The Watchers (47 page)

But the enormity of the plot was overshadowed by what lay behind it. U.S. and British intelligence officials agreed that the terrorist cell's command and control center was not in the UK, but in a remote, mountainous area of Pakistan. The terror gang that the CIA and U.S. soldiers had attacked in Afghanistan after 9/11 had slipped across the border and now found refuge in lands beyond the control of either country. There they regrouped, rejuvenated, and appointed new leaders. Al Qaeda 2.0 was in business.

After the British tipped off the Americans, the NSA was able to intercept e-mails that Ali sent to an apparent terrorist minder in Pakistan, an Al Qaeda operative named Rashid Rauf. The messages contained cryptic references to a plot, and that alarmed senior Bush administration officials. This seemed like proof that Ali's campaign was being directed from Pakistan. In Townsend's equation, Rauf was like that mysterious person C. And he revealed a lot, indeed. After the discovery of Rauf, U.S. officials leaned on the Brits to roll up the cell before it was too late. But the Americans refused to hand over the e-mails, which had been found on a Yahoo! server in the United States. These were the fruits of secret surveillance. And when Ali eventually went on trial, the British couldn't introduce the e-mails as evidence.

Had the attack succeeded, and been traced back to Al Qaeda in Pakistan, the Bush administration might never have survived the public condemnation. An inescapable narrative would form quickly and show that the nation's leaders were asleep at the watch: Al Qaeda terrorists, who had escaped U.S. military capture five years earlier, were allowed to gather their strength and form new contacts in the lawless regions of Pakistan under the noses of American intelligence officials and their allies. Their attack would have eclipsed 9/11 in scale and ambition, casting doubt on the Bush administration's conduct of a war on terror that had protected neither American lives nor American values.

This image of utter fecklessness would be compounded by the memory of the administration's dismal, at times unconscionable, response to Hurricane Katrina a year earlier.

Three months after the British authorities broke up the alleged cell, Dame Eliza Manningham-Buller, the head of Britain's MI5 security service, announced that officials were aware of nearly thirty active terrorist plots in the country. “These plots often have links back to Al Qaeda in Pakistan,” she told an assembly at Queen Mary's College in London, “and through those links, Al Qaeda gives guidance and training to its largely British foot soldiers here on an extensive and growing scale.”

A British army, trained by Al Qaeda in Pakistan to kill thousands, had turned its sights on the United States. That they actually had a chance to succeed was as terrifying as the plot itself. For all of the sacrifices of the past five years, all the risky, secret programs, the perilous authorities, and the moments when the government might break under the weight of a relentless war fought in the dark, the airline bombers could simply have gone undetected. Absent a tip, a gut-level suspicion, or a lucky signal, they could have slipped beneath the waves into the noisy depths, only to reappear in a fiery blast.

It took a combination of sophisticated surveillance and classic shoe-leather investigative work to untangle the bombing plot. Cameras, microphones, and the NSA's massive probe of the telecom system were merely tools. Effective though they were, machines could not solve problems without human ingenuity. It was another reminder of the limits of technology.

When the urgency of the moment subsided, some began to question how serious the threat had been. The CIA looked into Homeland Security's claim that the terrorists could bring down the planes with relatively small amounts of explosives. That was impossible, the agency's scientists determined. It would take a lot more material to down a jetliner. The Homeland guys were overreacting.

The seemingly arbitrary nature of the liquids restriction made airport security, already rife with inconveniences, appear even more pointless. And it reinforced the popular belief that the Bush administration was using fear of terrorism as a weapon, ratcheting up the nation's color-coded alert system whenever the public's overall confidence waned or its opposition to policies grew.

But for all that, the plot was still foiled. Luck and detective work might have played a bigger role than secret surveillance and technological wizardry. But who could say for sure? All anyone knew was that the bombs didn't go off. And as far as the U.S. government was concerned, that equaled success.

