May 30, 1996, Prepublication Copy
Subject to Editorial Correction
For most of history, cryptography--the art and science of secret writing--has belonged to governments concerned about protecting their own secrets and about asserting their prerogatives for access to information relevant to national security and public safety. In the United States, cryptography policy has reflected the U.S. government's needs for effective cryptographic protection of classified and other sensitive communications as well as its needs to gather intelligence for national security purposes, needs that would be damaged by the widespread use of cryptography. National security concerns have motivated such actions as development of cryptographic technologies, development of countermeasures to reverse the effects of encryption, and control of cryptographic technologies for export.
In the last 20 years, a number of developments have brought about what could be called the popularization of cryptography. First, some industries--notably financial services--have come to rely on encryption as an enabler of secure electronic funds transfers. Second, other industries have developed an interest in encryption for protection of proprietary and other sensitive information. Third, the broadening use of computers and computer networks has generalized the demand for technologies to secure communications down to the level of individual citizens and assure the privacy and security of their electronic records and transmissions. Fourth, the sharply increased use of wireless communications (e.g., cellular telephones) has highlighted the greater vulnerability of such communications to unauthorized intercept as well as the difficulty of detecting these intercepts.
As a result, efforts have increased to develop encryption systems for private sector use and to integrate encryption with other information technology products. Interest has grown in the commercial market for cryptographic technologies and systems incorporating such technologies, and the nation has witnessed a heightened debate over individual need for and access to technologies to protect individual privacy.
Still another consequence of the expectation of widespread use of encryption is the emergence of law enforcement concerns that parallel, on a civilian basis, some of the national security concerns. Law enforcement officials fear that wide dissemination of effective cryptographic technologies will impede their efforts to collect information necessary for pursuing criminal investigations. On the other side, civil libertarians fear that controls on cryptographic technologies will give government authorities both in the United States and abroad unprecedented and unwarranted capabilities for intrusion into the private lives of citizens.
At the request of the U.S. Congress in November 1993, the National Research Council's Computer Science and Telecommunications Board (CSTB) formed the Committee to Study National Cryptography Policy. In accordance with its legislative charge (Box P.1), the committee undertook the following tasks:
Given the diverse applications of cryptography, national cryptography policy involves a very large number of important issues. Important to national cryptography policy as well are issues related to the deployment of a large-scale infrastructure for cryptography and legislation and regulations to support the widespread use of cryptography for authentication and data integrity purposes (i.e., collateral applications of cryptography), even though these issues have not taken center stage in the policy debate.
The committee focused its efforts primarily on issues related to cryptography for confidentiality, because the contentious problem that this committee was assembled to address at the center of the public policy debate relates to the use of cryptography in confidentiality applications. It also addressed issues of cryptography policy related to authentication and data integrity at a relatively high level, casting its findings and recommendations in these areas in fairly general terms. However, it notes that detailed consideration of issues and policy options in these collateral areas requires additional study at a level of detail and thoroughness comparable to that of this report.
In preparing this report, the committee reviewed and synthesized relevant material from recent reports, took written and oral testimony from government, industry, and private individuals, reached out extensively to the affected stakeholders to solicit input, and met seven times to discuss the input from these sources as well as the independent observations and findings of the committee members themselves. In addition, this study built upon three prior efforts to examine national cryptography policy: the Association for Computing Machinery report Codes, Keys, and Conflicts: Issues in U.S. Crypto Policy, the Office of Technology Assessment report Information Security and Privacy in Network Environments, and the JASON encryption study.  A number of other examinations of cryptography and/or information security policy were also important to the committee's work.
The subject of national cryptography policy is quite complex, as it figures importantly in many areas of national interest. To keep the project manageable within the time, resources, and expertise available, the committee chose not to address in detail a number of issues that arose with some nontrivial frequency during the course of its study.
At the same time, this report does address certain conditions of the political, social, and technological environment that will affect the answers that anyone would formulate to these questions, such as the potential impact on policy of a world that offers many users the possibilities of secure communications.
For most of history, the science and technologies associated with cryptography have been the purview of national governments and/or heads of state. It is only in the last 25 years that cryptographic expertise has begun to diffuse into the nongovernment world. Thus, it is not surprising that much of the basis and rationale underlying national cryptography policy has been and continues to be highly classified. Indeed, in a 1982 article, then-Deputy Director of the Central Intelligence Agency Bobby R. Inman wrote that
Secrecy is a two-edged sword for a democratic nation--on the one hand, secrecy has a legitimate basis in those situations in which fundamental national interests are at stake (e.g., the preservation of American lives during wartime). Moreover, the history of intelligence reveals many instances in which the revelation of a secret, whether intentional or inadvertent, has led to the compromise of an information source or the loss of a key battle.
