The Watchers

by Shane Harris

  CHAPTER 15

  CALL TO ARMS

  Poindexter had known Donald Rumsfeld as long as anyone in Washington. They’d met almost twenty-five years earlier in the Ford administration, when their meteoric career paths crossed at the Pentagon. Rumsfeld was the new secretary of defense, the youngest in the department’s history, and Poindexter had come to work for Admiral James Holloway, the chief of Naval Operations, as his executive assistant and all-around right-hand man.

  Years later Poindexter and Rumsfeld met again at the White House. Poindexter was on Reagan’s NSC staff when Rumsfeld was the president’s chief emissary to the Middle East during the hostage crisis. They had battled the fires of suicidal terrorism, each from different positions. And while their professional courses had diverged after that, Poindexter and Rumsfeld had remained in similar orbits.

  A few weeks after Poindexter moved into his new office, he dropped in on Rumsfeld, who was back at the Pentagon, this time as the oldest defense secretary in history. Poindexter had been there for the welcome-back ceremony some months earlier, but they hadn’t had much time to chat. Rumsfeld arranged for a private luncheon in his office, so they could catch up.

  It was a friendly meal, and mostly a social call. Rumsfeld asked after Linda and Poindexter’s boys. They reminisced about Holloway, who, like Rumsfeld, was a wrestler, the former for Navy, the latter at Princeton. Rumsfeld liked to bring Holloway into his office for beers and bull sessions; they’d prop their feet up on the furniture and relive favorite matches while puffing on cigars. Poindexter enjoyed being near those casual moments and such men of great institutional power and professional wisdom.

  Poindexter had appraised Rumsfeld early on as a sharp bureaucrat and a climber, but one who possessed a refreshing genuineness that Poindexter found lacking in so many of his civilian superiors. These men merely played the part of statesman, he felt. Like John Warner, for whom Poindexter had worked when he was secretary of the navy. These were the days before Warner married Elizabeth Taylor, and before Virginia sent him to the Senate; and while his political future seemed momentous, Warner once confided to Poindexter that what he really wanted to be was an actor. Well, sir, the young officer had thought, you’re doing a pretty good job.

  Rumsfeld was a lot like Poindexter. In the crucible of the Pentagon they’d forged a mutual distrust of the military and intelligence bureaucracy, which so often resisted the president’s policies. And it extended to the Congress, whose incursion into the commander in chief ’s sacred realm of national security had been on a long and steady march since the fallout of Watergate.

  The suspicion culminated in political warfare during Iran-Contra. Poindexter believed that congressional Democrats had set their armies on him personally as an avenue to the president. And when they could not breach his wall—“The buck stops here with me”—they hated him for it. The sentiment was mutual, though. Poindexter despised Congress for launching the attack in the first place.

  Over lunch, Poindexter and Rumsfeld remembered the caldron of Lebanon and the attack on the Marines, the first shot in the intermittent series of battles that had brought them to their present positions. And they uttered the old lamentations—how the intelligence agencies had too long demurred in the fight over the years, hadn’t embraced the expansive powers that presidents had meant to give them. Look at how they’d failed to use Executive order 12333, Poindexter said. “By the time that policy gets down to the working level, every layer of command has backed away from the edge a little bit, to make sure that nobody stepped over,” Poindexter said. “They’re overly cautious,” he told the secretary.

  “You’re absolutely right,” Rumsfeld replied. And, he said, he’d had a similar problem trying to promulgate combat rules of engagement. Bush would agree to one set of rules, but when Rumsfeld asked the spear-carriers in the field what those rules were, their version bore no resemblance to what the president had agreed to.

  Poindexter and Rumsfeld both thought people were backing off, playing too cautiously, when they should have seized their hard-won authorities, unbowed and unafraid. The attacks of 9/11 were as much a result of weakness as they were a chance to erase it. Now each man, in his own way, was seizing the moment to bend the system to his will and toward the end he thought it ought to run—Poindexter in quiet, studious steps, Rumsfeld in cage-rattling bursts of bravado, as he sought to transform the military for the war on terror. The bookworm and the wrestler. But they were heading down the same path.

