Raymond A. Zilinskas, Ph.D., is director of the Chemical and Biological Weapons Nonproliferation Program at the Center for Nonproliferation Studies, Monterey Institute of International Studies, Monterey, CA.

Jonathan B. Tucker, Ph.D., is a policy analyst specializing in chemical and biological weapons proliferation and control. He is a 2002–03 senior fellow at the U.S. Institute of Peace in Washington, DC, on leave from the Monterey Institute’s Center for Nonproliferation Studies. Before joining the center in 1996, he served for six years in U.S. government positions at the Department of State, the Congressional Office of Technology Assessment, and the Arms Control and Disarmament Agency. In 1995, he was a UN biological weapons inspector in Iraq. He holds a B.S. in biology from Yale and a Ph.D. in political science from the Massachusetts Institute of Technology. In addition to numerous papers and reports, he is the author of Scourge: The Once and Future Threat of Smallpox and the editor of Toxic Terror: Assessing Terrorist Use of Chemical and Biological Weapons.

A report on the “Workshop on Guidelines for the Publication of Scientific Research Potentially Related to Biological and Toxin Warfare,” Washington, DC, 12 August 2002.¹

Summary

Some open scientific publications in microbiology and related fields contain information potentially relevant to the development, production, formulation, or delivery of biological weapons. A workshop held by the Monterey Institute’s Center for Nonproliferation Studies in August 2002 discussed possible approaches to minimize the risk that such “sensitive” research findings could be misused for biological warfare or terrorism. Most participants agreed that to limit damage to the scientific enterprise, restrictions on the dissemination of unclassified research results, if any, should be limited to the small proportion of findings directly relevant to malicious use. Moreover, the controls must not be so onerous as to encourage circumvention or cause scientists to abandon key areas of research. One proposal put before the workshop participants would consider placing restrictions on research that involves a Select Agent (as defined by the U.S. Centers for Disease Control and Prevention) and that aims to achieve one or more of six weaponization²-related goals:

1. Enhance pathogen infectivity, pathogenicity, antibiotic resistance, or resistance to host immunological defenses
2. Improve the ability of a microbial pathogen to remain viable and virulent during prolonged storage and/or after release into the environment
3. Facilitate the dissemination of biological agents as a fine-particle aerosol
4. Facilitate the dissemination of a biological agent by contamination of food or water sources
5. Create a novel pathogen or one with characteristics that have been altered to evade current detection methods or host immune defenses
6. Assemble oligonucleotides to synthesize the genome of a pathogenic microorganism

Participants identified significant problems with the proposed regulatory system, however, including issues that would have to be addressed by any scheme for controlling research or the information it produced. To be workable, a system for restricting publication of “sensitive” information requires the broad support of the international scientific community, which must perceive that the security benefits of restricting open publication outweigh the possible costs to science.

Introduction

Since the days of World War II and the Cold War, security analysts and policy makers have worried that information published in the open scientific literature could benefit military adversaries. More recently, research in the biosciences has become a focus of concern because of the dramatic advances in this field and their potential application to biological warfare and bioterrorism. Security experts fear that sophisticated terrorist groups or national biological warfare programs could exploit advances in microbial pathophysiology and genomics to develop advanced biological warfare agents and delivery systems capable of inflicting thousands of casualties. (Low-technology biological weapons, such as foodborne pathogens and toxins, are likely to inflict much lower levels of damage.) Because an advanced biological warfare capability requires extensive research and development, laboratory experimentation, and field-testing, it would most likely be pursued by states of proliferation concern or by wealthy terrorist organizations (such as the Japanese group Aum Shinrikyo) that can recruit skilled scientific and engineering teams.

Until fairly recently, information on converting pathogens and toxins into biological warfare agents, and on methods for their dispersion, was largely unavailable. Over the past decade, however, a Pandora’s box of bioweapons-related data has been opened. Technically trained individuals, including antigovernment activists on the far left and right of the political spectrum, have published lethal “cookbooks” such as Silent Death, which contain step-by-step instructions for the production and delivery of pathogens, natural poisons, and toxic chemicals.³ In addition, the U.S. government, as part of its routine declassification program, released a number of studies prepared by the pre-1969 offensive biological warfare program that contain sensitive information on the weaponization and dispersal of various agents.⁴

Since early 2002, the U.S. government has attempted to withdraw many of these reports from public access, but it may be too late to prevent the further dissemination of previously released data.⁵

In addition to these sources of sensitive information, a growing number of papers published in the open scientific literature contain findings from basic or applied research in microbiology and biomedicine that was conducted for legitimate purposes but may be relevant to the development, production, formulation, or delivery of biological weapons. The U.S. Department of Defense (DoD) is concerned that advances in biological research could be misused by those seeking to develop biological weapons. At the same time, DoD recognizes that constraints on such research could hamper the development of defenses against biological weapons, interfere with other useful applications, and impede the progress of science.

To help DoD better understand the costs and benefits of policies intended to address the weapons implications of biological research, the Monterey Institute of International Studies’ Center for Nonproliferation Studies held a workshop on 12 August 2002 with the primary aim of developing guidelines that DoD might use when reviewing proposed bioscience research projects and sensitive findings generated by the projects it supports. As the workshop proceeded, it became clear that for any guidelines to be effective, they would have to be applied more broadly, potentially by all funders of bioscience research. The guidelines would also have to be perceived as reasonable and fair so that the scientific community would accept them voluntarily. First Amendment issues, to the extent that they apply to publications that researchers produce after accepting government funds, were deemed not pertinent to the narrow scope of the workshop.

