The Watchers (17 page)

Kleinsmith and a colleague spread out a map of the United States on a table in the team's workspace. He scanned the terrain, not immediately sure what he hoped to find. Lots of cities. Lots of rivers.

Rivers
.

Without saying a word, Kleinsmith grabbed a stack of yellow Post-it notes and a pair of scissors. He cut the notes into triangles and affixed them to apparently random points on the map.

“What are you doing?” his puzzled colleague asked.

“I'm marking the areas where they're stealing this stuff.”

“How do you know?”

He drew a path down the St. Lawrence River from Canada, into New York, and onto the Great Lakes. Sixteenth-century explorers had used this route when they first came to North America, he explained. Invading French and British armies had used it. Even the Germans snuck u-boats into the river during World War II. The river was a natural entrance point to the continent, and on its banks, hubs of commerce and activity had sprung up. “If you go back and look at intel,” Kleinsmith told his colleague, “you'll see that the pattern of theft fits the river.”

She checked. Indeed, the intelligence indicated that the technology in question was either found at or associated with universities and corporate facilities located in the geographic corridor Kleinsmith had isolated. He never thought that Chinese spies were floating missile parts up the St. Lawrence River. But he knew that this place, over time, was a beehive of activity. People and goods moved in and out. They conducted trade. They waged wars. These were entry and exit points. People come in, they take things out, whether physical devices or useful information.

If the Pentagon wanted a way to model the espionage threat, here it was. Go back to history, formulate a hypothesis, and then see what the intelligence says. It all seemed perfectly logical.

 

To Kleinsmith and his tiny team, the conclusions of their analysis were inescapable. The Chinese had established a veritable underground network inside the United States. The analysis showed front companies probably controlled by government officials. Ostensibly legitimate Chinese academics, scientists, and businesspeople, some of them with contacts and even teaching posts at major research universities, were well positioned to send technology designs and other useful intelligence back home.

The leads needed to be vetted, no doubt. And as word of the China experiment spread throughout the intelligence community, Kleinsmith's team drew vehement detractors. One Defense Department analyst confronted Kleinsmith over his conspiratorial notions, the “connections” he had found sitting right out in the open: “You could find a connection between China and dog poop, the way you're representing this.”

“You're right,” Kleinsmith replied. “But that's not what we're showing.”

Kleinsmith and his team, acting on their own and without prior knowledge of the threat, had demonstrated that the congressional commission on Chinese spying was essentially right. And they'd done it all in a matter of days, with relatively little effort. So, if an entire congressional investigation had now validated the IDC's methods, why wouldn't the Pentagon eagerly pour money into more advanced tools? Why wouldn't they build more IDCs to stop spies and counter a whole range of new threats that befuddled traditional agencies?

They were good and fair questions. But Kleinsmith and his supersleuths had raised another that was far more troubling to the senior officials who'd put them on the case: How did these techniques, impressive though they were, not violate almost every privacy law on the books?

Kleinsmith knew the regulation well. He could rattle it off as instinctively as his home address. DOD 5240.1-R:

Kleinsmith didn't have to be a lawyer to know what that meant. Regulation 5240 was the guardrail against domestic spying by the military. And he was bumping up against it.

The government had been down this torturous road several times. After World War II, the National Security Agency began collecting all telegram traffic leaving the United States, a practice that continued well into the 1970s. The FBI had set up a covert spying operation against the Black Panthers and other perceived “hostile” groups, including political opponents of various administrations. Critical journalists had ended up on enemies lists; their phones were tapped, their movements were tracked, and in some cases their finances were audited by the IRS. The exposure of those and other scandalous operations prompted a near full stop on domestic intelligence work. For decades, America's spies had kept their noses pointed overseas and had left stateside investigations of security threats to law enforcement agencies. They operated under crystal clear rules about what information the government could collect on American citizens.

The IDC's techniques muddied those rules. As Kleinsmith and his analysts harvested thousands of Web pages, they unavoidably vacuumed up the names of American citizens—probably thousands of them—who were mentioned in news articles, in chat rooms, or on electronic bulletin boards along with people Kleinsmith actually wanted to know more about.

They were the innocent bystanders of the investigation. Ordinary people who had done nothing wrong, and whom the government had no reason to believe were helping the Chinese or any other hostile service. Their connections to potential targets were coincidental. A college president mentioned on the same page as a visiting Chinese scholar. A U.S. executive who visited China on a trade delegation. Kleinsmith had plenty of logical explanations, but in order to verify them, he'd have to dig deeper into the connections. And that meant, in effect, investigating an American citizen without legal cause.

Even if he'd had the authority to vet these names, he didn't have the time. There were simply too many names to sift through without slowing down the rest of the analysis. But it wasn't just names that the IDC team collected. The harvest revealed facts about a person's life. His job. His hobbies. Any trivial fact mentioned in an online newsletter, an annual report, a transcript, became part of an ever-widening profile. What could an industrious snoop have found out about Kleinsmith just by scanning the Internet? That he was thirty-three years old. That he was a Cub Scout den leader. That he liked to play online war simulations in his spare time. It was all out there for the taking.

Only with a duly authorized warrant, issued pursuant to an official investigation, could an intelligence agency start building files on U.S. persons. And by law, that designation covered not only American citizens but also legal residents, American corporations, and even unincorporated associations substantially composed of American citizens or resident aliens.

Regulation 5240 was the Defense Department's legal guide. It specified what kinds of information an agency could collect (it must concern foreign powers and governments), under what circumstances (generally, with a warrant or pursuant to a legal authorization), how long it could be retained (if it had no demonstrable intelligence value, no longer than ninety days), and how widely the collecting agency could share it.

