Spy Schools

by Daniel Golden

  Looking back, some of Liu’s former colleagues in the Duke lab feel that he violated their trust. “When you toil away in academia, only about ten people know it’s your idea,” one member of the lab, Jonah Gollub, told me. “Ideas were flowing from here to China. In retrospect, people feel they weren’t given the full picture.”

  The Liu case illustrates how vulnerable academic research is to foreign raiders, and how little universities do to protect it. Eager to attract international students and open branches abroad, universities are reluctant to offend China and other countries by cracking down on research theft. Yet, by looking the other way, they’re betraying the government agencies, and ultimately the American taxpayers, funding the research. We pay taxes for our military to defend us, only to have universities compromise that security by pursuing global prominence without acknowledging or addressing the collateral damage.

  Although Liu has never been charged with any crime, the FBI looked into his activities and briefed university presidents and law enforcement officials about them. At an October 2012 closed-door session of the National Security Higher Education Advisory Board at FBI headquarters, according to an agenda obtained through a public records request, Smith recounted how, “without his knowledge, a Chinese national targeted his lab and … created a mirror institute in China. The episode cost Duke significantly in licensing, patents and royalties and kept Smith from being the first to publish ground-breaking research.” An FBI video interview with Smith about the episode, shown to an invitation-only audience in September 2015, was titled simply “The Theft of a Great Idea.”

  Liu “was definitely filled with intent,” and his actions “could have tremendous economic impact in the future,” Smith wrote me in July 2015. “I think if people understood how something like this happens, and how those with potentially ill intent can take advantage of the natural chaos that occurs in US academic environments, they might become more aware and avoid things like this in the future.”

  * * *

  ACADEMIC RESEARCH OFFERS a valuable, vulnerable, and low-risk target for foreign espionage. Despite pursuing groundbreaking technologies for the Pentagon and the intelligence community, university laboratories are less protected than their corporate counterparts, reflecting a culture oriented toward collaboration and publication. Typically, university researchers aren’t required to sign nondisclosure agreements, which run counter to the ethic of openness.

  “There’s a lot less control than in a company like Boeing,” says John Villasenor, a professor of electrical engineering at the University of California, Los Angeles. “Universities are ripe pickings for anybody who’s interested in accessing intellectual property.”

  Ignorance about intellectual property safeguards, or even hostility toward them, is rife among science students and faculty. There is “zero instruction” on the topic outside law school, Villasenor told me. Significant proportions of UCLA engineering graduate students whom he surveyed couldn’t define a patent (21 percent), copyright (32 percent), trademark (51 percent), or trade secret (68 percent). Never contemplating the possibility of espionage, American professors sometimes comply with requests from acquaintances or strangers overseas for research advice, manuscript reviews, or unpublished data. A civil engineering professor at Penn State once phoned Graham Spanier, then the university’s president, to say that a foreigner had emailed him asking how to build an underground concrete structure that could withstand a megaton explosion.

  “I was about to hit send, when it dawned on me, I’d better ask,” the professor said. “I don’t know this person.” Spanier notified the FBI, which traced the request back through seven intermediary layers before losing the trail. The elaborately disguised source was never unmasked.

  The casual university attitude belies the growing threat. Academic solicitation, or “the use of students, professors, scientists or researchers as collectors,” tripled from 8 percent of all foreign efforts to obtain sensitive or classified information in fiscal 2010 to 24 percent in 2014, according to the Defense Security Service, a Defense Department agency that protects American technology.

  American college graduates with a flair for engineering or computer science typically join high-technology companies, or start their own, rather than continue their educations. As a result, international students dominate graduate programs in those fields at U.S. universities, forming the backbone of their workforce for cutting-edge research. In 2012–13, foreign students earned 56.9 percent of doctorates conferred by U.S. universities in engineering and 52.5 percent of those in computer and information sciences. They comprise more than 70 percent of graduate students nationwide in Smith’s specialty, electrical engineering.

  “Foreign intelligence services, foreign corporations and foreign governments often target these students in an attempt to have them provide the results of the research they are working on or other proprietary and intellectual property that belongs to the United States government or to United States corporations that are funding the research,” David W. Szady, former FBI assistant director for counterintelligence, wrote in a July 2014 newsletter. “Foreign militaries can develop state-of-the-art weapons systems by stealing research from colleges and universities that is sponsored by the United States Department of Defense.”

  * * *

  AMERICAN TAXPAYERS FUND a significant amount of academic research and development. The U.S. government spent $27.4 billion on it in 2014, up from $16.9 billion in 2000 and $9.1 billion in 1990. That includes $2.4 billion in 2014 from the Pentagon and intelligence agencies (not counting the CIA, which doesn’t report expenditures), up from $1.7 billion in 2000 and $1.2 billion in 1990.

  Some of this research is off-limits to foreign students. If it’s classified, only people with security clearances can work on it, usually in secure, off-campus facilities. If it’s export-controlled, the next level down on the secrecy scale, the university must obtain a license from the government for a foreign national to participate. Such licenses are typically denied for students from countries such as China and Iran.