Although not a consensus report, this article seeks to convey a sense of the discussions that took place during the Center for Nonproliferation Studies workshop and the general understandings gained by the participants concerning possible restrictions on the publication of sensitive research. The moderators kept the discussion focused on broad concepts rather than detailed procedures to avoid getting bogged down in minutiae. Instead, the primary aim was to identify policy approaches that offer a reasonable balance between enhanced security and scientific freedom.

This report covers the following topics:

1. Assessing the malicious uses to which sensitive information might be applied
2. Historical precedents for regulating scientific information
3. The current debate over whether to restrict research in the biosciences
4. The feasibility of such restrictions
5. Proposed regulatory guidelines for those who review proposals and draft publications
6. Brief thoughts on restrictions of sensitive findings
7. Critiques by workshop participants of the proposed guidelines
8. Some concluding ideas on what the workshop accomplished and possible next steps

Assessing the Threat of Misuse

Individuals or groups seeking to acquire biological weapons in violation of international law are unlikely to advertise their illicit activities in open scientific publications. It is therefore safe to assume that research papers submitted to scientific journals are intended for peaceful purposes and that the authors themselves are aboveboard. Accordingly, policies to restrict the dissemination of sensitive scientific information should not primarily aim to prevent deliberate abuses by academic scientists, but rather to make it more difficult for terrorist groups and states of proliferation concern to acquire techniques and data that could be used to enhance their biological warfare capabilities.

Lone terrorists or organizations with rudimentary scientific expertise do not present the main concern with respect to the misuse of basic research findings, because they are unlikely to peruse the scientific literature in search of recipes for biological weapons. The most serious threat arises from scientists working in sophisticated, well-funded national biological warfare programs or in affluent terrorist organizations such as Aum Shinrikyo, which recruited skilled scientists and engineers to produce biological weapons. Such experts do read the scientific literature and are capable of funneling basic research findings on molecular pathogenesis into efforts to “improve” existing pathogens or to pursue new lines of weapons-related development.

A major challenge facing those who seek to prevent the proliferation of biological weapons is the dual-use nature of applied microbiology and biotechnology. The same scientific knowledge, technical know-how, and equipment involved in the peaceful development and production of commercial products, such as food additives, biopesticides, biopharmaceuticals, and vaccines, could be misused by those seeking to acquire biological weapons. For example, because the growth characteristics of the harmless bacterium Bacillus thuringiensis, an agricultural biopesticide, are nearly identical to those of Bacillus anthracis, the causative agent of anthrax, an industrial production line for the pesticide could easily be diverted to manufacture anthrax. Moreover, methods used to develop attenuated live-virus vaccines could be applied to develop more virulent viral strains for weapons.⁶

For these reasons, understanding the intent behind a research activity may be the only way for an outsider to determine whether it is being undertaken for peaceful or illicit purposes.

This dual-use dilemma spills over into the scientific literature. When research papers are published that elucidate the infectious process, describe the molecular basis of pathogenesis, or elucidate the action of toxins on physiological systems, they add to the general body of scientific knowledge and facilitate the development of new means of medical prevention and therapy. Yet bioweaponeers may also access this information and use it to devise more deadly biological weapons. Examples of scientific research that pose dual-use dilemmas are

• Sequencing the genomes of human pathogens. This achievement is leading to new therapies against infectious disease, yet it could also facilitate the development of novel biological warfare agents. According to one assessment, “The ever-expanding microbial genome databases now provide a parts list of all potential genes involved in pathogenicity and virulence, adhesion and colonization of host cells, immune-response evasion and antibiotic resistance, from which to pick and choose the most lethal combinations.”⁷
• Construction of “fusion toxins” derived from two distinct toxins. This technique, developed to kill cancer cells, could be redirected to create toxins that target the normal cells of almost any human tissue for destruction.
• Genetic engineering of a Bacillus anthracis strain containing inserted toxin genes (for example, for cereolysine AB). This approach could be used to render anthrax bacteria resistant to existing protective vaccines.⁸
• The finding by Australian researchers that inserting the gene for interleukin-4 into the genome of ectromelia (mousepox) virus significantly enhances viral virulence and vaccine resistance in mice. Some observers fear that applying the same procedure to poxviruses that infect humans could result in vaccine-resistant strains.
• Development of “stealth” viruses that evade the human immune system. These viruses serve as molecular vectors to introduce curative genes into patients with inherited diseases, but they could also be misused to develop more effective viral weapons.⁹
• Publication of the molecular details of two virulent strains of influenza, the 1997 Hong Kong flu and the 1918 Spanish flu, the second of which killed 20 million to 40 million people worldwide. Such information could facilitate the deliberate engineering of lethal variants of influenza virus.¹⁰
• Generation of influenza A virus from cloned DNA segments. This new technique makes it feasible to generate attenuated live virus vaccines, but it might also be misused to engineer a variety of novel RNA viruses.
• Genetic engineering of the tobacco plant to produce subunits of cholera toxin. The same method could make it possible to produce large quantities of protein toxins cheaply and relatively easily.
• Studies of viral proteins that are similar to mammalian proteins, as tools to probe their function. One recent study revealed that a protein made by the variola (smallpox) virus inhibits a component of the human immune system more effectively than a similar protein produced by the vaccinia (smallpox vaccine) virus. This finding may partially explain why variola is a human pathogen, whereas vaccinia is not.
• Aerosol spray drug-delivery systems. Such systems have been developed to replace insulin injections for diabetes¹¹ and for intranasal delivery of a live virus vaccine against influenza. However, aerosolization is also the most efficient way of disseminating biological warfare agents.
• Synthesis of infectious poliovirus by assembling custom DNA strands ordered from a commercial biotechnology company.¹² Similar techniques might eventually be used to recreate more deadly viruses, such as Ebola or the 1918 strain of influenza, in vitro.¹³

