[Senate Hearing 116-122]
[From the U.S. Government Publishing Office]
S. Hrg. 116-122
EXPORT CONTROL REFORM IMPLEMENTATION: OUTSIDE PERSPECTIVES
=======================================================================
HEARING
BEFORE THE
COMMITTEE ON
BANKING,HOUSING,AND URBAN AFFAIRS
UNITED STATES SENATE
ONE HUNDRED SIXTEENTH CONGRESS
FIRST SESSION
ON
CONDUCTING OVERSIGHT ON IMPLEMENTATION OF THE EXPORT CONTROL REFORM ACT
(ECRA)
__________
JULY 18, 2019
__________
Printed for the use of the Committee on Banking, Housing, and Urban
Affairs
[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]
Available at: https: //www.govinfo.gov /
__________
U.S. GOVERNMENT PUBLISHING OFFICE
39-542 PDF WASHINGTON : 2020
-----------------------------------------------------------------------------
COMMITTEE ON BANKING, HOUSING, AND URBAN AFFAIRS
MIKE CRAPO, Idaho, Chairman
RICHARD C. SHELBY, Alabama SHERROD BROWN, Ohio
PATRICK J. TOOMEY, Pennsylvania JACK REED, Rhode Island
TIM SCOTT, South Carolina ROBERT MENENDEZ, New Jersey
BEN SASSE, Nebraska JON TESTER, Montana
TOM COTTON, Arkansas MARK R. WARNER, Virginia
MIKE ROUNDS, South Dakota ELIZABETH WARREN, Massachusetts
DAVID PERDUE, Georgia BRIAN SCHATZ, Hawaii
THOM TILLIS, North Carolina CHRIS VAN HOLLEN, Maryland
JOHN KENNEDY, Louisiana CATHERINE CORTEZ MASTO, Nevada
MARTHA McSALLY, Arizona DOUG JONES, Alabama
JERRY MORAN, Kansas TINA SMITH, Minnesota
KEVIN CRAMER, North Dakota KYRSTEN SINEMA, Arizona
Gregg Richard, Staff Director
John V. O'Hara, Chief Counsel for National Security Policy
James Guiliano, Professional Staff Member
Laura Swanson, Democratic Deputy Staff Director
Colin McGinnis, Democratic Policy Director
Cameron Ricker, Chief Clerk
Shelvin Simmons, IT Director
Charles J. Moffat, Hearing Clerk
Jim Crowell, Editor
(ii)
C O N T E N T S
----------
THURSDAY, JULY 18, 2019
Page
Opening statement of Chairman Crapo.............................. 1
Prepared statement........................................... 21
Opening statements, comments, or prepared statements of:
Senator Brown................................................ 3
Prepared statement....................................... 22
Senator Tester............................................... 4
WITNESSES
Eric L. Hirschhorn, Former Under Secretary for Industry and
Security, Department of Commerce............................... 5
Prepared statement........................................... 23
Responses to written questions of:
Chairman Crapo........................................... 35
Senator Brown............................................ 39
Senator Warren........................................... 39
Senator Cortez Masto..................................... 40
Senator Sinema........................................... 43
Nova J. Daly, Former Deputy Assistant Secretary of Treasury for
Investment Security (2006-2009) and Senior Public Policy
Advisor, Wiley Rein LLP........................................ 7
Prepared statement........................................... 27
Responses to written questions of:
Chairman Crapo........................................... 43
Senator Brown............................................ 47
Senator Warren........................................... 49
Senator Cortez Masto..................................... 51
Senator Sinema........................................... 52
Ben Buchanan, Ph.D., Assistant Teaching Professor, School of
Foreign Service Senior Faculty Fellow, Center for Security and
Emerging Technology, Georgetown University..................... 9
Prepared statement........................................... 32
Responses to written questions of:
Chairman Crapo........................................... 53
Senator Warren........................................... 54
Senator Sinema........................................... 55
Additional Material Supplied for the Record
Letter submitted by Dennis Ralston, Sr. Director--Government
Affairs and Cooperative R&D, KLA............................... 56
(iii)
EXPORT CONTROL REFORM IMPLEMENTATION: OUTSIDE PERSPECTIVES
----------
THURSDAY, JULY 18, 2019
U.S. Senate,
Committee on Banking, Housing, and Urban Affairs,
Washington, DC.
The Committee met at 10:04 a.m. in room SD-538, Dirksen
Senate Office Building, Hon. Mike Crapo, Chairman of the
Committee, presiding.
