In addition to the above laws, there is 8.1 Critical Technologies evidence that the Government is enadering the noe of contract and grant termo in federally sponsored research as a means of controlling the disarmination of research results. Such restrictions entered into voluntarily by grant recipients and contractors may avoid the constitutional double sur rounding ITAR and EAR and may become the most effective means for controlling the now of scientific information. The Defense Science Board Task Force on University Responsive nes has suggested (ore Section 17), and the Department of Defense is reportedly considering, a mechanism which would focus control of research results on the begotiation process at the time of the government grant or contract

Another potential controlling mechanim may evolve from the Immigration and Nationality Act which empowers the Government to refuse admission to an aben if there is reason to believe that the alien will "engage in activities which would endanger the welfare, safety, or Recenty of the United States." In addiben, an alien may be deported for failing to comply with the conditions under which he was admitted to date, the immigration laws have not been koed in any general way to contro! technology transfer. Recently, however, the Department of State has made formal inquiries to universities of the intended areas of research of certain foreign astionals from Communist Countries

For the purposes of this report we may regard university research activities as points within a cube, whose dimensions are: (1) basic to applied; (2) civilian to military; and (3) internationally noncompetitive to competitive. At one extreme, basic research of only civilian interest with no potential for interna tional competition presents no conflict and can be freely pursued at the university. Equally clearly, research at the opposite extreme does not belong in a university and most probably should be classified. The concerns addressed by this report involve research on so-called critical or sensitive technologies located somewhere in the middle of this cobe. Le baring basic and applied components, dual military civilian uses, and international competitive interest. This is indeed the case with both cryptology and VLS! research.

Starting with crypology, we foresee that during the next decade there will be • progressively increasing number of interconnected computers commeni cating both within and across organiza tional boundaries. Conceivably, if personal computers continue to grow at their current rate of some one million new unita/year the aggregate of geographically distributed systems will extend to individual bomes and small businesses as well in such a future Betting we believe that the protection and authentication of data is a mandatory requirement. Recall that by protection we mean that data may be

4.6 Interpretation and Conclusions communicated without interception,

As is evident from the above summaries of the relevant laws and regulations, the legal structure surrounding the control of scientific and technological information is bewildering. Nevertheless, we shall attempt to draw several conclusiona

The regulations appear to be impracti cal instruments not intended to regulate scientific exchange in the academic world, for the following reasons: 1. At least ITAR and EAR exclude banc research from their sphere of contrel

2. The scienter requirement of the Edier came-that the person discloning information know that the recipient of the information intenda to use it for a prohibited purpose in a foreign country – provides a legal precedent for a similar requirement for EAR

2 Nether EAR Dor ITAR can control

information which has already
reached the public domain. Thus
the Government would find it
difficult to prosecute a case where
the information disclosed in a
meeting had already appeared in
some generally available publica-

whereas by authentication we mean the existence of credible means for certifying the signatory of an electronic message

Data must be protected precisely because the power of computers makes possible the malicious use of these machines in breaking the defenses of a remote installation, selecting data of interest, copying it and finally erasing all traces of such an invasion. As the use of geographically distributed interconpected systems grows beyond today's banking and business applications to broader financial, legal, medical and governmental services the potential and penalty for such misuses becomes greater. It is for this reason that we assert a societal need for vigorous research in public cryptology. For it is only through cryptologic techniques that the protection and authentication of data can be effectively insured.

In addition, the unconstrained pursuit of cryptologic research is expected to have beneficial intellectual repercussions in allied fields of computer science. More importantly the uming of expected developments strongly suggests that now is the time to parove effectively such research and thus to gain on the frontiers of theoretical computer science knowledge. The above motivations are unfortunately in conflict with the possible damage that new publicly available cryptologic results can cause in governmental communications and in the acquisition of foreign intelligence

Proceeding to our second major area, we believe that VLS! research is instrumental to the future evolution of the electrical engineering and computer science disciplines. More specifically, progress in the formation of new systems depends critically on the structure and function of their subord nate components-which are VLSI circuita Thus, for example, the development of successful speech-compre bension systems appears to depend critically on the design of proper VLSI structures that will be word in large numbers to carry out simultaneously and in parallel many mmilar information processing operations.