  The bulk of federally funded university research, though, is fundamental, and open to all students. Since it can also be published without restriction, one might wonder, why bother stealing it? The answer: to save time, and avoid mistakes. Access can provide insights beyond the results published in an academic journal. “It’s great to know the solutions, but the process is arguably just as important,” says historian Vince Houghton, curator of the International Spy Museum in Washington, D.C. “You can see the paths not taken, the failures and dead ends.” With a mole in a U.S. university laboratory, researchers overseas can publish and patent an idea first, ahead of the true pioneers, and enjoy the consequent acclaim, funding, and surge in interest from top students and faculty.

  A foreign government may be eager to scoop up a fundamental breakthrough before its applications become so important that it’s labeled secret—and foreign students lose access to it. J. A. Koerner, the former head of counterintelligence in the FBI’s Tampa office, has a term for such promising science: pre-classified. “Once it becomes integrated into a military system, it will be classified and harder to get at,” he says.

  The very openness of U.S. universities denies them recourse against foreigners who siphon American ideas abroad. Economic espionage laws require the owner of stolen trade secrets to have taken reasonable precautions to protect them, like Coca-Cola’s famed vigilance in defense of its formula. Lacking nondisclosure and collaboration agreements to safeguard intellectual property, universities can’t meet that standard.

  Reflecting his prominence at Duke, Professor David Smith has two offices in different buildings that face each other across a well-manicured lawn. He runs both his own research group out of the Fitzpatrick Center for Interdisciplinary Engineering, Medicine and Applied Sciences, which has a stone façade and modern amenities like “smart bridges,” as well as the department of electrical and computer engineering out of older, red-brick Hudson Hall.


  When we talked in April 2016 in the Hudson Hall office, where a remodeling had left the walls and shelves bare, Smith displayed none of the self-importance that one might expect from a department chair and award-winning scientist often touted as a Nobel Prize candidate. Casually dressed, he came across as soft-spoken and unpretentious.

  The case of Ruopeng Liu, his former student turned Chinese billionaire, had taught him how easily the relationship of trust that scientific collaborations depend on can be abused, he told me. “No one has any training in intellectual property,” he said. “It’s something we’re all grappling with—where to draw the line.”

  * * *

  SMITH WAS BORN in 1964 in Okinawa, Japan, where his father served in the U.S. military. When he was a baby, his parents divorced, and he had no more contact with his father. His mother worked odd jobs, and they lived all over California: Riverside, San Diego, Carlsbad, San Francisco, Palm Springs, and finally Escondido, where he spent his last three years of high school.

  He earned his bachelor’s degree in 1988 and his doctorate in 1994, both from the University of California, San Diego. His hobby in graduate school was blackjack. He learned to count cards, a strategy that improves the player’s chances by predicting the value of cards remaining in the deck. “I ended up organizing a group that played blackjack in Las Vegas for a while.” Scammed by dealers using marked decks at a casino on an Indian reservation near San Diego, he sued to recover his losses. After five years of litigation, during which Smith “learned quite a bit about Indian gaming law,” a California appeals court ruled against him.

  Smith was a latecomer to academic stardom. At UCSD, he “was a fairly typical graduate student with no expectations on that level at all,” said David Schurig, who overlapped with him there and later was a researcher in Smith’s lab at Duke. “It’s quite an impressive rise.”

  As a postdoctoral student at UCSD, Smith became involved with a biotech company his adviser was starting, and hardly published any research. He decided to leave academia for industry, but thought that a few publications first would help his career. As it happened, his articles helped launch the field of metamaterials—artificial materials with properties not found in nature—and “I was lucky enough to turn things around.”

  Around 1998, he began collaborating with Sir John Pendry, an English physicist and professor at Imperial College London, who theorized that metamaterials could warp the path of light as it moves through space. At a scientific meeting in San Antonio in 2005, Pendry suggested what he regarded as an “amusing” application. “I said, ‘By the way, we can make something invisible,’” he later recalled. “I just gave a sketch and one slide with a formula on it and I sat down and I expect everybody to laugh. Straight faces.”

  Smith, who had joined Duke’s faculty in 2004, missed the conference, but two members of his lab attended Pendry’s speech. “Pretty soon the phone lines were hot, and David said, ‘We’ve got to build this stuff,’” Pendry recalled.

  “I and my group thought it would be a fun challenge and we realized right away we could do the experiment and the design,” Smith told me. “I never expected that interest in this topic would be so huge.”

  Invisibility has always fascinated humankind. After killing Medusa, Perseus donned an invisibility helmet to elude the Gorgons. Long before Harry Potter and Frodo Baggins, King Arthur and Tom Thumb wore invisibility cloaks, while Gyges, a shepherd featured in Plato’s Republic, sported an invisibility ring to murder a king and seduce the queen.