OPENING STATEMENT OF CHAIRMAN MIKE CRAPO
Chairman Crapo. The hearing will come to order.
No one can dispute that technological advances are of vital
importance to United States progress and development, where
progress in knowledge and innovations undergird the growth of
our U.S. productivity.
The U.S.-China Commission found that about half of the U.S.
GDP and two-thirds of its productivity gains is attributable to
U.S. technological improvements.
In August of 2018, the President signed the Foreign
Investment Review Modernization Act, called ``FIRRMA,'' and the
Export Control Reform Act, known as ``ECRA,'' into law.
FIRRMA is designed to strengthen the existing regulatory
architecture in significant ways to deal with inbound foreign
investments that would have the potential to threaten U.S.
national security interests.
ECRA importantly reauthorizes an otherwise moribund Export
Administration Act, continued only by annual reissuances of
Presidential national security declarations.
It authorizes the Bureau of Industry and Security, or BIS,
at Commerce to update controls on exports designed to prevent
certain U.S. dual-use technologies, lower-level military items,
and other things from ending up in the wrong hands.
These two important, hugely bipartisan bills were intended,
in no small part, to ensure that with proper controls in place
to establish highly guarded inward and outward regimes, a
productive relationship between the United States and China is
not only possible, but could be of the highest value in terms
of global prosperity and security.
Today's hearing picks up where the Committee left off when
it last looked at assessing investment controls on technology
in its June 4th hearing on ``Confronting Threats from China.''
On June 4th, we examined China's intention to secure global
technological leadership for itself, with a particular emphasis
on some of its inbound foreign direct investment strategies,
particularly into the U.S. semiconductor industry.
Today the Committee shifts gears slightly to examine
control issues surrounding exports of things outbound from the
United States and other re-exports or transfers that may occur
abroad.
Right now there is a raft of export control regulation on
the horizon at the Commerce Department.
So far BIS is actively engaged on two rulemaking fronts
covering ``emerging and foundational technologies,'' which
include technologies from such sectors as artificial
intelligence, computing, additive manufacturing, data
analytics, robotics, surveillance, and a long list of others.
Importantly, the items that BIS designates as ``emerging
technology'' will also be deemed to be ``critical technology''
under FIRRMA and subject many potential inbound investment
deals to CFIUS review notification requirements.
The current rulemaking under consideration at BIS is not
set in stone.
It is busy poring over a myriad of industry and
governmental comments that will inform its application of
strict controls over emerging technologies, which industry will
use to understand to whom it can transfer these technologies,
who can otherwise use them, and who can even research them.
The Committee has before it a very accomplished panel of
witnesses assembled to help us pull apart the underlying risks
associated with the United States continuing its robust
international economic relationships, including that with
China, against preserving U.S. technological leadership over
these emerging and foundational technologies and some of the
more sensitive items that that would produce.
In the past, export controls sometimes have not been able
to keep up with innovation, and this problem is exacerbated by
today's pace of advancements, particularly in ``artificial
intelligence,'' which owing to its nature is itself a difficult
sector to control.
Considering that BIS is very unlikely to designate all
artificial intelligence technology, we are fortunate to have
Dr. Buchanan here today to help the Committee better understand
what ``artificial intelligence'' means, how it works, and why
or why not certain aspects are more controllable than others.
Our professional export control experts, Mr. Hirschhorn and
Mr. Daly, are expected to offer their assessments on how BIS
may establish controls that address emerging and foundational
technologies, while preserving the innovative capacity of the
United States.
Before I turn to Senator Brown for his statement, let me
indicate that I am going to have to step out for hopefully not
too long to go to the Judiciary Committee where legislation
dealing with AML BSA issues and other aspects that are of great
interest and jurisdiction of this Committee are being
considered at this moment. So I am going to have to step down
there. I will turn it over to you, Senator Brown, and please
take charge while I am gong.
Senator Brown. [Presiding.] I will give my opening
statement, then call on you, and I will start with Senator
Toomey if Senator Crapo is not back for questions if you are
here. If you are not, then I will start.
Senator Tester. Mr. Chairman, before you leave, I have a
very quick opening statement after Senator Brown, a minute.
Would that be OK?
Senator Brown. Sure.
OPENING STATEMENT OF SENATOR SHERROD BROWN
Senator Brown. Thanks to Chairman Crapo for setting this
hearing up, and welcome, Mr. Hirschhorn, Mr. Daly, and Mr.
Buchanan. Thank you for your role in all of this over the
years.
