Pay Any Price (34 page)

Binney had argued with the lawyers, telling them that he had already built in strong protections to make certain that any data collected on Americans would remain encrypted and blocked by the software. Data on American citizens could never be viewed by NSA analysts using the system. But the lawyers had been adamant that Binney's program was too risky and would put the NSA in legal jeopardy. To the NSA lawyers, Binney was like a mad scientist who had developed a monster that had to be kept chained in the basement. As a result, Binney's system had never been allowed beyond the pilot project stage.

But now, in the wake of 9/11, it was a different story. Now the NSA not only wanted to deploy the system, but the agency wanted to unleash Binney's monster on the American public. The NSA wanted to do exactly what its lawyers had previously told Binney they feared most. Bill Binney's lab experiment was being turned into a coldly efficient weapon to spy on American citizens.

 

Jacobson told Binney that the NSA's new domestic spying operation was being set up in a big office space just down the hall from Binney's skunk works, on the third floor of Building 2B at the NSA headquarters complex. New AT&T lines were already being installed in the room.

After Jacobson told him about his secret orders and then quietly walked away, Binney immediately understood why top NSA officials had kept him in the dark. He had been making waves inside the NSA for years and had developed a reputation for being a loose cannon in an agency filled with quiet conformists. Binney was outgoing, talkative and curious, a man who found it easy to laugh and who was always eager to share what he knew with others inside the agency. He believed the agency had become too hidebound and was not keeping up with the revolution under way in information technology, and so he had gravitated to the skunk works in order to try to shake things up. He was constantly questioning the way things were done at the NSA.

That made Bill Binney stick out like a sore thumb in an agency of introverts. A stunning 80 percent of NSA personnel have been identified as ISTJ (Introverted Sensing Thinking Judging) types on the Myers-Briggs personality profile test. That meant that the NSA was filled with quiet people who valued tradition, order, and loyalty; who were organized and methodical; who believed in procedures and plans and respected rules. They were people who believed in going by the book. The joke was that an extrovert at the NSA was someone who looked down at your shoes while talking, instead of at his own.

The NSA did not use the Myers-Briggs test to determine whom it would hire. But the agency's screening process, its hunt for math and computer experts willing to work in a secret, highly compartmentalized organization where they would perform abstract, analytical functions that they could not discuss with anyone else, led to a high degree of uniformity. Many of the NSA's ISTJs were eclectic geeks just this side of Rain Man. One was known to park his car in exactly the same spot in the agency parking lot every day—no matter whether the lot was empty—and then walk precisely the same steps from that parking spot to his office. Another would buy secondhand pants, wear them every day to work for two weeks, and then throw them out and buy another pair, so that he never had to do laundry.

In addition to this disarming weirdness, there was a dark side to the predominance of this singular personality type within the agency. The introverts at the NSA never questioned authority. They kept to themselves and remained silent about the agency's secrets, for good or ill. Many NSA employees were married to other NSA employees, and often their children came to work there as well, reinforcing the agency's insular nature, enhanced by its geographic isolation at Fort Meade in suburban Maryland, far from the rest of official Washington.

This quietly obedient workforce, cramped into a zone of absolute secrecy, sometimes had the feel of a cult that was deeply suspicious of outside influences. That made the NSA ripe for corruption and abuse, an organization that wasted billions, refused to admit mistakes, and was a tempting target for leaders eager to wield its awesome technological power however they saw fit.

 

Bill Binney was one of the NSA's 20 percent who were not ISTJs yet had still found a home in its secret world. He grew up in central Pennsylvania, majored in math at Penn State, and, after joining the army in 1965, was assigned to the U.S. Army Security Agency, where he learned communications traffic analysis. He soon found himself at a U.S. base in Turkey, analyzing Soviet-bloc communications in the midst of the Cold War. In 1967, he was assigned by the army to NSA headquarters, and, after leaving the service in 1969, returned to the NSA as a civilian in 1970. He stayed for the rest of his career, and by the late 1990s, had risen through the ranks to become the agency's technical director for world geopolitical and military analysis, making him one of the top technical geeks in the agency's main operations directorate. He was also assigned to a special NSA panel that managed the agency's technical relationships with foreign intelligence services around the world.

But he had been frustrated throughout his career with the NSA's bureaucratic ways. For decades, the agency's flaws had been masked by the fact that the NSA's main adversary—the Soviet Union—was a hulking, slow-moving target that made the NSA look nimble by comparison. But after the end of the Cold War, the NSA began to drift, in search of new missions just as the Internet was triggering a digital revolution.

Binney had taken over the agency's skunk works, officially named the Sigint Automation Research Center, or SARC, in order to force-feed change into the agency's bloated system. At the SARC in the 1990s, he realized that the biggest problem for the NSA was that it still did not know what to do about the Internet and the surging growth of digital communications online. In the early 1990s, at the dawn of the Internet age, the NSA had been largely dismissive of the web. The agency had traditionally focused on cracking codes and secretly breaking into the secure communications of foreign governments and armies, and NSA officials saw little value in monitoring the new public websites that were starting to crop up all over the world. If the information wasn't secret, it couldn't be of much interest.

Binney, however, realized that the NSA was facing a paradigm shift but didn't know it yet. There was an ocean of information being created on the Internet, and the new challenge for the agency was not how to break in and collect a narrow band of data that revealed the Soviet order of battle, but how to sift through and analyze massive amounts of openly available information flooding through the world's computers.

The NSA's fetish for secrecy made things worse. The data that the agency did collect was streaming into hundreds of different databases scattered throughout the agency, all compartmented and closed off from each other. There were at least forty major databases used frequently by analysts, each one tied directly to a specific and highly secretive collection program somewhere in the world. If, for instance, the NSA managed to clandestinely access high-frequency Russian military radio traffic, the data would be fed into its own database, separate from data acquired through other collection programs. There were at least fourteen different databases for phone data alone.