Until recently, few if any editors of scientific journals considered whether research findings such as these could be misused. When the mousepox study was submitted to the Journal of Virology, the draft article made no mention of its potential relevance to biological warfare, and neither the journal editor nor two independent reviewers raised questions about its broader implications.¹⁴ In the aftermath of the 11 September terrorist attacks and the anthrax-tainted letters, public awareness and concern about bioterrorism have increased dramatically. As a result, proposed restrictions on the movements of foreign scientists and the publication of sensitive information have returned to the national agenda after a long hiatus.¹⁵ Several newly adopted or proposed regulations would affect the conduct of basic research:

• In February 2002, DoD circulated a draft directive proposing the establishment of new categories of controlled information and requiring certain DoD-funded scientists to submit publications for prior review. After strong objections from the scientific community, the draft directive was withdrawn and is being revised.
• On 12 April 2002, the Department of Agriculture announced that it would no longer sponsor visas for foreign scientists.
• On 7 May 2002, the White House Office of Science and Technology Policy declared its intention to implement “an enhanced mechanism for visa review” in sensitive areas of science and technology that are “uniquely available in the United States,” with the aim of identifying individuals who would use this knowledge to harm the United States or its allies. An Interagency Panel on Advanced Science and Security, chaired by officials from the Departments of State and Justice, will review visas case by case.
• On 26 July 2002, Rep. Dave Weldon introduced House Resolution 514, which expresses concern that Science published a paper describing the in vitro synthesis of infectious human poliovirus. The draft resolution, which was not reported out of committee, calls on scientific publishers and editors “to establish ethical standards to ensure that published material does not aid terrorists in the development of agents of bioterrorism.” It further requests the Executive Branch to “examine all policies, including national security directives, relevant to the classification or publication of federally funded research to ensure that … information that could be useful in the development of chemical, biological, or nuclear weapons is not made accessible to terrorists or countries of proliferation concern.”
• In August 2002, the White House Office of Management and Budget announced its intention to develop rules for “the discussion and publication of information that could prove sensitive to national security.” This new category of information would be unclassified but could not be publicly disseminated. The planned rules would apply only to research conducted by government scientists or under contract and would not cover research performed under government grants, the mechanism used for most university research. Nevertheless, this development clearly has implications for the effort to define sensitive information contained in unclassified research papers.

Historical Precedents for Regulating Scientific Information

The possibility of government-imposed restraints on the communication of scientific information is a matter of great concern to the academic research community. What are the historical precedents for this type of policy?

The Corson Report. In the early 1980s, Cold War tensions led the U.S. government to impose tight controls on foreign researchers and to limit the publication of certain types of basic and applied research by American scientists. At that time, it was feared that open publications in the fields of computer science, microelectronics, communications, and electro-optical sensors might be exploited for military purposes by the Soviet Union and its Warsaw Pact allies. In 1982, a U.S. National Academy of Sciences committee addressed this issue in a report titled Scientific Communication and National Security, better known as the Corson Report after its chairman, Dale R. Corson. The report concluded that the vast majority of university research should be unclassified, that some sensitive research might warrant formal security classification, and that a “narrow gray area” existed in which “limited restrictions short of classification” were warranted. Based on the findings of the Corson Report, President Ronald Reagan issued National Security Decision Directive 189 of 21 September 1985, which concluded: “No restrictions may be placed upon the conduct or reporting of federally-funded fundamental research that has not received national security classification, except as provided in applicable U.S. Statutes.” It is unclear whether this policy, or the Corson Report recommendations, can be generalized beyond their Cold War origins and primary applicability to the physical sciences and technology.

Classified Research at Universities. The U.S. government has long classified information that bears directly on national security, including technical data generated by universities performing government contract research. Since the 1960s, most U.S. universities have banned classified research on campus, but some have established subsidiary institutions that seek out and perform classified research contracts for the federal government. For example, in 1951 the Massachusetts Institute of Technology founded a second campus, Lincoln Laboratory, for classified research,¹⁶ and Johns Hopkins University operates a similar organization called the Applied Physics Laboratory. Many faculty members at these institutions hold joint appointments with academic departments at their home universities and accept the associated restrictions on their ability to publish contract work in the open literature. In most cases, however, the activities subject to classification or other publication restrictions do not involve basic research but rather applied research or technology development.

Secret Patents. The main purpose of the patent system is to give inventors exclusive control over commercialization of their intellectual property for a fixed period while permitting public disclosure. This arrangement preserves the economic incentive for an inventor to bring a discovery to fruition while allowing others to improve on the invention and avoid wasting time repeating work that has already been done. Although bioscientists may voluntarily delay publishing research results in order to file a patent application, the decision not to publish is made by the scientists themselves or the institutions they work for, not by personnel of the funding agency or by U.S. government officials.

Nevertheless, the Invention and Secrecy Act of 1952 (35 U.S.C. 181) empowers the U.S. Patent and Trademark Office to prevent the release of information contained in a patent on security grounds. The statute specifies that “whenever publication or disclosure by the grant of a patent on an invention in which the Government has a property interest might, in the opinion of the head of the interested Government agency, be detrimental to the national security, the Commissioner upon being so notified shall order that the invention be kept secret and shall withhold the grant of a patent therefor.” In 1977, the Patent and Trademark Office invoked this law to prevent the disclosure of a patent on an encryption device invented at the University of Wisconsin.¹⁷ Although the Patent and Trademark Office has developed guidelines for examiners to use when reviewing patent applications for their national security implications, the agency has declined to disclose the guidelines. In any event, because the guidelines were developed in 1972, when modern biotechnology was still in its infancy, they are probably outdated.

