---------- Forwarded message ----------
Date: Fri, 14 Feb 1997 08:28:28 -0800 (PST)
From: Declan McCullagh <declan(a)well.com>
To: fight-censorship(a)vorlon.mit.edu
Subject: USACM analysis of Commerce Dept crypto export proposals
February 12, 1997
Nancy Crowe
Regulatory Policy Division
Bureau of Export Administration
Department of Commerce
Room 2705
14th Street and Pennsylvania Ave., N.W.
Washington, D.C. 20230.
Docket No. 960918265-6366-03
Dear Ms. Crowe
The United States Public Policy Committee for the Association for
Computing (USACM) welcomes this opportunity to submit our views on the
Interim Rule issued by the Department of Commerce with regard to
"Encryption Items Transferred From the United States Munitions List to the
Commerce Control List." The USACM believes it is in the best interest of
the U.S. government to promote the widespread use of strong encryption.
>From our perspective the Interim Rule fails to recognize the legitimate
needs and interests of academic, professional, scientific, and ordinary
users of telecommunications technology. Thus, the Interim Rule must be
modified before it can resolve the many problems with the current export
controls on encryption technologies.
Introduction and Summary
The Association for Computing is an international professional
society whose 76,000 members (60,000 in the U.S.) represent a critical
mass of computer scientists in education, industry, and government. The
USACM provides a means for promoting dialogue on technology policy issues
with United States policy makers and the general public. We have
identified a number of serious problems with specific provisions of the
Interim Rule.
As a professional society of computer scientists which produces a
number of peer-reviewed technical journals, we are concerned that the
Interim Rule will hamper both communication and education in our field.
Part 7.34.3 (b)(3) which refers to the distinction between printed and
electronic publications of cryptographic materials is unworkable under the
new paradigms of electronic publishing and communications. Electronic
media, including the World Wide Web, listserves, Usenet news groups, and
video conferencing are becoming the prominent means by which scientists
communicate. Provisions of the Rule, specifically Parts 7.34.9 and 744.9,
which affect teaching cryptography to foreign students are vague and
contradictory. Educational environments are not limited to academic
institutions but also occur in national and industry labs and by distance
education. Restrictions on cryptography exports must not interfere with
the traditional freedom of access over digital networks which is
indispensable to maintain motivated and effective academic and research
communities.
We also believe that the development of public policies and
technical standards for communications technologies, such as a Key
Recovery Infrastructure (KRI), raise vital issues of privacy,
competitiveness, and scientific innovation. Parts 740.8 and 742.15 raise
a number of troubling issues for the computing community. We believe it
is unwise for the Commerce Department to link relaxing export controls on
56-bit encryption to the development of a KRI as both the desirability and
the feasibility of such a system remains uncertain. Key recovery products
have not yet been subject to the vigorous testing necessary for a proposed
standard and there is little understanding of how such a system would
operate and what controls would be needed to ensure that it remained
secure. Also, Supplement No. 7 to part 742 (which requires that
businesses who wish to export 56-bit encryption before 1998 submit a
biannual business plan for developing key recovery products) will stifle
the innovation of new cryptography technologies and discourage the process
of scientific innovation. We believe the Commerce Department should not
promulgate regulations which prohibit U.S. research and development from
responding to market demands and limit the ability of Americans using new
on-line services to protect their privacy.
Analysis
The USACM has identified electronic publication, education,
research and development, key recovery, and privacy as problematic areas
which need addressing. We have outlined our concerns below:
Electronic Publishing
It is unreasonable and unconstitutional to distinguish between
printed and electronic distribution of encryption source code as set forth
in the note to Part 7.34.3 (b)(2) and (b)(3). A Federal Court in
California has ruled in Bernstein v. U.S. Department of State that source
code is speech and is thus protected under the first amendment. This
distinction is also currently being challenged in a federal court in the
District of Columbia in Karn v. U.S. Department of State. The USACM
joined the Electronic Privacy Information Center, the American Civil
Liberties Union, and the Internet Society in submitting an Amici Curiae
brief in the case which argued that such language is an impermissible
regulation aimed at the suppression of expression. As computer scientists
we see no practical reason why the Commerce Department should insist on
creating a distinction when one does not exist.
The ACM is the publisher of numerous scientific publications and
conference proceedings. They range from our flagship journal
Communications of the ACM (CACM) to the on-line, peer-reviewed journal
Experimental Algorithms. All 76,000 members of ACM, including 15,000
members overseas, receive CACM by mail and have access to ACM's on-line
publications. ACM foresees a time when all its publishing will be
electronic and on-line. At that time, it will need interoperable
encryption technology available in the U.S. and in its mirror sites abroad
to dispense its material. Its subscribers worldwide will need access to
secure, commercial encryption as well.