Scouring the databases for information was cumbersome and time-consuming. One of the NSA's largest data repositories was known as Pinwale, and its search function was called Dictionary Search. Pinwale was so massive and poorly organized, and Dictionary Search so rudimentary, that it could take hours for the system to provide answers to many basic questions.

In the late 1990s, Binney, along with Ed Loomis and a few other NSA experts in the SARC, began to work on programs to bring the NSA kicking and screaming into the digital age. Their first attempt was a research project called Grandmaster, which was later refined and developed into a program called Thin Thread. Thin Thread was really four programs in one. The most important of the four was called Mainway. It was the primary analytical tool included in Thin Thread, and was a graphing and social network building process that was years ahead of its time. It applied chaining and link analysis to the data that was streaming through Thin Thread, providing one of the most powerful data search tools devised by the NSA up to that time. It allowed intensive web-based data searches without requiring the NSA to store the data first.

And so, at about the same time that Stanford University graduate students Larry Page and Sergey Brin were working on a research project on search engines that they turned into a start-up company named Google, Bill Binney and Ed Loomis were working in a small government lab on a project they thought would revolutionize the way the American intelligence community collected and analyzed data in secret.

Just before the turn of the century, Binney, Loomis, and their team at the SARC were convinced that Thin Thread provided a leap forward for the NSA and would put the agency back in the forefront of web-based technology. They expected NSA's top managers to embrace their program and give them the modest funding required to deploy it throughout the NSA system.

Instead, their ideas and program were rejected. First, the NSA's in-house lawyers raised objections, saying that Thin Thread would violate the law by collecting too much data on U.S. citizens, dismissing Binney's claims that the protections built into his system would comply with the law. Next, NSA managers said that Thin Thread would not “scale,” meaning that it could never handle the enormous volume of data searches that NSA's analysts conducted on a daily basis. Eventually, it became clear to Binney and his team that the real reason for the opposition to Thin Thread was that top NSA officials were already backing a different approach to dealing with the Internet—a huge new program called Trailblazer. While Thin Thread was a small, in-house pilot project developed on the cheap by a few NSA employees, Trailblazer was a sprawling multibillion-dollar program that involved large outside contractors, led by SAIC, a national security consulting firm that was deeply intertwined with the NSA and its management. SAIC was the prime contractor on Trailblazer, and in 2000, SAIC executive Bill Black was named deputy director of the NSA.

Stonewalled by management, Binney decided to go around his NSA bosses and take his case for Thin Thread to Congress, and specifically to Diane Roark.

 

Diane Roark was a staffer on the House Permanent Select Committee on Intelligence, assigned to handle oversight of the NSA. Born on a farm in Oregon in 1949, Roark had graduated from Catholic University in Washington, earned a PhD in political science from the University of Florida, and then began working for the government in 1981 when she joined the Energy Department. She rose quickly over the next few years, moving first to the Department of Defense and then the National Security Council at the White House in the Reagan administration. She had been at the House committee since 1985. By the late 1990s, her oversight work on the NSA had made her increasingly skeptical of the agency and its hodgepodge approach to coping with the digital revolution. The NSA had no strategy to make sure that its technical research would provide useful tools for its intelligence operations.

“There wasn't any coherent approach to dealing with the Internet and the digital age,” recalled Roark. “Everybody just did what they were interested in. There was a big separation between the technical and the operations people, and the technical people didn't seem to care about whether what they were doing helped the operations people. Nobody tracked how they were spending their money. It was pretty bad.”

Worse, Roark realized that the NSA didn't really want to change. “They were so used to believing that they were ahead on technology, they didn't realize that they had fallen behind. There was just about no relationship between the NSA and Silicon Valley at that time. They had extreme insularity. I was really alarmed. But they just kept saying we are okay, just give us money and everything will be okay.”

A massive computer crash at the NSA that lasted for three days in January 2000 only increased Roark's skepticism, and made her realize that the agency had to undergo fundamental change.

Her doubts made Roark a natural ally for a brilliant maverick like Bill Binney. She had first met him when Binney briefed her on the SARC's work. She had been impressed and stayed in touch with him as she began to investigate the NSA's weaknesses. And so in his battle with NSA management to save Thin Thread in the years just before 9/11, Binney decided to turn to Roark for help.

After Binney briefed her, Roark became excited by Thin Thread's potential, and she began asking top NSA officials uncomfortable questions about the program's status. She was frustrated that the program had not been used before the millennium, when there were reports of possible terrorist plots.

She also began to look more closely at Trailblazer. She realized to her horror that the NSA liked Trailblazer so much in part because it was designed to try to connect the agency's old, existing analog technology to the new digital revolution. Roark insisted on briefings from Trailblazer managers and came away convinced that the program was doomed to become a costly failure.

“Trailblazer was supposed to build an Internet software-based system on top of an analog hardware system, and it just wasn't going to work,” she recalled. “They had always felt comfortable with their existing systems. They wanted to use pre-Internet technology for the Internet age. I told them right away that would fail. It was just common sense.” (Roark proved prescient. Years later, the NSA abandoned Trailblazer. After spending billions of dollars on the program's development, the agency was finally forced to admit that it would not work.)

 

By early 2000, Roark's intervention began to infuriate NSA Director Michael Hayden. He had already decided to go with Trailblazer and SAIC over Thin Thread, and he wanted Congress to give the agency the billions of dollars that Trailblazer would demand, no questions asked. He certainly did not want to have to explain himself to some lowly congressional staffer.