The Current Debate

Biological scientists and the professional associations that represent them generally argue that the government should not restrict the dissemination of unclassified fundamental research, even when the findings of the research are potentially hazardous. Other authorities counter that national security restrictions on the publication of sensitive research are warranted in cases where such information could be directly relevant to the development or production of biological and toxin weapons.¹⁸ Such restrictions, they argue, would not be merely a political gesture but would offer tangible benefits for national security. The main arguments laid out by the two sides are as follows:

Arguments Against Restricting Publication of “Sensitive” Research

• Restricting the publication of valid research by creating a gray area between open and classified information would run counter to the norms of the scientific community and inevitably hinder valuable advances in the biosciences, including those that offer therapeutic benefits.¹⁹ Moreover, restrictions on publication would deter academic researchers from making long-term intellectual investments in research areas that could be subject to censorship.²⁰ The risk of forgoing scientific advances may be greater than the risk that sensitive information could fall into the wrong hands.
• Controls on scientific communications would create a false sense of security and hamper research needed to develop effective countermeasures against bioterrorism. For example, the discovery that poxviruses can be genetically modified to become more virulent and vaccine-resistant has alerted the scientific community to the need to develop antiviral drugs to counter future engineered strains.
• Embargoing sensitive information would delay its spread only temporarily. Whereas man-made plans or designs can be kept secret indefinitely, scientific information about facts of nature is capable of being rediscovered at any time by investigators working outside the regulatory regime.
• Numerous technical and practical hurdles stand between the information generated by basic research, such as the DNA sequences of pathogens, and the ability of bioweaponeers to develop novel biological warfare agents.
• Eliminating the “materials and methods” sections of scientific papers to protect sensitive information would prevent scientists from replicating the claimed results to prove or disprove their validity. This step would undermine the empirical basis of scientific progress by making it impossible for scientists to validate the work of their colleagues.
• Controls on scientific communication could include restrictions on patents, preventing pharmaceutical companies from learning of intellectual property relevant to drug development. For example, if the discovery of a microbial virulence factor remains secret, drug companies could not develop a drug to disable it. The ultimate losers would be patients.

Arguments for Restricting Publication of “Sensitive” Research

• Scientific knowledge is not inherently beneficial, and some types of information may be inherently dangerous. According to bioethicist Arthur Caplan, “We have to get away from the ethos that knowledge is good, knowledge should be publicly available, that information will liberate us.… Information will kill us in the techno-terrorist age.”²¹
• Scientific advances published in the open scientific literature could facilitate the development of new types of biological weapons that may not be countered effectively by defensive applications of the same research. Novel biological weapons might be developed rapidly and at relatively low cost, whereas therapies, preventatives, and other countermeasures—even if they prove feasible—would probably take several years and considerable resources to develop and deploy. Moreover, defenses cannot be developed for all conceivable biological warfare threats. The inefficiencies imposed by restricting the dissemination of some types of scientific research may be a price worth paying if the restrictions hamper offensive applications more than defensive ones.
• As with military deterrence, the benefits of restricting sensitive scientific communication may be underestimated because of the difficulty of demonstrating a clear cause-and-effect relationship between controls on sensitive information and the non-occurrence of a bioterrorist event.
• Scientific research is already subject to numerous ethical guidelines with respect to animal welfare, human subject research, recombinant DNA guidelines, and other professional norms. Because of concerns over the possible hazards of recombinant DNA technologies, the 1975 Asilomar Conference established a mechanism for regulating certain types of potentially dangerous research.²² Voluntary adoption of these guidelines by the scientific community led to the creation in 1976 of a scientific review board at the National Institutes of Health known as the Recombinant DNA Advisory Committee. The Asilomar and committee guidelines, however, govern the conduct of research rather than restricting its subsequent publication. Moreover, they are directed at accidental or unanticipated consequences of research, not at intentionally malicious applications of the knowledge that research generates.
• Nearly all journals have a written requirement that data and methodologies that support the author’s interpretations must be published. Recently, however, several microbiologists have voluntarily requested that journal editors withhold key pieces of information from the “materials and methods” sections of their papers because of concerns about giving away sensitive details to terrorists.²³ For example, publishing the DNA sequences of genetic probes used to identify B. anthracis and other pathogens might enable bioweapons scientists to develop non-detectable strains.

Feasibility of Restricting Scientific Communications

The original aim of the Center for Nonproliferation Studies workshop was to suggest guidelines that DoD could consider for reviewing research contracts that it supports in the field of microbiology. After some initial discussion, it became clear that the issue affects other government funding agencies as well. Had the purview of the study been limited to applied biodefense research, it might have made sense to develop a set of guidelines for DoD alone. But because a broad range of scientific research funded by both military and civilian agencies could be misused, workshop participants decided to broaden the focus to cover all government-funded research in microbiology and related fields.

Reflecting this new scope, workshop participants agreed that if research is to be assessed for its potential national security implications, the review should take place not at the end of the process, when a manuscript has been submitted for publication, but rather at the beginning, when the project proposal is being considered for funding by a government agency. In addition, the agency could require as a condition of funding that any papers resulting from the research project undergo security review prior to publication. At present, such conditions can legally be imposed on research contracts but not on grants, the primary form of support for university research. Under a possible future system of security review, however, researchers and journal editors might agree in advance to respect recommendations to keep “sensitive” data out of open circulation.

