Balancing Scientific Openness and National Security Controls at the Nuclear Weapons Laboratories (1999)

Balancing Scientific Openness and National Security

There has been extensive discussion in recent months of the possible loss to the People's Republic of China of important nuclear-weapons-related information from certain national laboratories of the Department of Energy (DOE). This aroused concern about the leakage of weapons-related information through interactions of scientists employed by the laboratories with foreign nationals. As a result, moratoria on foreign visits and tighter controls governing interactions with foreigners have been proposed.

To provide an expedited examination of some of the issues surrounding such proposals and policies, in the hope of influencing the current debates as to how to proceed, the National Academies assembled a Committee on Balancing Scientific Openness and National Security. This committee has examined the roles of the national laboratories, the contribution of foreign interactions to the fulfillment of those roles, the risks and benefits of scientific openness¹ in this context, and the merits and liabilities of the specific policies being implemented or proposed with respect to contacts with foreign nationals. Of course, this broad agenda could not be covered in depth in the time available, but the committee did benefit from the prior involvement of its individual members in a wide variety of studies

and activities touching on these topics. The committee also had the benefit of a symposium held at the National Academy of Sciences on August 2–3, 1999, in which a diverse array of knowledgeable observers addressed the issues.² The committee's statement of task, brief biographies of its members, and the agenda for the symposium are attached as appendixes.

The committee reviewed unclassified portions of the reports dealing with the possible losses of weapons-related information from certain DOE national laboratories.³ Although the committee has not reviewed the factual foundations for the allegations of espionage, it starts from the premise that protection of information that relates to the construction of nuclear weapons is of the highest importance to national security. One of DOE's most important responsibilities is to safeguard such information for the protection of the Nation and, indeed, of all mankind. Viewed in this light, it might appear simple and obvious that such information would best be protected by limiting access by foreign nationals to the laboratories at which the information is found and by isolating the scientists and engineers who work on such matters from contact with foreigners. Indeed, the initial proposed reaction to the alleged losses was along these lines. As will be seen, the committee believes that a less sweeping approach is required.

In response to the allegations of the loss of secrets to China, DOE has not sought to bar all foreign contact by laboratory personnel, but has sought to tighten significantly the policies governing interactions by laboratory staff with foreign nationals.⁴ DOE has adopted organizational changes that are intended to give heightened prominence to security and counterintelligence,⁵ revised the order governing access to the laboratories by foreign visitors, provided new guidance governing “exports” of unclassified information through communication with foreign nationals, and undertaken various actions to strengthen the protection of classified and sensitive information. (Documents related to the new and enhanced DOE security policies are listed in Appendix E.) The effects of these initiatives

are still somewhat uncertain because the policies are evolving and the details of implementation are important. This report is intended to set out the committee's view of the conflicting objectives that must be balanced in the development and implementation of such policies.

Although there have now been several studies of possible recent espionage at DOE's laboratories, those of which the committee is aware do not identify scientific exchange with foreign nationals—either visitors or employees—as the source of losses of classified information. Although all possible channels of loss warrant attention, past experience suggests that overly strict action with respect to foreign nationals is neither necessary nor appropriate. Indeed, a response that focuses on foreign nationals may result in a misallocation of effort at best and a highly damaging reaction—without locating the true source of the espionage—at worst. Close and careful evaluation is required as policy is developed and as implementation proceeds.

This report focuses on the three so-called “weapons laboratories”—Los Alamos National Laboratory, Lawrence Livermore National Laboratory, and Sandia National Laboratory—because these laboratories are centrally involved in nuclear weapons programs and are the most directly and immediately affected by the various proposed policies.⁶ These laboratories received some 6,398 foreign visitors or assignees (individuals who stay from 30 days to 2 years) in 1998, including approximately 1,824 visitors from sensitive countries (countries that are seen to present proliferation or national security risks). In addition, employees from the laboratories travel to foreign countries to engage in scientific meetings, including travel to sensitive countries, and interact with foreigners at foreign laboratories or at scientific conferences. In 1998, weapons laboratory personnel engaged in approximately 5,799 trips to foreign nations, including 1,814 trips to sensitive countries, on official business.⁷

The weapons laboratories engage in work in three principal areas: weapons-related research, open scientific research, and cooperative programs with other nations. Most of this work—perhaps 70 to 75 percent —is unclassified and the performance of all the work requires laboratory personnel to draw on contributions from the outside scientific world. ⁸ Indeed, because DOE's laboratories conduct only 1 to 2 percent of the world's research and development, the effectiveness of the laboratories depends substantially on the capacity to access and apply the 98 to 99 percent of the work that is performed elsewhere.⁹