On the other hand, secrecy has sometimes been used to stifle public debate and conceal poorly conceived and ill-informed national policies, and mistrust is therefore quite common among many responsible critics of government policy. A common refrain by defenders of policies whose origins and rationales are secret is that "if you knew what we knew, you would agree with us." Such a position may be true or false, but it clearly does not provide much reassurance for those not privy to those secrets for one very simple reason: those who fear that government is hiding poorly conceived policies behind a wall of secrecy are not likely to trust the government, yet in the absence of a substantive argument being called for, the government's claim is essentially a plea for trust.
In pursuing this study, the committee has adopted the position that some secrets are still legitimate in today's global environment, but that its role is to illuminate as much as possible without compromising those legitimate interests. Thus, the committee has tried to act as a surrogate for well-intentioned and well-meaning people who fear that the worst is hiding behind the wall of secrecy--it has tried to ask the questions that these people would have asked if they could have done so. Public Law 103-160 called for all defense agencies, including the National Security Agency, to cooperate fully with the National Research Council in this study.
For obvious reasons, the committee cannot determine if it did not hear a particular piece of information because an agency withheld that information or because that piece of information simply did not exist. But for a number of reasons, the committee believes that to the best of its knowledge, the relevant agencies have complied with Public Law 103-160 and other agencies have cooperated with the committee. One important reason is that several members of the committee have had extensive experience (on a classified basis) with the relevant agencies, and these members heard nothing in the briefings held for the committee that was inconsistent with that experience. A second reason is that these agencies had every motivation and self-interest to make the best possible case for their respective positions on the issues before the committee. Thus, on the basis of agency assurances that the committee has indeed received all information relevant to the issue at hand, they cannot plausibly argue that "if the committee knew what Agency X knew, it would agree with Agency X's position."
This unclassified report does not have a classified annex, nor is there a classified version of it. After receiving a number of classified briefings on material relevant to the subject of this study, the fully cleared members of the committee (13 out of the total of 16) agree that these details, while necessarily important to policy makers who need to decide tomorrow what to do in a specific case, are not particularly relevant to the larger issues of why policy has the shape and texture that it does today nor to the general outline of how technology will and policy should evolve in the future. For example, the committee was briefed on certain intelligence activities of various nations. Policy makers care that the activities of nation X (a friendly nation) fall into certain categories and that those of nation Y (an unfriendly nation) fall into other categories, because they must craft a policy toward nation X in one way and one toward nation Y in another way. But for analytical purposes, the exact names of the nations involved are much less relevant than the fact that there will always be nations friendly and unfriendly to the United States. Committee members are prepared to respond on a classified basis if necessary to critiques and questions that involve classified material.
As for the time line of this study, the committee was acutely aware of the speed with which the market and product technologies evolve. The legislation called for a study to be delivered within 2 years after the full processing of all necessary security clearances, and the study committee accelerated its work schedule to deliver a report in 18 months from its first meeting (and only 13 months from the final granting of the last clearance). The delivery date of this study was affected by the fact that the contract to fund this study was signed by the Department of Defense on September 30, 1994.
The title of this report is Cryptography's Role in Securing the Information Society. The committee chose this title as one best describing our inquiry and report--that is, the committee has tried to focus on the role that cryptography, as one of a number of tools and technologies, can play in providing security for an information age society through, among other means, preventing computer-enabled crimes and enhancing national security. At the same time, the committee is not unaware of the acronym for this report--CRISIS--and it believes that the acronym is apt.
From my own standpoint as chair of the NRC Committee to Study National Cryptography Policy, I believe that the crisis is a policy crisis, rather than a technology crisis, an industry crisis, a law enforcement crisis, or an intelligence-gathering crisis.
It is not a technology crisis because technologies have always been two-edged swords. All technologies--cryptography included--can be used for good or for ill. They can be used to serve society or to harm it, and cryptography will no doubt be used for both purposes by different groups. Public policy will determine in large measure not just the net balance of benefit and loss but also how much benefit will be derived from constructive uses of this remarkable technology.
It is not an industry crisis, nor a law enforcement crisis, nor an intelligence-gathering crisis, because industry, law enforcement, and the intelligence establishment have all had to cope with rapid technological change, and for the most part the vitality of these enterprises within the nation is a testament to their successes in so coping.