These were the rules, and people in Kleinsmith's business were trained to follow them. But technology had outpaced such restrictions. Now they were holding him back. His team had obtained no warrants because they weren't specifically targeting anyone. And their task was an experiment, not an operational mission. To confuse matters, the team was combing through publicly available information, which was not absolutely off-limits to intelligence agencies. But after they harvested those Web pages, they deposited them in storage with classified
government
data. The names of U.S. persons were being commingled with those of people actually under investigation.

Kleinsmith and his analysts risked breaking just about every rule spelled out in 5240 and the laws from which it flowed. They collected information. They stored it. They were searching it after they stored it. They intended to craft reports, which would be shared, perhaps widely. Their data mining was some bizarre hybrid of covert monitoring and public research.That's how the Army's lawyers saw it, and they conveyed that concern to the Pentagon's senior leadership.

But there was something else about the China experiment that proved far more troubling and politically perilous. The IDC wasn't just collecting information on ordinary people. The names of many prominent Americans popped up as well. For starters, there was Condoleezza Rice, the provost of Stanford University and a former member of George H. W. Bush's National Security Council. The hit was another by-product—Stanford, an elite research institution, hosted a number of Chinese scholars and delegations.

But the connections didn't stop there. William Cohen, the secretary of defense, also appeared. Was this the promise of “intelligence on steroids”? Condoleezza Rice and Bill Cohen implicated in a military smuggling ring? Hamre and other Pentagon leaders blanched at the political implications. The controversy over technology leaks to China had inflamed the White House. Republicans accused Bill Clinton of weakening U.S. export control laws in exchange for smoother diplomatic relations with Beijing and after generous campaign contributions from U.S. missile and satellite companies. The whole affair was radioactive. And so a chill shot through the Pentagon's upper ranks when Kleinsmith's China harvest churned up the name of First Lady Hillary Rodham Clinton.

The Pentagon placed an urgent call to the House Intelligence Committee. Hamre wanted to come up to the Hill and talk to the staffers personally. A number of them had heard of the IDC already, but they didn't know much more than that the analysts were using computers in new ways.

Hamre explained the problems the IDC had encountered with “U.S. persons” information. He wanted to know, Should the Pentagon shut down the program? The legal implications were obvious to everyone, but the IDC had shown real potential. They had to weigh the competing concerns of an intelligence breakthrough and privacy law. The Chinese were gaming the United States, but the country's laws arguably were making it harder to fight back.

The committee's lawyers had an easy gauge for deciding whether to continue a promising program like this or to kill it: Does it pass the
Washington Post
test? In other words, if the details of this operation ended up on the front page of the nation's most important political newspaper, would you—or more to the point, your boss—feel comfortable explaining it? Could you live with the headline “Defense Department Collecting Information on Americans in Espionage Investigation”?

No way. Even if the IDC analysts operated in a legal gray area, and there was some indication they did, this operation would never pass the test. The committee staff also wondered why the IDC was on this job in the first place. The National Security Agency was the expert in handling U.S. persons information. The Army intelligence command didn't have a clue about this stuff. It seemed to some staffers, as they listened to Hamre and worked through the legal implications, that the IDC team was utterly unaware that they even had a legal problem.

In October 1999, two months after the IDC had started the China experiment, the Department of the Army sent down new orders to the team. They should conclude the experiment and then purge all the data from their systems. That included the harvest and any reports they'd created, with visualization tools or by hand. The IDC must also return all information obtained from other agencies. And under no circumstances could they retain any nonessential U.S. person information for more than ninety days. Since they had no time to vet, everything would have to go.

Kleinsmith tried not to fret. Hamre was just one of many customers, he told himself. The IDC was still a hot ticket across the military commands. They had plenty of work to keep them busy.

Kleinsmith wasn't oblivious to the lawyers' concerns. He knew that the IDC's methods skirted the edge. But he also knew that most of his customers weren't bothered.

Over the next several months Kleinsmith gave more than one hundred briefings on the IDC to members of Congress, generals, and senior government officials. He could tell almost immediately whether someone got it. The look in his eyes. How he leaned forward. If he nodded as Kleinsmith elaborated on the technology's potential. Hamre got it, he thought. And so did two officers from the Army's Special Operations Command, who showed up at the IDC in December 1999 looking for information about a little-known terrorist organization called Al Qaeda.

 

 

 

 

A pair of officers showed up unannounced, a Navy commander and a Marine captain. Ordinarily, military visitors got the VIP reception, with a formal welcome and nickel tour. But these two had meant to come in quietly.

They explained that Special Operations Command, headquartered in Florida, had heard about the IDC's work on the China experiment. People were impressed. Word was that the analysts here had developed some unique capabilities. Could they ask Kleinsmith some questions about that?

Kleinsmith brought the officers into the IDC's main conference room. He was used to the dog and pony show by now, and the China presentation was as good a way as any to introduce people to his new breed of analysis. But as he walked the officers through those results, they interrupted him with pointed questions that had nothing to do with tracking military hardware.

These men wanted to track people. And, presumably, kill them. They never revealed a specific mission, but Kleinsmith knew that Special Operations guys were hunters. Their elder brothers had taken down the hijackers of
Achille Lauro
. Special Ops went places no one else would, or could. And based on the places that seemed to interest these officers most—“Can you track a vehicle moving real-time through the streets of Karachi?” one asked—Kleinsmith got the drift. They were hunting terrorists.