The Center for Nonproliferation Studies workshop sought to define some notional guidelines for (a) the security review of research proposals and (b) restricting the distribution of sensitive research results so that only bona fide scientists are granted access to it. Both sets of guidelines must be judged workable and effective if they are to be accepted voluntarily by the scientific community. At the same time, a degree of flexibility will be needed on the part of government funding agencies and research institutions to craft an information regime that achieves a reasonable balance between scientific freedom and national security.

Guidelines for the Security Review of Research Proposals

It was the sense of workshop participants that guidelines for reviewers of scientific research proposals should focus on the small subset of projects and findings that are directly relevant to the military or terrorist use of pathogens and toxins. Restrictions on the publication of sensitive research could not be implemented if the definition of “sensitive” is either so broad as to be unmanageable or so narrow as to be meaningless because too little information is captured. Any proposed restrictions should focus exclusively on methods or data that facilitate the development, formulation, and delivery of “advanced” biological weapons by technically sophisticated terrorists or states of proliferation concern. The guidelines must also be clear and unambiguous so that they can be applied in essentially the same way by grant administrators in various government agencies and, indeed, in different countries. Otherwise, the security review process would appear inconsistent and arbitrary.

It is also important to consider what fraction of the scientific literature would be captured by the proposed security review criteria. Ideally, one would want to restrict only a small fraction of scientific communications. A second factor is the uniqueness of the research area being reviewed. If a large number of scientists and institutions in several countries are pursuing a potentially dual-use field of study, effective control over the resulting data may be practically impossible. In contrast, research conducted by a discrete group of scientists at a few readily identifiable institutions may be more amenable to governance.

Finally, security reviewers should keep in mind that a biological weapon is more than a microbial pathogen or a toxin (a chemical poison made by a living organism). Instead, a biological weapon is a system consisting of a formulated microbial or toxin agent, a container holding the agent, and a dispersal mechanism. Weaponization involves five basic steps:

1. A pathogen or toxin is obtained from a natural source or a laboratory culture collection, then selected or genetically modified to render it more virulent or stable in the environment.
2. The agent is produced in significant quantities, either as liquid slurry or a dry powder.
3. The agent is “formulated,” or combined with a medium in which it is suspended or dissolved, along with chemical additives that extend its stability, shelf-life, and virulence in storage and promote its dissemination as an infectious aerosol.
4. The formulated agent is placed in a container in which it can be safely stored and transported.
5. A portion of the formulated agent is disseminated, either as a food or beverage contaminant or fine-particle aerosol, by an appropriate delivery mechanism. The container and dispersal mechanism may form one unit, constituting a biological munition.

A Notional Definition of “Sensitive” Research

Working from these principles, the organizers of the Center for Nonproliferation Studies workshop proposed, for consideration by the participants, a set of criteria for reviewing proposed research projects for their potential national security impact. Of greatest concern is research on human pathogens and toxins, including those on the Select Agents List developed by the U.S. Centers for Disease Control and Prevention,²⁴ which has subdivided Select Agents into three classes, with Category A agents assessed to pose the highest threat in terms of public health impact and Category C agents the lowest. Thus, any research proposal or publication involving a Category A agent would immediately warrant greater scrutiny. The level of concern would be substantially reduced for Category B and C agents, and even further for unlisted agents. Although “concern” is not synonymous with “restriction,” the greater the potential hazard associated with the agent under study, the more care should be taken when conducting the security review.

Nevertheless, limiting the definition of sensitive research to work involving Select Agents would be an imperfect filter because of the possibility that “surrogate” microorganisms could be used to develop weaponization techniques or that non-pathogenic microbes could be engineered to become virulent. For example, the vaccinia, monkeypox, and camelpox viruses, which are not Select Agents, are close genetic relatives of variola (smallpox) virus, so efforts to enhance their virulence could have direct relevance to the weaponization of variola. Similarly, influenza viruses are not Select Agents—yet if they are weaponized, they could pose serious biological warfare threats. Accordingly, the proposed criteria might in rare cases be extended to cover sensitive research involving non-Select Agents. It is also important to note that states and terrorist organizations could employ animal or plant pathogens as economic weapons. In the past, a number of pathogens have been developed for this purpose, and the U.S. Department of Agriculture is preparing a list of select agents that can kill or harm animal and plant populations. The Center for Nonproliferation Studies workshop, however, focused on the potential use of pathogens and toxins against human targets.

Given the importance of the weaponization steps described above, reviewers of scientific research proposals under the proposed scheme would need to go beyond an exclusive focus on Select Agents to consider how the results could facilitate or enhance their weaponization and delivery of pathogens and toxins—for example, by increasing their virulence or survival time in the environment or by improving the means of dispersal over a target population. The types of sensitive scientific data subject to embargo or control would have to be unambiguously and narrowly defined. One approach discussed at the workshop would be to subject to security review any research proposal that (a) involves a Select Agent (or obvious surrogate) and (b) entails one or more of the following research activities:

