that he was playing by the rules was his word.

 

 

 

Hayden had come up with a plan, and Bush personally felt it was a good idea. The NSA director's system was technologically feasible. He felt it would produce useful intelligence—perhaps the most useful intelligence the government could expect, since it had so few human spies on the ground in Afghanistan and throughout the Middle East. One question remained: Was Hayden's system legal?

The crux of the answer actually preceded Hayden's pitch. Yes. The president of the United States had the authority to order electronic surveillance of foreign terrorists. And he didn't need a court's permission.

That was the conclusion of John Yoo, an expert on presidential war powers who found himself in high demand after the terrorist attacks. Yoo worked as a politically appointed attorney in the Justice Department's Office of Legal Counsel, the font of wisdom to which all presidents turned when they needed to know if a particular course of action was, in fact, legal. On September 25, Yoo sent a memo to a senior official in the Justice Department, who had asked whether the administration might change some language in FISA in such a way that made it easier for the government to secure warrants in terrorism cases. (That question was up for public debate as Congress and the administration hammered out major changes to surveillance rules in the USA PATRIOT Act.)

Yoo, who had no experience arguing cases before the court that granted FISA warrants, determined that it was not unconstitutional to change the wording of the law. The administration could propose that “a purpose” of surveillance was to collect foreign intelligence, as opposed to “the purpose,” which was what the law required. The administration wanted to lower the bar, and Yoo said that was fine.

But then Yoo offered an unsolicited assessment. The president didn't really need to go through FISA at all. The nation was at war, and the president had to gather intelligence to fight that war. “The Fourth Amendment declares that ‘[t]he right of the people to be secure in their persons, houses, papers, and effects, against
unreasonable
searches and seizures shall not be violated.' ” The emphasis was Yoo's own.

“Thus, the touchstone for review is whether a search is ‘reasonable,'” he continued. And the Supreme Court had made clear that the government didn't always need a warrant in order to conduct a reasonable search, Yoo noted. Quoting a case that upheld the constitutionality of an Oregon school district's program of random drug testing, Yoo wrote that “a warrantless search can be constitutional ‘when special needs, beyond the normal need of law enforcement, make the warrant and probable-cause requirement impracticable.'”

Yoo was going well beyond the question that had first been posed—whether or not FISA could be amended. Now he was articulating a basis for warrantless surveillance conducted “beyond the normal need of law enforcement” because the Fourth Amendment's requirement of probable cause to search was “impracticable.”

The president could decide whether those conditions existed, Yoo determined. And the commander in chief should rest comfortably in his authorities: “[T]he Court has found warrantless searches reasonable when there are ‘exigent circumstances,' such as threat to the safety of law enforcement or third parties.”

Yoo was a scholar, and had pulled only a few previous stints in Washington—in the midnineties, as a general counsel to the Senate Judiciary Committee, and in a clerkship for appeals court judge Lawrence Silberman, who happened to be the presiding judge on John Poindexter's appeal of his Iran-Contra convictions. Yoo had come to Justice shortly before the 9/11 attacks. But in this memo and others he wrote in coming weeks, he laid the foundation for a sweeping new regime of warrantless electronic surveillance. From these memoranda, President Bush justified his decision to grant Hayden the authority for his new mission.

The key question hanging over Hayden's surveillance system was whether or not it touched the arch of FISA. Yoo had concluded that that didn't matter. Even if the NSA's surveillance targeted U.S. persons, including those inside the United States, the administration decided that any communication involving foreign parties made the entire communication “foreign intelligence.” The president's authority here was unchallenged, Yoo concluded in a later memo. “Unless Congress made a clear statement in the Foreign Intelligence Surveillance Act that it sought to restrict presidential authority to conduct warrantless searches in the national security area—which it has not—then the statute must be construed to avoid [such] a reading.”