An article which described the development of a new cryptographic
algorithm would likely appear in one of the many technical journals or
conference proceedings published by ACM or the Institute for Electronics
and Electrical Engineers (IEEE), another international professional
society. In fact a number of the groundbreaking articles in the field of
cryptography science were originally published by ACM and IEEE.
Publication of encryption algorithms is extremely important to the field
of cryptography. In order for an algorithm to be trusted, it must be
challenged. To do that, the code must be made widely available. Foreign
members of ACM will be unable to access in electronic format the same
articles they currently receive in the printed journal. And, it is
technically impossible, at this late date, to partition ACM's publications
into distinct paper and electronic (hence encryptable) media.
Electronic communications, including the World Wide Web, list
serves, Usenet news groups, and video conferencing are becoming the
prominent means by which scientists communicate. Science is a global
pursuit and there exists a open communications network between scientists
in different countries. Part 734.2 which prohibits making cryptographic
software available outside the U.S. will not only eliminate this
international communication but also technical communication among U.S.
scientists. In electronic communications it is not always clear to whom
the information is being transmitted. WWW sites and Usenet news groups
are accessible by anyone with a modem. Video conferences can be
retransmitted overseas and moderated listserves are difficult to control.
The Interim Rule refers to an individual taking "precautions adequate to
prevent unauthorized transfer of such code outside the U.S." It is our
belief that it would be impossible to be certain of any precautions taken.
This will effectively eliminate all communications on electronic media
that describe or discuss cryptographic source code.
We believe the interim rule must be revised to eliminate the
distinction between printed and electronic source code and to allow for
open communications within scientific communities. Restricting these
communications will retard the evolution of the science and the
development of new algorithms and cryptographic devices.
Education
Many ACM members are computer science professors and teachers, so
we are concerned about the contradictions in the proposed regulations with
regards to education. A number of fields and sub-fields address
cryptography as part of their curricula. Part 734.9 states that
"Educational Information" is not subject to the new regulations if it is
"released by instruction in catalog courses and associated teaching
laboratories of academic institutions." Computer science, mathematics,
engineering, and electronic security may all include technical instruction
in encryption technologies and would be covered in U.S. university
classrooms. However, questions arise with regard to distance and home or
overseas education because of Part 744.9. It states that "No U.S. person
may, without a license from BXA, provide technical assistance (including
training) to foreign persons with the intent to aid a foreign person in
the development or manufacture outside the United States..." While Part
744.9 defines a U.S. person it does not define "technical assistance" or
"training." It is uncertain whether a U.S. professor teaching a course in
which foreign students are registered, or teaching a course in
cryptography overseas would be "training" a foreigner to develop a
cryptographic device if the course work was more detailed than "a
discussion of information about cryptography." This would affect course
studies as disparate as 'number theory' and 'local area networks'.
Also, educational environments are not limited to academic
institutions but are also found in national and industry labs. Many
computer scientists receive their first hands on training after they
graduate from their University. It is unclear whether this "training" or
"technical assistance" is in violation of the Interim Rule. The intent of
the training is give the new employee the practical tools necessary to
participate in the field of cryptography science, and is not necessarily
intended to be project or employer specific. While the General
prohibition in Part 744.9 discusses the meaning of intent as applied to an
academic setting, it is not clear if "academic setting" can be applied to
instruction which occurs outside of the University environment.
The argument made previously with regard to digital media also
applies to education. As part of their course work, students often use
electronic media as resources (WWW, digital libraries, CD-ROMs), as a
communication device for the class outside the classroom (electronic mail,
listserves), and to learn from listening to the discussions among research
scientists (Usenet groups, listserves). Part 7.34.3 (b)(3) which covers
encryption source code in electronic form or media will restrict these
types of educational instruction. Instructors will be unable to take
advantage of digital media in their courses. Students studying
cryptography will be unfairly disadvantaged as they will be unable to
access valuable resources even in the process of furthering their
education.
The USACM believes the contradictions in Parts 7.34 and 744.9 must
be resolved in a clear manner so educators are not required to reduce the
quality of their courses for fear of misinterpreting the Interim Rule.
Specifically, "academic setting," "training," and "technical assistance"
must be defined, and distance education, and academic research and
communication must be addressed.
Research and Development
Encryption policies must reflect the needs of the global market.