  Journalists crave invisibility for professional benefit; we long to be unnoticed observers, flies on the wall, albeit flies with notebooks, pens, cameras, and tape recorders. Similarly, the obvious advantages of invisibility in war and espionage have long intrigued strategists. The British Army employed a stage magician and a filmmaker as invisibility consultants in World War II. For a 2002 Wall Street Journal article about the CIA’s resurgence at the Rochester Institute of Technology, I sat in on a meeting where the agency’s chief scientist, John Phillips, suggested projects for college seniors. High on his list was bending light rays to keep a spy in the shadows.

  “Make me invisible,” the six-foot-three, 250-pound Phillips exhorted.

  In June 2006, Pendry, Schurig, and Smith coauthored an article in the prestigious journal Science explaining how to make an invisibility cloak. The following October, in a report published online by Science, those scientists, together with other members of Smith’s lab, unveiled the first successful cloak. Composed of thousands of copper circuits, it “can make light curve around an object, and then emerge just as if it had passed in a straight line,” Smith later wrote. “Think of it like water flowing past a rock in a stream.”

  There was one major caveat. The cloak only concealed objects from microwaves, not the human eye. Since visible light waves are 10,000 times shorter than microwaves, the metamaterials would need to be correspondingly tinier. Such a difference in scale creates a practical difficulty that remains to be solved: the metals used in those smaller metamaterials mostly absorb light, rather than diverting it. Still, the discovery had a range of potential applications, from improving cell phone reception by bending waves around a building or other obstacle, to reducing interference by antennas in military signaling.

  The two Science publications spurred a media blitz. The diffident professor of electrical engineering found himself famous, with his research admired—and, in some quarters, coveted—worldwide.

  “I would never have imagined that someone would believe our group to be so important,” he told me. “I could barely find students to join our group, and finding funding is a nonstop struggle that is never certain. It just would never have seemed a possibility that anyone would seek to obtain intellectual property from us.”

  * * *

  IN AUGUST 2006, as Smith was adjusting to sudden celebrity, he welcomed a graduate student from China into his lab. He had high expectations for Ruopeng Liu. While Chinese candidates, who made up more than 80 percent of the department’s applicant pool, were “very difficult to evaluate,” Liu stood out as an “outstanding prospective student,” Smith recalled.

  The professor and his new pupil had contrasting personalities. While Smith was diffident, thoughtful, and precise, Liu was gregarious, eager, self-confident, and prone to hyperbole. As the youngest member of the group, “he came across as a lovable, bumbling, enthusiastic kid,” Smith said.

  Liu was “a very high-energy guy, almost in a cartoonish sort of way,” another professor recalled. “Pretty outgoing, very friendly, bopping around. Awkward at times, but endearing. When he introduces himself, he’ll bounce into a room. Easy to be around. He was remarkably open to different and new ideas. He was willing to take a crazy idea and see how far it could go.”

  Where Smith was a late bloomer, Liu was something of a prodigy. Born in Shaanxi Province, in northwest China, he moved south at the age of nine to Shenzhen. A fast-growing manufacturing and financial center across a river and a bay from Hong Kong, Shenzhen is home to telecommunications giant ZTE Corporation, where Liu’s parents worked. As a sophomore at Zhejiang University, he encountered the hot new subject of metamaterials. A year later, Tie Jun Cui, one of the leading Chinese scientists in the field, became a visiting professor at Zhejiang, and Liu’s mentor. Liu began taking a train each weekend to work in Cui’s lab at Southeast University in Nanjing. “He conducted his research in my group” for two years, Cui told me. Liu also assembled a group of Zhejiang students to explore mathematical modeling.

  “He’s very energetic and ambitious, and he knows how to organize people,” said Da Huang, a friend of Liu’s at Zhejiang who later joined Smith’s lab. “When he’s focused, he thinks fast.” In his leisure time at Duke, Liu enjoyed grilling meat for potluck dinners and watching a Chinese television drama about Mongol emperor Genghis Khan.

  Liu came to Duke with “a thousand ideas” and wasn’t shy about sharing them, recalled Jonah Gollub, another member of Smith’s
lab. Within a week, Liu organized an hour-long seminar to discuss his research in China. “He came in with this grand unified theory of metamaterials,” Gollub said. “It’s not how science is usually done, but there was the sense he was brilliant. He doesn’t sleep, he’s one hundred percent focused on science, one hundred percent enthusiasm for what he was working on.”

  “Ruopeng was a very unusual graduate student,” another lab member recalled. “He hit the ground running. He was having debates with David Smith and postdocs on day one. That’s something you normally see at the end of the first or second year.”

  When he arrived, Liu was one of only five students in Smith’s group. It would swell to half a dozen graduate students and three or four postdoctoral fellows by 2010. They worked primarily on fundamental research, though occasionally a project was export-controlled. They exchanged ideas at weekly meetings. The lab relied on federal funding, especially from the U.S. military’s research branches.

  As the boss, Smith was supportive but not intrusive. He expected researchers to be independent and solve problems on their own. “At the time, I found David’s approach to be very frustrating, very hands-off,” Gollub said. “You didn’t necessarily feel that you were strongly guided in any direction. If you were struggling, you couldn’t necessarily count on him to notice.”