If automatic speech comprehension is

achieved, it will have, besides the obvious civilian and military applica bons, international competitive reper cussions. Japan, for example, bas already embarked on a ten-year project to achieve this goal (the Fifth Generation Computer Project) with the express intent of eventually dominating the world's computer marketa. As before, we De bere the presence of a commercial conflict, which becomes particularly important in VLSI, because circuits can be easily copied and replicated.

The same argument is applicable to VLSI design techniques, an area where US. dominance is expected. In this area, we are concerned with computer based techniques for effectively designing very mall VLSI circuits (e.g., five mm on each side) containing ever 100,000 circuita a design process that cannot be carried out manually because of inberent complexities. A successful" system capable of such design complexity, along with its software components, can be easily copied, perhaps by foreign scholars who participate in this research The conflicts of purpose characterized by these examples are further compounded by a confusion of boundaries: In both cryptologic and VLS! research the boundary between what is basic research of civilias interest and what can be used for military or international competitive purposes is very diffuse. Consider, for example the RSA EDCTYP tion scheme discussed in Section 3.2 That scheme can be viewed as: (1) a mathematical theory that defines certain functions; (2) an algorithm that implements these functions using subordinate functions like multiplication; and (3) one or two VLS! circuits that implement in hardware that algorithes. Here, the creative and most difficult part has been the discovery of the theory-a basic research activity. The conversion of this theory to an algorithm and the subsequent conversion of that algorithm to VLSI arcuits are relatively straight forward development activities that can be effectively carried out by any good team of domestic or foreign engineers. Can this research activity be clearly partitioned into non-critical and critical parts?

In conclusion, certain new research activities are by their very nature multifaceted, Le, they have basic and applied components, and are significant in the international commercial and military arena. The pursuit of such research is essential for the fundamental growth and leadership of US technology, yet the easy "export" of this research is believed undesirable for it may tend to weaken the nation in military and commercial terms.

staggering and probably not subject to

financial measures. The NSA, in view of these national security implications, feels that university research on public cryptology should be controlled through prepublication reviews.

In the case of VLSI, the governmental concerns center on the leakage of this critical technology through foreign scholars, as well as through the unconstrained publication of research resulta. Since approximately one third of the engineering graduate students of our major universities are not US. citizens, the fear appear to rest en a sound numerical basis. Since VLSI circuits are used in a large variety of applications, including the control and instrumenta tion of weapons, the Government fears that we are weakening through such a leakage of technology the military strength of the US.

In addition, the easy export of forefront VLSI research gives an unfair commercial advantage to the international competition. In blant terms, the Government asks why we freely export precious technology, while paying dearly for formiga im ports such as ol and

The imposition of constraints on research invariably resuhe in a loss of effectiveness. Progress is slower and fewer people, brace fewer good prople, are attracted to persue such research. In addition, certain future opportunities are foreclosed and reaths that would have otherwise been achieved by a wider, intercommunicating community are either never realized or postponed.

The inherent coupling of research and education in our universites means that constraints on the former necessarily lead to constraints on the latter. As a result, in such a constrained environment we cannot train as effectively the scientists and engineers who after moving to industrial and academic settings will generate this nation's future technological progress.

In addition, and from a geo-political viewpoint, the control of US research in certain critical technologies will tend to weaken our albes and is likely to lead third-world and neutral countries away from the US, toward possibly adversary countries for the acquisition of beaded technology.

The imposition of constraints on foreign scholars whe participate in US. university research will lead to reduction of the effective research workforce, since • large number of researchers, hence of good researchers, are not US citizens. In addition, sech restrictions will reduce

& The Problems of Technology the number of foreign scholars who Export

The export of technology that is the focus of governmental concerns involves the leakage of university research either (1) in the form of research results; or (2) through the training of foreign scholars in critical technologies on US.compuses. Is cryptology, the NSA in charged with the responsibility of (1) insuring the security of governmental communica bons; and (2) gathering intelligence brom foreign communications. These two pursuits have the technically conflicting objectives of desiring good cryptography for the US and bad cryptography for other countries. Thus, public cryptology work, if successful, usually helps one of theer objectives while hindering the other, depending on whether it enables better code making or better code breaking.