Congress passed ECRA last year, the Export Control Reform
Act, to strengthen our country's ability to protect technology
that is critical to our national security from being stolen by
countries like China. We did that through creating a permanent
statutory basis for U.S. export controls, which we passed
alongside FIRRMA, and thank you, Mr. Hirschhorn, especially for
your work on that, to get CFIUS more authority to look at a
broader range of transactions. We passed both of these to
strengthen our national security and give us stronger tools to
protect ourselves from countries trying to get their hands on
our most sensitive technologies.
Today, a year later, this hearing will help us to assess
ECRA is being appropriately implemented and enforced and
whether the system has the resources to get the job done. That
oversight of implementation is a very important function of
this Committee.
In ECRA, we included provisions designed to address
emerging and foundational technologies. We know how fast
technology changes. We know we needed tools that would evolve
with those changes. Congress also wanted to make sure that the
identification of these technologies remains an ongoing process
and that new controls would be targeted to technologies that
are considered essential to our national security.
The law also directed Federal agencies to take into account
foreign development and availability of those technologies and
the effect controls would have on the development of
technologies within the United States. We want to protect U.S.
national security priorities through tough and appropriate
export controls. Ultimately, important national security and
law enforcement considerations should, of course, be paramount,
but kept separate from trade and economic considerations.
Unfortunately, as with its treatment of ZTE and Huawei,
this Administration seems to be failing that crucial test.
Although export control decisions can appear to be simple, each
requires a complex policy and legal analysis, as you know, ones
that evolve statutes, regulations, international commitments,
intelligence and law enforcement, industrial base implications,
license administration, foreign availability, and multilateral
and bilateral foreign policy issues. The technologies we are
looking at are often complex, and they are constantly evolving.
Technology that were once sensitive become ubiquitous.
Commercial technologies that are not normally sensitive can
still be applied to new uses or by end-use users of concern in
ways that could threaten our national security. Concerns about
destinations and users and end uses vary widely and change
consistently.
This, in other words, as you all know better than others,
is complicated stuff, and we need to get it right.
As Commerce proceeds with its rulemaking process in
emerging and foundational technologies, this Committee must
ensure that Commerce hews to the standards established in ECRA.
It is hard to have a conversation about export controls and
emerging technologies without addressing the role that China
plays in these areas.
Through its Belt and Road Initiative, its Made in China
2025 Initiative, China executes ambitious plans to develop new
technology and manufacturing capabilities. It is investing
heavily in artificial intelligence and 5G infrastructure. It is
reported to be investing up to $10 billion in a national
quantum information lab, and it is 2 years into an additive
manufacturing plan to create a $3 billion industry by next
year, and we see what additive manufacturing has done in places
like Youngstown, Ohio, and elsewhere in this country.
China is focused on dominating the technology and
manufacturing sectors in the decades to come. That should have
us worried, especially when we remember China's history of
using the same technologies it develops for economic purposes
to also help modernize its military, a key driver of our
efforts in the last couple years to update CFIUS and export
controls. They should remain a focus of our executive agencies
as they set controls and issue licenses under new export
control laws and regulations.
China's sometimes illegal acquisition strategies require a
forceful response from our Government and our allies. In that
sense, the United States is not alone in the issues it faces
from China. That is why as Commerce and other agencies identify
and consider controls when foundational and emerging
technologies, it is important that any new unilateral controls
be implemented with an eye toward multilateral agreements.
Multilateral controls like multilateral sanctions are much more
effective if they are imposed by and with our allies and if
control standards are harmonized as much as possible.
I think all of us on this Committee in both parties are
concerned with the unilateral nature of so much that our
country is doing internationally. This is a case where it
cannot be so.
Senator Tester.
STATEMENT OF SENATOR JON TESTER
Senator Tester. Thank you, Ranking Member Brown, and I want
to thank you and the Chairman for having this hearing. Very,
very quickly, I want to thank all the folks on the panel who
are about to testify. I think it is interesting that there is
nobody from the Administration here, and the fact is that
export control reform implementation is critically important.
Its impacts on national security are important. How we strike a
balance between national security and export competitiveness is
critically important, yet the Administration is not here for us
to ask questions of.
I think this panel is great, and I think you should be
here. But the Administration needs to be here to answer
questions. If we are going to do our job as the legislative
branch with the checks and balances, I do not think Democrats
or Republicans or Independents should tolerate the fact--and
this is not the first Committee hearing this has happened to me
that the Administration does not send somebody here at our
request. And I will make the assumption that the Chairman and
Ranking Member did request people from the Administration to be
here.