  Poindexter didn’t lay out all the technical details of TIA for Rumsfeld. But he gave the secretary the overview, explained the concept and his goals. He wanted Rumsfeld to understand what he was up to, why he had returned. Rumsfeld expressed his support for the research and affirmed to Poindexter that it was a good idea. And that was enough for him. Poindexter left lunch feeling that he had an ally, an old and faithful friend, in the highest place.

  Poindexter launched a broad campaign among the intelligence and national security agencies to enlist support for TIA. Rumsfeld had given his verbal support, but Poindexter wanted more. He thought that individual agencies should join in his research, by experimenting with the data collection and analysis tools that his team soon would be fielding. As they brought new tools on-board, Poindexter hoped that the agencies’ own analysts would put them through their paces. Only agencies like the CIA, the FBI, or the NSA could truly discover what worked and what failed. After all, they were Poindexter’s ultimate customers. Once he built a TIA prototype, it would be up to them to use it in the real world.

  Although Poindexter distrusted what he saw as the bureaucracy’s calcified traditions and narrow mind-set, he knew that without broad support TIA would languish on a shelf. Yet he was not prepared to let them tinker with his masterwork in a vacuum. And so, Poindexter hatched a plan to enlist the agencies in his effort and, at the same time, influence them.

  Any agency that wanted to test-drive the TIA tools would have to install a special node on a data network he set up exclusively for that purpose. It was a digital laboratory where the agencies could tinker and toy, then share their results. The tools lived on the network. But the nodes—consisting of little more than a laptop computer—would sit at the agencies themselves.

  Poindexter would install the central node for the network in a place of his choosing. When the time came to pick it, Tony Tether suggested that Poindexter go meet with a technologically astute Army general named Keith Alexander to ask if he could borrow space in his facility outside Washington. Though not well known to Poindexter at the time, the place shared TIA’s spirit of rebellious innovation. It was the Information Dominance Center, former home of Erik Kleinsmith and the Able Danger team.

  Alexander, the commanding general of the Army’s Intelligence and Security Command at Fort Belvoir, was a career military intelligence officer and West Point graduate. He was a warrior geek like Poindexter and his crew. Alexander earned two master’s degrees from the Naval Postgraduate School, one in systems technology, with a specialization in electronic warfare, and the other in physics. He was also the intelligence officer for the 1st Armored Division during the 1991 Gulf War.

  Poindexter had come to know Alexander during the Genoa years. The general had attended the technology demonstrations, so he was already up to speed on Poindexter’s thinking. Alexander agreed to give Poindexter and his team some space on the sci-fi set of the IDC, and to let them use its existing data network, which ran on the Joint Worldwide Intelligence Communications System. This was the same information source that Kleinsmith and his analysts had tapped to harvest classified messages and traffic during Able Danger, and the intelligence agencies already had access to it.

  Above all, Poindexter wanted real users and real data to run his experiments. The IDC gave him both. It already had the physical infrastructure to link agencies together and let them share intelligence. More nodes could be installed as more agencies joined. The IDC provided a natural setting. But there too space was scarce. Withi
n feet of where Kleinsmith had worked two years earlier, Poindexter and his team received an office about the size of the broom closets they were working in back at DARPA headquarters.

  But again it wasn’t physical space that Poindexter was after. Now he had his virtual lab. He called it the TIA network. And as winter melted into spring, Poindexter enlisted many agencies to join it.

  He made a sales tour. He felt that he couldn’t emphasize enough the urgency that propelled him. He evangelized about TIA, delighting his audiences with both his enthusiasm and the thoroughness of his approach. He seemed to have thought of everything. Poindexter remained convinced that another attack was either imminent or already well along in the planning sequence. Fortunately for him, the senior officials he was meeting with were gripped by the same fear, and they were primed for new ideas.

  Langley, Virginia, was on Poindexter’s whistle-stop tour. He knew the CIA’s capabilities intimately, and its limitations, and so he felt particularly emboldened to tell the agency’s top men why they should join his quest.