The financial costs to our Government of the disclosure of critical code breaking techniques can be measured in billions of dollars. For example, the NSA Lee that encryption devices for use in our governmental communications mort operate securely for several decades to insure: (1) certification of the encryption scheme, (2) operation in the field, and (3) immunity for some time beyond their removal from the field. Moreover, the political costs of losing the security of our own communications, or our ability to collect foreign intelligence, could be

acquire first-band knowledge of our system-a loss for the US, bether such scholars say in this country or return to their own countries. In addition, the exclusion of foreign scholar from certain research activities is impracticable and creates an unpleas ant chmate. Impracticable because it in difficult for a university which is predicated on the free pursuit of ideas to police who persem what ideas, and enpleasant because of the evident discrimination associated with such a restriction

Finally, in today's arsting of multinational corporations, is not at all clear that control of university research of of its access by foreign scholars will be effective in reducing the overall leakage of critical technologies.

Takes together, these consequences suggest that such constraints are biely to reduce our overall technological leadership and weaken the very strength that they are introded to protect

61 Introduction

The arguments of the preceding section have led us to search for effective means that can allay governmental concerns while preserving the funda mental strength of unconstrained research. There is clearly so perfect solution that can thoroughly satisfy both sides of this conflict. Accordingly. we have approached this serious problem with a degree of flexibility and a tolerance for less than perfect solutions.

The balance on which we have settled is embodied in our recommendations for a relevant MIT policy which is presented bext along with summary of our rationale

6.2 Recommended MIT Policy

We believe that one important part of MIT mission is to prepare our society for a technologically advanced future. To do this, we must continue to pursue leading-edge research in areas that we believe to be of future significance, and to prepare the future professionals in forefront technologies. Current and expected developments in communica tion, information systems and other areas of science and technology call for continued intensive research and development efforts on our part In some areas, eg, very large scale integrated circuit design (VLSI), continued academic involvement is also central to the evolution of the underlying disciplines, in this case electrical engineering and computer science. At the same time, Derious concern has been expressed

about some technology transfer resulting from normal university activities in these areas. We therefore believe it becessary to state our policy with respect to this issue.

Freedom of inquiry and freedom to communicate are essential features of a university. Accordingly, we must be able to teach and perform research in an atmosphere where ideas are freely pursued and onchanged. MIT's role in advancing technology should continue in this open atmosphere. It is also true we believe, that scientific and technological progress are best secured in an open atmosphere, and that the scientific costs to the nation of Imposing restrictions outweigh the benefica. Openness also requires that as a general policy MIT not undertake classified research, or research whose results may not be freely published without prior permission. We believe that openDess of the university also requires that, once they are among us, foreign students, faculty and schol. are should be on an equal basis with their U.S. counterparts in their access to MIT academic and research project. Moreover, restrictions on access to ideas or places within a university are difficult to enforce and likely to be ineffective.

Exceptions to these policies regarding publication, classification and foreign students and scholars may be made, but only in those very rare instances where the area of work is crucially important to MIT educational mission and the exception is demonstrably necessary for the national good. If these conditiods are Dot wel, MIT will decline or discontinue the activity and, if appropriate, propose it for consideration off-campus or alorwbers.

MIT, bike other universities, has a responsibility to the national interest. When sensitive but enclassified research at MTT is important to the national security we will take appropriate steps to ensure that the relevant government agencies are informed of the results. For example, it is our current policy to inform the US. Government of research in information protection and authenti cation by sending prepublication mater ial in this area to the NSA at the same time that we grad it to our close colleagues for technical comment. As a further example, we have also agreed with the NSF that in those rare

La view of the significance of the Corson Report, (see Section 3.7) and the similarity of the issues and recommendations addressed by both the Corson and MIT reports we make the following com media:

We are in full agreement with the

principles of the Corson Report Our concerns stem from the possible misinterpretation of the Report's specific criteria under which "gray area" research may be subject to restrictions and of its specific methods for imple menting such restrictions. In particular, we fear that its detailed recommendations may be interpreted and implemented in ways which ignore their accompanying qualifications. If indeed such qualifica tions are ignored, the Corson Report recommendations could be read as restrictive imperatives.