But a policy crisis is upon the nation. In the face of an inevitably growing use of cryptography, our society, acting as it must through our government as informed by the manifold forums of our free private processes, has been unable to develop a consensus behind a coherent national cryptography policy, neither within its own ranks nor with the private stakeholders throughout society--the software industry, those concerned with computer security, the civil liberties community, and so on. Indeed, the committee could not even find a clear written statement of national cryptography policy that went beyond some very general statements.
To be sure, a number of Administration proposals have seen the light of day. The best known of these proposals, the Clipper initiative, was an honest attempt to address some of the issues underlying national cryptography policy, but one of its primary effects was to polarize rather than bring together the various stakeholders, both public and private. On the other hand, it did raise public awareness of the issue. In retrospect, many Administration officials have wished that the discourse on national cryptography policy could have unfolded differently, but in fairness we recognize that the government's task is not easy in view of the deep cleavages of interest reviewed in this report. In this context, we therefore saw it as our task, commanded by our statutory charge, to analyze the underlying reasons for this policy crisis and the interests at stake, and then to propose an intelligent, workable and acceptable policy.
The Committee to Study National Cryptography Policy is a group of 16 individuals with very diverse backgrounds, a broad range of expertise, and differing perspectives on the subject. The committee included individuals with extensive government service and also individuals with considerable skepticism about and suspicion of government; persons with great technical expertise in computers, communications, and cryptography; and persons with considerable experience in law enforcement, intelligence, civil liberties, national security, diplomacy, international trade, and other fields relevant to the formation of policy in this area. Committee members were drawn from industry, including telecommunications and computer hardware and software, and from users of cryptography in the for-profit and not-for-profit sectors; serving as well were academics and think-tank experts. The committee was by design highly heterogeneous, a characteristic intended to promote discussion and synergy among its members.
At first, we wondered whether these different perspectives would allow us to talk among ourselves at all, let alone come to agreement. But the committee worked hard. The full committee met for a total of 23 days in which we received briefings and argued various points; ad hoc subcommittees attended a dozen or so additional meetings to receive even more briefings; members of the committee and staff held a number of open sessions in which testimony from the interested public was sought and received (including a very well attended session at the Fifth Annual Conference on Computers, Freedom, and Privacy in San Francisco in early 1995 and an open session in Washington, D.C., in April 1995); and the committee reviewed nearly a hundred e-mail messages sent in response to its Internet call for input. The opportunity to receive not only written materials but also oral briefings from a number of government agencies, vendors, trade associations, and assorted experts, as well as to participate in the first-ever cryptography policy meeting of the Organization for Economic Cooperation and Development and of its Business Industry Advisory Council, provided the occasion for extended give-and-take discussions with government officials and private stakeholders.
Out of this extended dialogue, we found that coming to a consensus among ourselves--while difficult--was not impossible. The nature of a consensus position is that it is invariably somewhat different from a position developed, framed, and written by any one committee member, particularly before our dialogue and without comments from other committee members. Our consensus is a result of the extended learning and interaction process through which we lived rather than any conscious effort to compromise or to paper over differences. The committee stands fully behind its analysis, findings, and recommendations.
We believe that our report makes some reasonable proposals for national cryptography policy. But a proposal is just that--a proposal for action. What is needed now is a public debate, using and not sidestepping the full processes of government, leading to a judicious resolution of pressing cryptography policy issues and including, on some important points, legislative action. Only in this manner will the policy crisis come to a satisfactory and stable resolution.
The full list of individuals (except for those who explicitly requested anonymity) who provided input to the committee and the study project is contained in Appendix A. However, a number of individuals deserve special mention. Michael Nelson, Office of Science and Technology Policy, kept us informed about the evolution of Administration policy. Dorothy Denning of Georgetown University provided many useful papers concerning the law enforcement perspective on cryptography policy. Clinton Brooks and Ron Lee from the National Security Agency and Ed Roback and Raymond Kammer from the National Institute of Standards and Technology acted as agency liaisons for the committee, arranging briefings and providing other information. Marc Rotenberg from the Electronic Privacy Information Center and John Gilmore from Cygnus Support provided continuing input on a number of subjects as well as documents released under Freedom of Information Act requests. Rebecca Gould from the Business Software Alliance, Steve Walker from Trusted Information Systems, and Ollie Smoot from the Information Technology Industry Council kept the committee informed from the business perspective. Finally, the committee particularly acknowledges the literally hundreds of suggestions and criticisms provided by the reviewers of an early draft of this report. Those inputs helped the committee to sharpen its message and strengthen its presentation, but of course the content of the report is the responsibility of the committee.