  Poindexter drew an impressive crowd for his first TIA briefing. The agency’s executive director, the third in command, convened the meeting. A. B. “Buzzy” Krongard was a former Marine and longtime executive banker who’d only joined the CIA a few years before the 9/11 attacks. He was the top-ranking official there, joined by John Brennan, the deputy executive director and a onetime station chief in Riyadh. (Brennan spoke passable Arabic and recently had served as CIA director George Tenet’s chief of staff.) Donald Kerr, the deputy director of the agency’s science and technology directorate, and Alan Wade, the chief information officer, rounded out the group. Since the CIA was spydom’s central management office, Wade was also the top computer systems official for the entire intelligence community.

  This wasn’t the first time Poindexter had been to Langley since the terrorist attacks. Before the Information Awareness Office was formally established, Kerr had set up a special committee on science and technology and asked Poindexter to talk to the members about Genoa. As he listened to them kibitz, his frustration swelled. These were the same people who had failed to adopt his ideas years earlier. Now they’d invited him out to give a speech he’d made countless times. Poindexter pounded his fist on the table. “We have got to change the way we do business!” he boomed.

  It was a rare and notable venting from the usually placid admiral. So, a few months later, when Poindexter was back at CIA headquarters briefing Kerr and several of his senior colleagues on his new vision, he had their attention.

  Poindexter talked about TIA, its animating concept and its goals. This was bold stuff by any agency’s standards. But Poindexter’s expansive view of the terrorism intelligence problem struck Wade in particular as unique. Indeed, Poindexter was one of the only people Wade had heard talk about the challenge in such broad terms, and as an end-to-end concept. He’d thought out how to obtain information, how to use it, what limits to place on that use. In the first, urgent days after the 9/11 attacks, Wade mostly had heard his colleagues discuss just the first part—how to obtain information, and more of it. With the threat of another attack looming, the agencies began to covet one another’s most precious resource—their secrets—and to resent one another’s stinginess when so many refused to share. They could justify holding back out of an overabundance of caution; they didn’t want to expose vital sources and methods for acquiring intelligence lest they be lost forever. But in truth, information was one of Washington’s most valuable currencies. The intelligence agencies measured their status by it and were loath to diminish their influence by spreading the wealth.

  But Poindexter had figured out a compromise. He explained to Wade and the others that he would let the agencies come to the TIA network and bring as little or as much information as they wanted. In exchange they would be the first to experiment with his new tools. As the agencies’ haul of intelligence piled ever higher, they became desperate for ways to make sense of it. Poindexter could fill that need, with his test bed for the electronic war on terrorism.

  Wade liked the idea, but he heard something even more intriguing in Poindexter’s pitch, a concept that he hadn’t heard in any of the tech briefings he’d sat through since 9/11: the words “protect privacy.” Wade thought that Poindexter’s was the first ambitious information architecture that included privacy from the ground up. He described his privacy appliance concept, in which a physical device would sit between the user and the data, shielding the names and other identifying information of the millions of innocent people in the noise. The TIA system would employ “selective revelation,” Poindexter explained. The farther into the data a user wished to probe, the more outside authority he had to obtain. An intelligence analyst mining and moving information would only encounter individuals represented as numbers, or as some anonymous marker. Poindexter also proposed an “immutable audit trail,” a master record of every analyst who had used the TIA system, what data they’d touched, what they’d done with it. The system would be trained to spot suspicious patterns of use: say, an analyst poking around in domestic data sources that had no bearing on a terrorism investigation. Poindexter wanted to use TIA to watch the watchers.

  The CIA team liked what they heard. “Alan, you follow this,” Krongard advised. As TIA evolved, Wade and Poindexter held several more meetings, and the agency eventually installed a node on the TIA network.

  Poindexter was glad for the enthusiastic response. But he wanted someone to challenge him. To poke holes in his theory and his reasoning, particularly his novel concept of selective revelation. He wanted someone to tell him he was wrong.