Our own policy in Section 6.2 above calls for openness of communication and equality among foreign nationals and their US peers. "Exceptions may be made but only in those very rare instances where the area of work is

crucially important to MIT educational

mission and the exception is demonstrably necessary for the national good." By not specifying, priori, the criteria for or the general nature of such exceptiona, we feel that the danger of converting qualified examples into rules is mitigated.

The Five Presidents' Letter

February 27, 1981

The Honorable Malcolm Baldrige Secretary of Commerce 14th Street

Washington, D.C. 20230

The Honorable Alexander M. Haig, Jr. Secretary of State

2201 C Street, N.W. Washington, D.C. 20520

The Honorable Caspar Weinberger Secretary of Defense

The Protagon

Washington, D.C. 20301

and conferences where foreign national were present, employ forego Donal to work in the laboratores, or publish research findings in the open burrature Nor could universities, in effect, and foreign nationals to graduate studim in those areas. Such rernictions would conflict with the fundamental precepts that define the role and operation of th nabon's universities.

The regulations could be interpreted in cover instructive and research which although potentially useful in military applications, have much broader stility in such other areas as medical systema and communication rent. Such interpretations of the regulations. coupled with their orvere criminal penalties, could have a very rea and anintended chilling effect of legitimais

Dear Macro. Baldrige, Haig, and academic exchange.
Weinberger.

We are writing to request clarification of the applicability of certain export restrictions to teaching and research activities conducted by American universition. We are deeply concerned about recent attempts to apply to universition the International Traffic in Arms Regulations (ITAR) and the Export Administration Regulations (EAR) Examples of such efforts by government agencies include a December 12, 1980, memorandum by the Director of the Very High Speed Integrated Circuit (VHSIC) Program Offer, attempts to restrict publication of unclassified university research results arising from DOD-sponsored projecta, and a Depart ment of Commerce mandate to at least so university barring certain foreign scholars from that university's spon sored research activities due to their citizenship. Unfortunately, these initia tives appear to be only the first of many sach actions to follow.

The ITAR and EAR regulations have existed for a number of years, and have not until now been applied to traditional aniversity activities. The new construe tion of these regulations appears to contemplate government restrictions of research publications and of discourse among achoiers, as well as discrimina Lion based on nationality in the employment of faculty and the admission of students and visiting scholars. In the broad scientific and technical areas defined in the regulations, faculty could not conduct classroom lectures when foreign students were present, engage in the exchange of information with foreign visitors, present papers or participate in discussions at symposia

Restricting the free Dow of inform among scientions and engineers would aher fundamentally the system that produced the sostic and technological lead that the government is now bying to protect and leave with nothing w protect in the very bear future. The way to protect that lead is to make sure that the country's best talent is encouraged w work in the relevant areas, not to try i build a wall around past discoveries

It should be recognized that the only realistic way to "contain" VRSIC research is to classify the who program. In our view this would be m defeating effort the science undergring high technologies cannot be put into the bottle. Furthermers, universities have conceded that p formance of classified research a incompatible with their essential parposms University acestuia esil prefer, for the most part, to change the Beld of interest rather than here the research and teaching so consound Forcing high technology research was d universities would decrease our ebm's competitive position, since the d would have to be carried out mare sewly and less effectively in classified atmosphere. Moreover, we would close a continuous Bow of new gradu from the university programs which have been flourishing up to a pat Elimination of such waching and research from academic in would endanger the future of gradua programs in engineering. science, and related Selda, and result in a tremendous imo of pour high technology otherwise avalabi a American Industry. The new ONCE A