It was hard to argue that a disaster hadn't been averted. Ali, who denied the breadth of the plot at his trial, nevertheless admitted that he at least planned to set off a bomb in Heathrow Airport. He and two codefendants also admitted to planning a series of small-scale bombings in order to generate publicity for an Internet documentary he was making to protest British foreign policy in Iraq, Afghanistan, and Lebanon. A jury convicted Ali of conspiracy to murder, although they could not reach a conclusion on the major charge in the case, that he'd engaged in a conspiracy to destroy airliners.

Only after a second trial, in 2009, were Ali and two others finally found guilty of that crime. At this trial, NSA's intercepted e-mails were introduced as evidence, and they apparently persuaded the jury. Rauf, the terrorist minder in Pakistan with whom Ali corresponded, had reportedly been killed by a CIA missile strike in November 2008. Just two months later, the Justice Department issued a request for the e-mails with a court in California, where Yahoo! was headquartered. American officials then gave the e-mails to the British. Perhaps with Rauf dead, there was no longer any reason to monitor his e-mail account or to keep the messages secret.

The initial failure to convict Ali and his conspirators on the plane-bombing plot underscored the current of uncertainty that ran through any terrorist investigation. It could be difficult to prove an accusation that seemed so readily supported by the facts. And indeed, officials had falsely accused some people of horrific acts before. In 2004 the FBI incorrectly linked an Oregon attorney, Brandon Mayfield, to the Madrid train bombings that killed nearly two hundred people. And in June 2006, only weeks before the London plot unfolded, U.S. law enforcement officers arrested seven men in Florida on the dubious charge that they were plotting to blow up the Sears Tower. Investigators' own records showed that an FBI undercover operative, posing as an Al Qaeda member, might have stoked the accused's appetites for violence, and that the men had few means of carrying out their fanciful plan. Two years later, all federal authorities had to show for their case was one acquittal and two mistrials after juries deadlocked. Ultimately, in May 2009, five of the original seven defendants were convicted of providing “material support” to a terrorist organization.

This was the nature of the game. A string of partially completed plays, the occasional fumble. And no one could say for sure who was winning. And where exactly the ball was on the field. When the rare victories came, when the hunters could claim a point, the celebration was ecstatic and fleeting. Though few would have heard it, the echoes of
Achille Lauro
bounced off the White House walls on the morning of August 10. People had worked. The system had worked. Not perfectly, just dependably. Maybe that's what John Poindexter had in mind all along.

But there was one final lesson. A harbinger, really. The UK was the closest thing to a surveillance state in the Western, democratic world. Perhaps no city in the world was more heavily monitored by security cameras than London. Citizens' every steps were recorded and monitored by an army of algorithms, which were in turn overseen by human officers. The cameras had started going up years earlier, after the kidnapping of a British child, a crime that both shocked and angered the public. Now, though, the cameras had been turned on another homegrown problem—immigrant Muslims and their children. Unlike such populations in the United States, who tended to assimilate into American life and culture, these new citizens were highly “disaffected,” security officials liked to say, angry at the disparities in wealth and social mobility in a stratified British society. And their ranks were growing. The British had dealt with this kind of insurgency before, in Northern Ireland. There too the fight had been one of national survival, and the tactics and techniques employed far exceeded anything that the United States had ever countenanced in the name of homeland security. The fact was, the UK had established a domestic intelligence society, both in law and in culture. If the British rules applied in the United States it would be as if the CIA had supplanted the FBI. Customary standards of evidence and reasonable suspicion would no longer apply to investigations. You could be followed, bugged, and rounded up because someone
believed
you were up to no good. Given the UK's long history with domestic terrorism, it was perhaps easy to understand their fears of Muslim extremists. But the United States clearly didn't have the same level of homegrown threat. And yet, American intelligence agencies were moving closer to the British model. What would it take to finally tip the balance? In averting the planes' bombing, the United States had avoided more than just another massive attack. It had avoided the answer to that question.