  Fran Townsend seemed like she would know for sure. The senior official in charge of intelligence at the Coast Guard, which now played a pivotal role in homeland security, she was already on Poindexter’s list of potential TIA participants. But he was most intrigued by her previous jobs and her peculiar area of expertise. Townsend was an authority on surveillance law.

  Townsend had spent much of her career at the Justice Department. She started out in 1988 at the U.S. attorney’s office for the southern district of New York and then spent the late 1990s working for Attorney General Janet Reno in Washington. Townsend had worked most of the major terrorism cases of the decade, including the Africa embassy attacks and the USS Cole bombing. She was among the few career officials in Justice who were following Al Qaeda and increasingly worried that law enforcement alone was ill equipped to halt its advances.

  Like Poindexter, Townsend was seared in the early fires of the terror wars. She had been in law school at the time of the Beirut bombing, and as a prosecutor she became convinced that Hezbollah was responsible for the attack and had never paid for its crimes. At the tail end of the Clinton administration, Townsend oversaw the preparation of warrant applications to the Foreign Intelligence Surveillance Court, the secretive panel of judges that oversaw the wiretapping of suspected terrorists and spies. Townsend was in charge of making sure the government’s requests complied with the letter and spirit of FISA. She knew the ins and outs of the law, as well as the temperament of the judges and the strict barriers that had been imposed between the law enforcement and intelligence arms of government. (Foreign and domestic. Those same bright lines that had kept the NSA outside America.) By the time Poindexter sought out her counsel, Townsend was one of the few people alive who legitimately could call herself a FISA expert.

  When she got the phone call, Townsend was taken aback. She had known Poindexter only by reputation. The idea that a former national security adviser would want a meeting with the intelligence director of the Coast Guard struck her as flattering and unusual. She wasn’t about to say no. Townsend invited him to Coast Guard headquarters to brief her personally, as well as the service’s intelligence and legal experts.

  Poindexter gave the standard TIA presentation. Townsend could tell that he’d already anticipated having to cross a variety of legal hurdles. There was a gap between what Poindexter envisioned and what wa
s doable under current law. But his proposals, specifically selective revelation, appealed to her. If TIA could actually “anonymize” the data it examined, and only reveal names through a legal process, then Townsend thought it would be an extraordinarily valuable intelligence tool. She imagined that it could help FBI agents find the most valuable leads to investigate, filtering out the dead ends and the background noise. That made TIA a worthwhile experiment, she thought.

  But she knew that Poindexter wanted advice, not plaudits. So she tallied up a list of his vulnerabilities. “You’re going to need advocates on the Hill,” Townsend said. The right staffers on the key committees, and he’d need enough face time to fully explain what TIA was and wasn’t. Nothing in his proposal struck her as legally problematic for the moment, since it was still just a research project. But Poindexter was thinking years into the future, to a day where perhaps the government would have access to now off-limits sources of information.

  Townsend knew how seriously the Justice Department and intelligence community obeyed “the wall” that separated their two worlds, in which information was acquired under different legal standards and used for different ends. Poindexter’s vision would merge those two worlds. He was taking on two cultures at once.

  Townsend had seen others try that, and the public reaction was explosive. She shared a cautionary tale with Poindexter. During the Clinton administration the FBI had launched a new e-mail monitoring technology named Carnivore. It was a packet sniffer, a tool that could track a surveillance target’s online messages. But a combination of factors, not least of which was its unfortunate name, overwhelmed any chance for widespread acceptance. The public harbored abiding suspicions of law enforcement in general. Now the FBI wanted to read people’s e-mail? Technological privacy activists were also outraged. Throughout much of the 1990s the Justice Department and the National Security Agency were making big public power plays to control the emergence of new Internet technologies. They’d pushed through a law that required all telecommunications companies to build their network in such a way that they could be instantly and secretly tapped. To a lot of people Carnivore was one more instance of relentless government excess. It attracted all the wrong kinds of attention. And Townsend had a front-row seat to what she considered a public relations train wreck.