The conduct of weapons-related research is a central mission of the three laboratories. The nature of the research, however, has changed significantly in recent years as a result of the end of the Cold War, the Comprehensive Test Ban Treaty,¹⁰ the termination of the development of new physics packages (i.e., the nuclear components) for nuclear weapons, and the broadening responsibilities of the laboratories for nonproliferation programs. In the absence of testing, the United States has launched a science-based Stockpile Stewardship Program. The program relies centrally on the development and maintenance of the scientific skills and tools at the laboratories to assure the dependability and safety of nuclear weapons. Without undertaking the detonation with nuclear yield of weapons in tests, computer modeling and other tools are used to understand better the aging of materials and the complexities of weapons performance. Although much of this work is and must remain classified, weapons researchers must also maintain contact with related fields of open scientific research. Open communication in rapidly moving technical fields works to the advantage of the United States if it has the ability to exploit new ideas quickly regardless of the source of those ideas. For example, cutting-edge weapons-related research is heavily dependent on the broader, open research in materials, nuclear physics, computer science, hydro-dynamics, lasers, and many other fields. Because of the international nature of science, this necessitates substantial international engagement by weapons scientists.

The laboratories also engage in open and broad-ranging scientific research in a wide variety of fields, both fundamental and applied. In the latter category, for example, the weapons laboratories contribute to the program of fusion energy research that is aimed at harnessing nuclear fusion as a commercially viable

energy source.¹¹ This research is of great importance because fusion offers the possibility of a virtually inexhaustible fuel supply, in combination with favorable safety, waste management, and nonproliferation attributes, as well as freedom from the air pollution and climate change risks associated with fossil fuels. The research budgets of the Europeans and the Japanese in fusion energy exceed the U.S. budget and even the Russians have maintained a significant program. In light of the substantial programs that exist outside the United States, the case for meaningful engagement of the United States in international fusion activities is overwhelming.¹² The United States, as well as its international partners, gains by working to solve such common problems collectively.

The laboratories also carry out cooperative programs that serve U.S. national security interests. For example, the United States has embarked on a cooperative program with the Russians to upgrade the protection, control, and accounting of weapons-grade nuclear material in Russia. ¹³ This program is in the direct and immediate national security interest of the United States because of the risks that would arise if even a relatively small amount of the large and growing stocks of this material were to become available to a proliferant nation or a terrorist group.¹⁴ The knowledge and skills necessary to carry out this program rest in the national laboratories and their involvement in the program is central to the program's success. Their participation, however, necessarily requires the laboratory staff to travel to sensitive countries and to engage in discussions with relevant counterparts in those countries, as well as to host visits by their counterparts. ¹⁵

There are other aspects of international engagement by the laboratories that should also be taken into account. First, a spirit of reciprocity, transparency, and cooperation with scientists and technologists from sensitive countries is essential if the United States is to implement and verify nuclear arms reductions agreements that are in its national security interest. Some access, albeit carefully

controlled access, to each other's weapons-related facilities is a necessary step in providing confidence in the arms control process. If the United States denies even controlled access to its weapons laboratories or staff, it should anticipate denial of U.S. access to facilities in sensitive countries. Such a step would constitute an unfortunate retreat in arms control.

Second, the weapons laboratories must attract and retain first-rate scientific and engineering talent. One of the biggest challenges in the science-based Stockpile Stewardship Program is the maintenance of a cadre of talented scientists and technologists to deal with the potential changes in the demands on the nation's nuclear-weapons-related skills in an unpredictable world. The laboratories have recruited and continue to recruit such personnel by providing them the opportunity to engage in open scientific work that is related to the maintenance of capabilities in classified weapons work. The ability to attract and maintain such talent will be compromised if the open work is overly constrained; a life “behind the fence” is unattractive, and less productive, for most of the front-rank researchers that the laboratories need to employ.

Finally, openness serves not only scientific advance, but also policy formation in the national security sphere. Secrecy can make it difficult for policy makers of all countries and the publics to whom they are accountable to know what they need to know in order to make sensible judgments on important public matters bearing on national security. For example, open international discussion of the usability of reactor-grade plutonium in nuclear weapons has been essential to effective policy-making on the management of plutonium for both military and civilian use worldwide. There is a fundamental tension between secrecy and openness in such matters. The basic point remains, however, that there are costs in extending the scope of secrecy too extensively.