1. Enhancing the infectivity, pathogenicity (virulence), antibiotic resistance, or resistance to host immunological defenses of pathogens in vitro or in vivo. This might be done through the insertion of virulence factors and other genes into a non-pathogenic or opportunistic pathogen. (Basic research on infectivity, pathogenicity, and immunological resistance would not face restrictions if it did not have direct applications to weaponization.)
2. Improving the ability of a microbial pathogen to remain viable and virulent during prolonged storage, and/or after release into the environment, by such means as the genetic manipulation of cell wall properties, sporulation, or microencapsulation. (Basic research on cell wall properties or sporulation would not be subject to restrictions if it did not have direct applications to weaponization.)
3. Facilitating the dissemination of a microbial or toxin agent as a fine-particle aerosol, including studies of the kinetics and dynamics of particulate aerosols containing microorganisms, microbial simulants, or toxins.
4. Facilitating the dissemination of a microbial or toxin agent by the contamination of food, beverage, or water sources—for example, by enabling the agent to persist for a protracted period in the presence of chlorine or some other disinfectant or to survive cooking temperatures.
5. Creating a novel pathogen or one with characteristics that have been altered so that it can evade current detection methods and/or host immunological defenses.
6. Assembling oligonucleotides (DNA fragments) in vitro with the aim of synthesizing a virus or other microorganism with pathogenic properties.

Under this notional security review process, reviewer(s) would place a research proposal in one of two categories. Studies involving a Select Agent that did not contain any of the six weaponization criteria listed above would be designated “Unrestricted,” with no limits on public release. Studies that involved a Select Agent (or, in rare cases, a surrogate agent) and one or more of the six weaponization criteria would be considered for placement in the “Restricted” category, which would entail certain controls on public release of the research results. The “Restricted” designation could also be applied in the later stages of a research project. Investigators thus would be obligated to notify the funding agency if an experiment was modified to include listed criteria or if the research generated findings having unexpected relevance to biological warfare and terrorism.

Using the review criteria described above, it is expected that only a small fraction of academic research projects would be designated “Restricted.” (One workshop participant suggested that in order to assess the efficacy of the proposed screening criteria, it would be useful to review microbiology research projects funded by the National Science Foundation, the National Institutes of Health, and DoD in a recent year and determine how many projects would have fallen into the “Restricted” category.) Compliance of research institutions with the security review process could be encouraged by withholding federal funding from those that refuse to participate. Non-governmental funders could apply the same security review criteria voluntarily, as they have done with the Asilomar and Recombinant DNA Advisory Committee guidelines.

Thoughts on Restricting Dissemination of Sensitive Results

According to the notional process discussed in the Center for Nonproliferation Studies workshop, when a paper containing the results of a “Restricted” research project was submitted to a scientific publisher, it would be accompanied by a letter from the funding agency denoting which portions of the paper were sensitive and warranted restrictions on distribution. Dissemination of the embargoed material to legitimate scientists (identified through a simple vetting process) would then be controlled by the journal editor, in cooperation with the funding agency. For example, access to sensitive data might be provided through secure, password-controlled websites, with substantial fines and other sanctions (such as denial of access) imposed in cases of unauthorized transfers.

Given the wide range of potential outlets for the communication of research findings (for example, published journals, abstracts, e-mail, postings on the World Wide Web, circulation of preprints, and oral presentations at scientific meetings), procedures for restricting the dissemination of sensitive research results would have to go far beyond the decision of one journal editor not to publish. Sensitive information could not be embargoed effectively unless the author was persuaded not to distribute it publicly in any form. Moreover, because scientific publishing is an international enterprise—editorial boards have international memberships, authors come from many countries, and journals are often owned by multinational corporations—a system focused exclusively on U.S. scientists or publishers would be ineffective.

It is also important to note that a system for restricting the publication of dangerous research findings would not “destroy” the sensitive information but would have the effect of placing it under embargo. Implementing this embargo would require determining who should be given access to the information and who should not.²⁵ In addition, guidelines for controlling the dissemination of sensitive information will need to consider how long the embargo would restrict public access to the sensitive information. Would the embargo be indefinite or only temporary? What mechanisms should be established to make constructive use of the embargoed information in a restricted setting—for example, for the development of defensive measures?

Critiques of the Proposed Review Process

Workshop participants identified serious drawbacks with the proposed security review process. The following points represent a compendium of comments from individual participants, although not all points were necessarily supported by a majority or consensus.

• The proposal appears to establish a system for national censorship in which scientific journals would be the guardians of the censored material. Yet most journals have international editors, not just American ones. Furthermore, most journals are not published by U.S. entities, nor are most published papers supported solely by U.S. funding agencies. Embargoing sensitive data submitted by American authors to U.S.-based journals may be feasible, but without effectively internationalizing the system, how would journals deal with non-American authors who would not be subject to prior security review? How effective would the system be if major non-U.S. journals did not participate?
• Research universities and scientific journals would not agree to the proposed mechanism. A recent MIT report flatly opposes government review of unclassified research to determine whether it contains “sensitive” information. Such a review process, the report states, “opens the Institute and its faculty, students, and staff to potentially arbitrary dictates from individual government contracting agents—however well intended.”²⁶
• The proposed review process ignores the fact that the intellectual knowledge generated under grants (as opposed to contracts) is the property of the investigator and that many research projects involve students and foreign nationals. Is government censorship to be applied to dissertations written by foreign students, all of which are published and available electronically?
• Scientists may be unwilling to take on the responsibility of reviewing a potentially “sensitive” paper, either because they are opposed in principle to the review process or concerned over potential liability for making the wrong security decision. Journals, and science more broadly, will suffer if some of the best scientists decline to review their colleagues’ work.
• Scientific journals are not set up to deal with the segregation of sensitive data or to provide for secure means of review and publication. Most scientific journals published by professional societies are priced at cost, with little overhead. Moreover, many journals are run by academic editors hosting websites with one or two technicians, some on a volunteer basis or housed at universities with minimal technical support. Because of the hassles and costs associated with handling embargoed information, leading scientific journals may simply avoid the problem by not accepting any “Restricted” papers.