 

 

 

 

Technology had its limits. And every new battle in the terrorist war offered a palpable reminder of that unassailable fact. But that didn't stop agencies across the government from reaching for a brass ring.

Years after Poindexter had been pushed out at DARPA his system lived on, both in a physical sense and, more powerfully, as an idea. What had started as a research project in the Pentagon's think tank had morphed into operational systems being built on the fly in silos across the government.

Over at the Homeland Security Department officials had launched a new program with a curiously familiar pedigree. They called it ADVISE, for Analysis, Dissemination, Visualization, Insight, and Semantic Enhancement. It was a mouthful, and not especially catchy. Perhaps the program's managers would have had more luck with the original name—the BAG.

Back in early 2003, officials in the new Homeland Department's science and technology directorate went looking for data-mining and visualization tools to assist their counterterrorism mission. Poindexter was going through the wringer at precisely the same time, but the potential for political backlash didn't stop them. In fact, the law establishing the department had specifically required it to field new technologies for identifying terrorist threats within the United States, and it authorized officials to tap the expertise of the national laboratories.

Officials went to Lawrence Livermore, onetime home of Dr. J. C. Smart, father of the BAG and expert on graph theory. That's where they were first turned on to the power of graphing tools, and arguably they were oversold on their potential. Because at the same time that the Homeland Security Department was investigating the BAG, their fellow terrorist hunters at Fort Meade were struggling to make the recalcitrant technology work.

The BAG was simply ill equipped to handle the volume and complexity of information the NSA was dealing with; in technical terms, it could not “scale.” But
any
government agency in the terror fight was going to be drowning in data. It was unlikely that the BAG, or tools like it, would solve their problem.

And yet it had been insinuated into other agencies in largely unseen ways. In September 2005 a team that had formed to run Homeland Security's ADVISE program and build a working prototype huddled at the Johns Hopkins Applied Physics Laboratory, in Maryland. There a new center had been created to help guide homeland security research. Steve Dennis, a technology official from Homeland Security, offered up the history of the program. Participants saw a diagram that conveniently illustrated how the work of Dr. Smart, mentioned by name, found its way into the present system's “architecture,” or underlying blueprint. Smart's research had been part of an architecture called “Nebraska,” which was eventually replaced by an updated version, dubbed “Kansas.” This was the current plan that formed the core of ADVISE. (Technology researchers usually recycled one idea into the next and devised a consistent naming pattern to mark the evolution. The same way that Poindexter's team used the names of winds, the ADVISE builders used the names of states.)

The meeting participants saw the BAG labeled on their diagram. A simple flow chart showed it leading out of Nebraska and onto another pair of program names: “KSP” and “NSA Operational System.” KSP stood for Knowledge System Prototype, a major initiative under way at the signals agency. The map showed a steady progression: What began at Livermore filtered over to the NSA, and now it had found its way to the Homeland Security Department. Lest there be any confusion about the roots that the department's data system shared with the terror-hunting apparatus at the NSA, the word “ADVISE” on the chart was followed by a big arrow pointing at “NSA Operational System.”

The message was clear.
If this stuff was good enough for the NSA—well, it's surely good enough for the Homeland Security Department!
Never mind that the BAG had failed to identify terrorists dependably. The ADVISE team had been sold on the tool already.

Though it was still incomplete, ADVISE was no theoretical concept. So far versions had been deployed in support of four “mission areas” of the Homeland Security Department, including customs and border protection, biological weapons and disease defense, and weapons of mass destruction countermeasures. The fourth area, designed to protect critical infrastructure such as electrical grids, dams, and nuclear reactors, was also using the ADVISE architecture. Known as the Threat and Vulnerability Information System, it included a “threat mapper” that gave counterterrorism analysts a way to depict information geographically across the United States and then search underlying databases of information about possible terrorist actors and agents.