The committee concludes that there are many aspects of the work at the laboratories that benefit from or even demand the opportunity for foreign interactions. Thus, the establishment of a sensible policy should be guided by a net assessment of the risks and benefits that such interactions allow.

The committee has not had access collectively to the classified assessments and thus cannot offer an independent judgment of the magnitude of the risks presented through interactions by laboratory personnel with foreign nationals. It is certainly plausible that foreign nations might seek to exploit visits to the laboratories or contacts with laboratory personnel as a means of obtaining classified information. Indeed, this is an information age and protections against espionage must remain an important element of U.S. national security in the years ahead. The apparent absence of any significant losses of classified information in the recent past either by weapons-lab personnel traveling abroad or through espionage by foreign visitors suggests, however, that the existing procedures gov-

erning such foreign interactions, although needing to be reexamined, have been generally adequate. Of course, continued vigilance is needed to assure adequate protection of weapons-related information.

In striking a balance, it must also be recognized, as explained above, that significant curtailment of foreign interactions by laboratory staff will impose substantial costs. The committee concludes that U.S. national security will not be served by severely restricting international interactions at the laboratories in an effort to hide secrets more carefully. Rather, national security can be better assured by a system that seeks to avoid losses, but that also enhances the capacity to advance knowledge in relevant fields and to achieve the other benefits of openness through international contacts by laboratory personnel. In short, a balanced policy should not only allow, but also facilitate the opportunity for foreign dialogue in appropriate areas.

If viewed properly, the protection of weapons-related secrets and the need for openness are not necessarily in conflict. Rather, they should be viewed as two aspects of a proper national security strategy. The weapons-related matters of espionage interest primarily relate to the detailed designs and the technological know-how that enable the fabrication of items of importance to national security. The sphere of information of principal concern to national security does not overlap with the areas in which the maintenance of openness is essential; hence, a proper policy can and should accommodate both tight secrecy in some areas and openness in others. Key to a proper balance is a clear and precise definition of the information that is to be sequestered and education of laboratory staff so as to assure that it is knowledgeable of the boundaries surrounding that information. ¹⁶

The need for such an approach is a common element of the modern technological world in other contexts. For example, microelectronics companies perceive it is in their interest to share information, through full involvement in open scientific dialogue, in connection with research on the properties of materials, while simultaneously maintaining tight secrecy with regard to the design and means of fabrication of a particular microelectronic device. Similarly, biotechnology companies may encourage their technical staff to publish in the open scientific literature about scientific advances, while maintaining tight secrecy about particular products that are under development. It is now common practice in the industrial world to leverage technical capability by stimulating and participating in the scientific contributions of the general scientific community, while exercising strict control over the specific integrative and “know-how” capability. This same model applies to the work of the weapons laboratories.

The committee draws a number of conclusions for the implementation of policy.

1. Maintain balance. Policy governing international dialogue by laboratory staff should seek to encourage international engagement in some areas, while tightly controlling it in others. It is through the encouragement of both aspects of policy that U.S. national security will be best advanced. Indeed, there is a very real danger that misdirected security crackdowns could have widely pernicious effects by inhibiting international interactions that serve U.S. national security interests, as well as research and development in general.

   Overly strict constraints on foreign interactions could affect not only foreign visitors and international meetings at the weapons laboratories, but also visits and meetings taking place at other U.S. institutions where weapons-lab personnel and visitors from sensitive countries might come into contact. Moreover, constraints on international meetings in this country or on travel abroad could affect not only the full-time employees of the weapons labs, but also a much wider array of security-cleared consultants and advisors to the laboratories and the various branches of government. Of course, some constraints already exist and have for years. The danger that concerns this committee is the possibility that the mechanisms to constrain interactions may be strengthened in ways that could cause significant damage in exchange for small benefits.¹⁷

2. Educate staff. Security procedures should be clear, easy to follow, and serve an understandable purpose. The key ingredient is the development of an awareness and appreciation throughout the laboratory staff of the nature of the threats and of the methods that should be used to prevent espionage. Education and training are essential. Top-level scientific and program management should develop a definition of the boundaries of acceptable interactions with foreign experts and should assure understanding of those boundaries by the technical personnel. Prescriptive rules are not an effective substitute for informed personal vigilance.