• Once a document has been designated “Restricted,” it must be tracked and its security monitored. Secure websites are never truly secure; indeed, the proposed system would provide an inducement for hackers. Moreover, unique access codes, or the information obtained by using them, could be passed along to other people, although putting a time limit on each access code would reduce the risk somewhat.
• The proposed mechanism for controlling access to sensitive information would require scientific journals (including private and nonprofit publishers) to bear the burden of maintaining and administrating a confidential archive forever. Why would any journal voluntarily jump through these hoops, which would involve regulations, paperwork, and up-front costs, to host only a few “Restricted” papers each year? The proposed system would impose high legal and professional risks on journals, with minimal benefits.
• A dissemination mechanism that publicized the fact that sensitive information was available only to a limited group of authorized scientists might call unwanted attention to such information. It would also be unwise to concentrate a library of potentially dangerous information in one place.
• In view of the possibility that some reviewers of submitted research proposals could be competitors of the scientist(s) whose work is being reviewed and could have ulterior motives for classifying research as “Restricted,” the review system must be organized in such a way as to avoid possible conflicts of interest. Even if reviews were conducted anonymously, scientists might be sufficiently familiar with their competitors’ work to identify who had submitted the proposal.

Conclusions

The Center for Nonproliferation Studies workshop was stimulating, informative, frustrating, and, ultimately, fruitful, although not in the way the organizers had expected. The workshop was stimulating because it brought together two groups of professionals who usually do not communicate, the first consisting of scientists, journal editors, and grant administrators and the second of intelligence and security experts. The workshop was informative because each side learned about the concerns and perceptions of the other, yet it was frustrating because the two sides could not reach a consensus on how sensitive information from scientific research should be handled. At the end of the day, however, the workshop was fruitful because it identified broad guidelines for the development of future policies on restricting scientific publication, guidelines that may be refined and broadened by future efforts. In so doing, the participants went some way toward distinguishing between policy actions that are simply unworkable and those that would impose costs and inefficiencies on the scientific enterprise but might still be judged worthwhile because of their benefits for national security.

Most workshop participants agreed that under rare circumstances, the open communication of unclassified research could pose such a high risk of substantial harm as to warrant controlling the distribution of that information. Moreover, unless the scientific community is proactive in addressing the “dark side” of research in the biosciences, it may become vulnerable to political forces demanding the imposition of excessive and inappropriate controls on publication. Restricting the communication of sensitive data in a way that does not cause undue harm to the viability of the scientific enterprise is clearly a difficult balancing act. If restrictions are pursued, they should be narrowly targeted on the small fraction of basic research directly relevant to the weaponization of biological and toxin agents and applied only under exceptional circumstances. Further, the regulations must not be so onerous that they will encourage circumvention or force scientists and journals to neglect or abandon important areas of research.

Finally, any workable system for restricting the publication of “sensitive” information must have the broad—if not wholehearted—support of the international scientific community, which must perceive that the benefits to national security outweigh the costs to science. Researchers and their professional societies are in the best position to provide guidance on how this difficult balance can be achieved, and the results of the Center for Nonproliferation Studies workshop provide a useful first step in that process.

Should the scientific community ultimately conclude that restricting the publication of sensitive research is not a productive way of preventing its misuse for bioweapons development, scientists will need to suggest alternative strategies for achieving this goal. For example, instead of creating a new category of “Restricted” information, it might be desirable to formally classify the small number of papers that contain truly sensitive data and, at the same time, find ways to foster research, reward faculty, and attract students in areas where publication is not required for advancement.

Another alternative would be for expert groups of scientists to assess how new research breakthroughs might be misused by those who wish to acquire weapons of mass destruction—biological, chemical, or nuclear. These groups would then seek to counteract offensive applications by identifying the applications’ likely “signatures” and developing countermeasures against them. In any event, it is essential that bioscientists and their professional organizations take the lead in informing security experts how best to meet the threats of biological warfare and terrorism in a practical and realistic manner. If they cannot, who can?

Acknowledgements

The authors are grateful to Gerald L. Epstein, Ph.D., and Richard M. Pilch, M.D., for useful comments on earlier drafts.

Funding Support

The workshop discussed in this article was funded entirely by the Defense Threat Reduction Agency, a component of DoD, as one task under the Integrated Proliferation Prevention, Research, and Education Project. The branch office of the Monterey Institute’s Center for Nonproliferation Studies in Washington, DC, provided logistic support.

Disclaimer

The interpretations and views expressed in this article are the authors’ own and do not necessarily represent those of the Monterey Institute of International Studies or the Defense Threat Reduction Agency. Also, the workshop participants may or may not agree with the authors’ interpretation of the workshop proceedings and results.

Contact

Raymond A. Zilinskas at rzilinskas@miis.edu.

References

Click on an end note number to return to the article.