The committee also received a high level of support from the National Research Council. Working with the Special Security Office of the Office of Naval Research, Kevin Hale and Kimberly Striker of the NRC's National Security Office had the complex task of facilitating the prompt processing of security clearances necessary to complete this study in a timely manner and otherwise managing these security clearances. Susan Maurizi worked under tight time constraints to provide editorial assistance. Acting as primary staff for the committee were Marjory Blumenthal, John Godfrey, Frank Pittelli, Gail Pritchard, and Herb Lin. Marjory Blumenthal directs the Computer Science and Telecommunications Board, the program unit within the National Research Council to which this congressional tasking was assigned. She sat with the committee during the great majority of its meetings, providing not only essential insight into the NRC process but also an indispensable long-term perspective on how this report could build on other CSTB work, most notably the 1991 NRC report Computers at Risk. John Godfrey, research associate for CSTB, was responsible for developing most of the factual material in most of the appendixes as well as for tracking down hundreds of loose ends; his prior work on a previous NRC report on standards also provided an important point of departure for the committee's discussion on standards as they apply to cryptography policy. Frank Pittelli is a consultant to CSTB, whose prior experience in computer and information security was invaluable in framing a discussion of technical issues in cryptography policy. Gail Pritchard, project assistant for CSTB, handled logistical matters for the committee with the utmost skill and patience as well as providing some research support to the committee. Finally, Herb Lin, senior staff officer for CSTB and study director on this project, arranged briefings, crafted meeting agendas, and turned the thoughts of committee members into drafts and then report text. It is fair to say that this study could not have been carried out nor this report written, especially on our accelerated schedule, without his prodigious energy and his extraordinary talents as study director, committee coordinator, writer, and editor.
Kenneth Dam, Chair
Committee to Study
National Cryptography Policy
March 29, 1996
A Channel for Feedback
CSTB will be glad to receive comments on this report. Please send them via Internet e-mail to CRYPTO@NAS.EDU, or via regular mail to CSTB, National Research Council, 2101 Constitution Avenue NW, Washington, DC 20418.
 Susan Landau et al., Codes, Keys, and Conflicts: Issues in U.S. Crypto Policy, Association for Computing Machinery Inc., New York, 1994.
 U.S. Congress, Office of Technology Assessment, Information Security and Privacy in Network Environments, OTA-TCT-606, U.S. Government Printing Office, Washington, D.C., September 1994.
 JASON Program Office, JASON Encryption/Privacy Study, Report JSR-93-520 (unpublished), MITRE Corporation, Reston, Va., August 18, 1993.
 These works include Global Information Infrastructure, a joint report by the European Association of Manufacturers of Business Machines and Information Technology Industry, the U.S. Information Technology Industry Council, and the Japan Electronic Industry Development Association (EUROBIT-ITI-JEIDA), developed for the G-7 Summit on the Global Information Society, GII Tripartite Preparatory Meeting, January 26-27, 1995, Brussels; the U.S. Council for International Business statement titled "Business Requirements for Encryption," October 10, 1994, New York; and the International Chamber of Commerce position paper "International Encryption Policy," Document No. 373/202 Rev. and No. 373-30/9 Rev., Paris, undated. Important source documents can be found in Lance J. Hoffman (ed.), Building in Big Brother, Springer-Verlag, New York, 1995; and in the cryptography policy source books published annually by the Electronic Privacy Information Center in Washington, D.C.
 Computer Science and Telecommunications Board, National Research Council, Computers at Risk: Safe Computing in the Information Age, National Academy Press, Washington, D.C., 1991.
 Bobby Inman, "Classifying Science: A Government Proposal . . . ," Aviation Week and Space Technology, February 8, 1982, p. 10.
 For example, following press reports of deciphered Libyan messages before and after a bombing in West Berlin in which an American soldier died, Libya changed its communications codes. A senior American official was quoted as saying that the subsequent Libyan purchase of advanced cryptographic equipment from a Swiss firm was "one of the prices [the United States is] paying for having revealed, in order to marshal support of our allies and public opinion, that intercepted communications traffic provided evidence that Libya was behind the bombing of the Berlin disco." See "Libyans Buy Message-Coding Equipment," Washington Post, April 22, 1986, p. A-8.
 The point of contact within the National Research Council for such inquiries is the Computer Science and Telecommunications Board, National Research Council, 2101 Constitution Avenue, N.W., Washington, D.C. Telephone 202-334-2605 or e-mail CSTB@NAS.EDU.
 Note that the committee was quite aware of potential financial conflicts of interest among several of its members. In accordance with established National Research Council procedures, these potential financial conflicts of interest were thoroughly discussed by the committee; no one with a direct and substantial financial stake in the outcome of the report served on the committee.
Address questions/comments to firstname.lastname@example.org
Last Updated on 05/30/96