3. Streamline procedures. Good science is compatible with good security if there is intelligent line management both at the labs and in Washington, which applies effective tools for security in a sensible fashion. It is important in this context to put in place procedures that establish needed security protections, while minimizing interference with work. For example, the DOE policy governing foreign visitors requires that the identities of certain visitors be checked with various intelligence services so as to enable an informed decision as to whether or how to allow visits to occur.¹⁸ Such a system should be implemented so that these so-called “indices checks” can be performed without undue delay. (The committee understands that the relevant DOE office aspires to a 10-day turn-around for such checks, but that the delays have been far longer in recent months.) Efforts should be made to assure that security procedures serve their purpose, but do not become bottlenecks. If the procedures sweep too widely or are too cumbersome, the achievement of compliance will be compromised.

4. Focus efforts. DOE should focus its efforts governing tightened security for information. The greatest attention should obviously be provided to the protection of classified information by appropriate physical and cybersecurity measures, and by personnel procedures and training. In response to the recent reports, DOE has revised its policies governing disclosures of “sensitive but unclassified” subjects, such as controls on exports of technical data through communication with foreign nationals (see Appendix E). Although the concrete effects of the revised guidance are not yet fully apparent, the directives in this area are creating significant ambiguity and can intrude on communications that are far distant from DOE' s security mission. As a result, the policy will have a detrimental effect on communications that serve U.S. national security interests. Instead of creating new control regimes, DOE should rely on the control systems that are already in place across the government —such as the controls on classified information and on export of certain technical data. The protection of truly important information is likely to be more effective if DOE devotes attention to the information of central importance—if it builds high fences around narrow areas —rather than by allowing the effort to be diffused and diluted by encompassing unclassified information of marginal (or non-existent) significance to national security.

   Physical security requirements should also be tailored to the level of risk presented by the specific location at which the work is performed. As DOE has recognized, its university-like laboratories (e.g., Fermi National Accelerator Laboratory) conduct no classified research and, as a result, should not be subject to requirements relating to military security. Even at the weapons laboratories, heightened levels of protection should be directed at those areas in which classi

   18	U.S. Department of Energy, Unclassified Foreign Visits and Assignments, DOE N 142.1, July 13, 1999, paragraph 4.d.

fied work is conducted, with more limited security elsewhere. In short, there should be a graded regime that follows the basic principle that the level of security should be tailored to the magnitude of the risks.

4. Beware of prejudice against foreigners. Over the past half-century foreign-born individuals have contributed broadly and profoundly to national security through their work at the national laboratories. During World War II, the foreign-born contributors to U.S. war efforts included many giants of science —Enrico Fermi, Eugene Wigner, Leo Szilard, Edward Teller, Hans Bethe, John Von Neumann, Stanislaw Ulam, and others. Many of the current employees of the laboratories are foreign-born and some are foreign nationals, including individuals from sensitive countries. These persons are making very significant contributions to the laboratories ' work.¹⁹

   To the committee's knowledge, no such employees have ever been identified as the source of losses of classified information in recent decades.²⁰ Efforts to prevent losses of important weapons-related information cannot fairly be directed at individuals of any particular national or ethnic background. The most notorious recent spies—for example, Aldrich Ames and the Walker family—were native-born U.S. citizens. Under the circumstances, there can be no justification for focusing counterintelligence and security efforts on those individuals who happen to be foreign nationals or who are of any particular national or ethnic background.

   Moreover, such a focus not only would be unfair, but also would be counter-productive. Approximately 52 percent of all doctoral students in the U.S. science and engineering programs are non-U.S. citizens, with many from China and India (both identified as sensitive countries). ²¹ To limit access to the laboratories by such researchers or to engage in practices that are less than fully welcoming of

them would deprive the laboratories of a major source of scientific and technical talent, with adverse consequences both for the laboratories and for U.S. national security.

We live in an age in which information is of central importance to U.S. national security; the United States has know-how and capabilities related to weapons that must be safeguarded carefully. We should expect that espionage aimed at obtaining this information will be attempted. But this does not mean that we should terminate all channels of international information flow through which weapons-related information might arguably leak. The world is awash in scientific discoveries and technological innovations. If the United States is to remain the world's technological leader, it must remain deeply engaged in international dialogue, despite the possibility of the illicit loss of information. Furthermore, international cooperation and communication related to nuclear arms control, nuclear nonproliferation, and the protection of nuclear materials bring substantial direct benefits to U.S. national security. As a result, the objective of U.S. policy should be to assure that the gains from international networking always offset the losses. In the case of the national laboratories, this is best assured by a policy of informed protection of truly important information, while preserving openness in other areas.