The international demand for products which incorporate strong
cryptographic tools is growing. Such products are widely available and
produced by a number of nations. U.S. scientists have been prominent in
the development of current encryption technologies. The field has
developed though research and development efforts along many different
tangents, only one of which describes key recovery products. There is
little evidence that the demand for cryptography tools is limited to those
products which incorporate key recovery protocols. Part 742.15 (which
states that businesses must submit a business plan for the development of
key recovery products before they may export 56- bit software; the
license must be renewed biannually until 1998 when only key recovery
products will be allowed for export) will restrict the U.S. to producing
only products which incorporate KRI protocols.
Mandating that businesses develop key recovery products will also
impede the natural market development of novel and innovative systems.
Part 740 hypothesizes that a worldwide KRI will be desirable, feasible,
and in place by 1998. However it is unclear whether key recovery is the
best alternative. Research along new tangents will continue in non-
industry and non-U.S. settings. A new protocol may be discovered which is
considered a better choice for a worldwide infrastructure. There will
exist a great market demand for variety in choosing a security system to
fit the needs of the distinct commercial group. If this happens U.S.
scientists and industry will be at a disadvantage as they will have only a
core competence in key recovery protocols as per Part 740.8.
There are a variety of commercial groups interested in utilizing
the Internet for business interactions and transactions. Without
interoperable encryption programs, commercial needs in an increasing
global environment cannot be met. Supplement No. 4 to Part 742 states
that a product can not interact with another product whose key recovery
system has been "altered, bypassed, disabled, or otherwise rendered
inoperative." This will be a major source of problems for researchers and
educators, as well as government and commercial institutions. The result
of a system not being able to talk to another system because of an
intentional or accidental disabling of the KRI protocols can have a very
large impact on telemedicine, research, government operations, and
commercial enterprises.
The USACM believes the Interim Rule should be rewritten to avoid
dissuading innovation and development and eliminating the U.S.'s core
competency in cryptography. It should also recognize the need for
consistency in interoperable systems.
Key Recovery
The USACM recognizes that there is a real market demand for key
recovery products from business and government employers. However, the
viability of a KRI has not yet been determined. It has not yet been
subject to the vigorous testing necessary for a proposed standard. There
is little understanding of how such a system would operate and what
controls would be needed to ensure that it remained secure. Part 740
describes the development of a Key Recovery Infrastructure within two
years. We believe it is unwise for the United States to insist on the
development of a untested, unproved technology for a worldwide
infrastructure. The National Research Council report stated that a
feasibility study needed to be performed on a smaller scale before key
escrow could be seriously proposed for commercial applications. We
believe this warning applies to KRI as well. While key recovery tools may
be appropriate in some settings, we believe it would be wrong to impose
such restrictions on users or businesses and the Interim Rule should not
dictate that businesses limit their research to a potentially unworkable
system.
Privacy
The USACM believes that certain principles should be reflected in a
national cryptography policy. Encryption should be used for privacy
protection and to encourage the development of technologies and
institutional practices which will provide real privacy for the future
users of the NII and real security for the protection of the system. The
USACM believes that transferring the regulation of cryptography to the
Commerce Department could establish United States leadership in protecting
the privacy rights of its citizens. However the Interim Rule fails to do
that.
Conclusion
We recognize that the government has a legitimate interest in
protecting national security. However, whether or not the worldwide
infrastructure is achieved, the role of national security agencies will
remain difficult. The government's proposal to balance national security,
business, and privacy interests by creating a Key Recovery Infrastructure
within the next two years is overly aggressive. We suggest that the
development of a policy that serves the long term interests of our
nation's security will not be one based on a Key Recovery Infrastructure,
but rather one that anticipates the widespread availability of strong
encryption and the multifaceted demands of a global economy. Toward that
end, the interests in protecting open research within the U.S. academic
community will be crucial.
Sincerely,
Barbara Simons, Ph.D
Chair, United States Public Policy Office for the
Association for Computing
The ACM, founded in 1947, is an international non-profit
educational and scientific society dedicated to the development and use of
information technology, and to addressing the impact information
technology has on the world's major social challenges. The Association's
activities include the publication of scholarly journals and the
sponsorship of special interest groups (SIGS) in numerous disciplines. ACM
has 76,000 members. The 60,000 who reside in the United States are
academic, professional, scientific, and ordinary users of
telecommunications technology and have a strong interest in the
development of sound encryption policies. The USACM provides a means for
promoting dialogue on technology policy issues with United States policy
makers and the general public. We respond to requests for information and
technical expertise from United States government agencies and
departments, seeks to influence relevant United States government policies
on behalf of the computing community and the public, and provides
information to the ACM on relevant United States government activities.