1. The Workshop on Guidelines for the Publication of Scientific Research Potentially Related to Biological and Toxin Warfare was held in Washington, DC, on 12 August 2002. Workshop participants were Ronald M. Atlas, American Society for Microbiology; Beverly Berger, U.S. Department of Energy; Eileen R. Choffnes, National Academy of Sciences; Charles Cooke, Defense Threat Reduction Agency; Eric Eisenstadt, Defense Advanced Research Project Agency; Lynn W. Enquist, Princeton University and Editor, Journal of Virology; Gerald L. Epstein, Defense Threat Reduction Agency; Mark S. Frankel, American Association for the Advancement of Science; Clem Gaines, Defense Threat Reduction Agency; Asha M. George, Nuclear Threat Initiative/Biological Programs; Janet R. Gilsdorf, University of Michigan Medical Center; Brooks Hanson, Science Journal; Elisa D. Harris, University of Maryland; Maryanna Henkart, National Science Foundation; David Heyman, Center for Strategic & International Studies; Peter B. Jahrling, U.S. Army Medical Research Institute of Infectious Diseases; Jonathan Knight, American Association of University Professors; Rachel E. Levinson, White House Office of Science and Technology Policy; Jeffrey Milstein, Defense Threat Reduction Agency; Dorothy B. Preslar, ILIAAD; Ariella M. Rosengard, University of Pennsylvania; Jeffery A. Schloss, National Institutes of Health; Ron Sentell, U.S. Department of Energy; Leonard Spector, Monterey Institute of International Studies; Ernest Takafuji, National Institute of Allergy and Infectious Diseases; Jonathan B. Tucker, Monterey Institute; Gillian Woollett, Pharmaceutical Research and Manufacturers of America; Raymond A. Zilinskas, Monterey Institute; and one participant who did not wish to be identified.

2. The term “weaponization” refers to the research and development steps involved in transforming a naturally occurring pathogen or toxin to the point where it becomes suitable for use in a biological weapon system.

3. Paul Zielbauer with William J. Broad, “In Utah, a Government Hater Sells a Germ-Warfare Book,” New York Times, 21 Nov. 2001.

4. R. A. Zilinskas, “Open Publications as Sources of Biological and Chemical Terrorism: Defining the Problem and Applying Remedies,” in Implications of 9/11 on National Security and the Path Forward to Peace, Proceedings of the Twelfth Annual International Arms Control Conference, Sandia National Laboratories, Albuquerque, NM, 18–20 April 2002, pp. 144–160.

5. A. H. Card, Jr., “Memorandum to heads of executive departments and agencies. Action to safeguard information regarding weapons of mass destruction and other sensitive documents related to homeland security” (Washington, DC: The White House, 2002).

6. D. L. Huxsoll, C. D. Parrott, and W. C. Patrick III, “Medicine in Defense Against Biological Warfare,” Journal of the American Medical Assn., Vol. 262, No. 5, 1989, pp. 677–679.

7. C. M. Fraser and M. R. Dando, “Genomics and Future Biological Weapons: The Need for Preventative Action by the Biomedical Community,” Nature Genetics, Vol. 29, Nov. 2001, pp. 253–256.

8. W. L. Broad , “Gene-Engineered Anthrax: Is It a Weapon?” New York Times, 14 February 1998; N. Wade, “Tests With Anthrax Raise Fears That American Vaccine Can Be Defeated,” New York Times, 26 March 1998.

9. P. Aldous, “Biologists Urged to Address Risk of Data Aiding Bioweapons Design,” Nature, Vol. 414, 2001, pp. 237–238.

10. Nell Boyce, “Speak No Evil: Should Biologists Publish Work That Could Be Misused?” U.S. News & World Report, 24 June 2002; www.usnews.com/usnews/issue/020624/misc/24bioterror.htm.

11. Ibid.

12. Jennifer Couzin, “Active Poliovirus Baked From Scratch,” Science, Vol. 297, 2002, pp. 174–175.

13. Sylvia Pagān Westphal, “Ebola Virus Could Be Synthesised,” New Scientist, 17 July 2002; www.newscientist.com/news/print.jsp?id=ns99992555.

14. Jon Cohen, “Designer Bugs,” Atlantic Monthly, Vol. 290, July/Aug. 2002, pp. 113–124.

15. A. Dembner, “Some Say Science Must Restrict Access to Research Data,” Boston Globe, 18 November 2001.

16. Alan Leo, “Science and Secrets,” Technology Review, 20 June 2002; www.technologyreview.com/articles/wo_leo062002.asp.

17. National Academy of Sciences, Scientific Communication and National Security (Washington, DC: National Academy Press, 1982).

18. George Poste, quoted in Dembner.

19. A. A. Salyers, “Science, Censorship, and Public Health,” Science, Vol. 296, 2002, p. 617.

20. R. A. Rosenbaum, M. J. Tenzer, S. H. Unger, W. Van Alstyle, and J. Knight, “Academic Freedom and the Classified Information System,” Science, Vol. 219, 1983, pp. 257–259.

21. Quoted in R. M. Atlas, “Bioterrorism: The ASM Response,” ASM News, Vol. 68, 2002, pp. 117–121.

22. Gregory A. Petsko, “An Asilomar Moment,” Genome Biology, Vol. 3, No. 10, 25 September 2002, pp. 1014.1-1014.3; http://genomebiology.com/2002/3/10/comment/1014.

23. Richard Monastersky, “Publish and Perish?” Chronicle of Higher Education, Vol. 49, No. 7, 11 October 2002, pp. A16–A19.

24. U.S. Centers for Disease Control and Prevention, Strategic Planning Group, “Biological and Chemical Terrorism: Strategic Plan for Preparedness and Response” (Atlanta: CDC, 2000), 49(RR04).

25. Gerald Epstein, “Controlling Biological Warfare Threats: Resolving Potential Tensions Among the Research Community, Industry, and the National Security Community,” Critical Reviews in Microbiology, Vol. 27, No. 4, 2001, pp. 321–354.

26. Massachusetts Institute of Technology, “Report of the Ad Hoc Faculty Committee on Access and Disclosure of Scientific Information” (Cambridge, MA: MIT, 12 June 2002).