Senator Brown. Senator Tester, thank you. I will also
emphasize that to Senator Crapo, to Chairman Crapo. You are
right. When I mentioned in my opening statement about the
importance of oversight, that always should include the people
who are actually administering the laws. Not all of you--some
of you have done that in the past, and your expertise is
really, really important, but that is a big part of it. So
thank you, Jon.
Let me introduce the three panelists, and we will begin.
Mr. Hirschhorn is former Under Secretary for Industry and
Security at Commerce, worked on FIRRMA, worked on ECRA. Thank
you for that.
We will turn to Mr. Daly then as former Deputy Assistant
Secretary for Investment Security at Treasury, and then
conclude with Mr. Buchanan on behalf of the Center for Security
and Emerging Technology.
Mr. Hirschhorn, begin please.
STATEMENT OF ERIC L. HIRSCHHORN, FORMER UNDER SECRETARY FOR
INDUSTRY AND SECURITY, DEPARTMENT OF COMMERCE
Mr. Hirschhorn. Thank you, Senator Brown, Senator Toomey.
It is an honor to be here.
The export control system's job is what I always describe
as ``the other side of the coin'' from that of the Department
of Defense. Defense's job is to make sure that if our soldiers
must go onto the battlefield, they carry the most advanced,
most reliable weapons we can give them. The job of BIS and its
sister agencies is to ensure that our adversaries on that
battlefield do not have the very best. That long has been the
central aim of our export control system, and we seek this
objective by controlling the transfer of sensitive technology
to those who might employ it against our interests.
ECRA governs exports and re-exports of so-called dual-use
technology, technology having recognized civilian as well as
military applications, and of low-level military items. The
existing control system has worked well, and ECRA will improve
it further.
ECRA continues the system's traditional emphasis on
military security and foreign policy. The statute also
expresses a preference for multilateral over unilateral
controls, as Senator Brown mentioned, and cautions against
controls that will adversely affect the U.S. competitive
position in global markets.
Importantly, ECRA requires the executive branch to identify
and control exports of emerging and foundational technologies
that are essential to the national security.
In reality, the executive branch has been controlling
emerging technologies for decades. The perennial problem is
that until a new technology is being applied in fairly specific
ways, it is difficult to write regulations that are
sufficiently precise to be meaningful. For one thing, due
process requires the kind of specificity that one sees in
entries on the Commerce Control List and the U.S. Munitions
List.
And beyond legal considerations, if we unilaterally control
any technology too tightly, whether it is emerging or not,
there is a good chance that we will drive research and
development, and ultimately production as well, offshore. So
the bottom line is that if and when potential military
applications of a new technology begin to jell, it is those
applications that we should control and do so multilaterally,
if that is at all possible.
Foundational technologies are at the other end of the
developmental spectrum in that it may be too late, rather than
too early, to control them effectively. By definition, their
uses are widespread and they typically are available outside
the United States. Often, most or all export restrictions on
them--unilateral as well as multilateral--have been removed or
sharply curtailed.
A frequently cited example is that of semiconductors being
sold to China. Yes, China is seeking cutting-edge chips for
military purposes. Those chips are subject to tight,
multilateral controls, however, and China cannot obtain them
legally.
But China also seeks large volumes of chips and other
commodities whose technology is several generations old,
principally for use in consumer products in furtherance of its
Made in China 2025 effort. These items, and the technology for
their production, are subject to reduced controls, or even de
facto decontrol, by the multilateral groups to which the United
States belongs.
We can recontrol the U.S.-origin technologies unilaterally
and thereby cutoff the sale of the resulting commodities, but
it is far from certain that our allies would agree to do the
same. China prefers U.S. technology. We know that. But if U.S.-
based supplies were unavailable, China doubtless would buy
elsewhere.
I am not saying we should not do this, but I do not think
we should kid ourselves about how difficult it is to do it
effectively.
Given where I spent 7 years until 2 years ago, I am not
going to comment on particular China enforcement cases. As a
general matter, though, I do not think it is sound policy to
treat export controls, which are imposed for military and
foreign policy reasons, as an element of our commercial trade
policy to be bargained over along with sales of beef, chicken,
soybeans, and the like. And it is even worse to treat the
enforcement of export controls in that manner. It sends the
wrong message to those who would violate our laws and put our
country at risk. It places the lives of our uniformed men and
women in jeopardy as well as undercutting our law enforcement
agencies and respect for the rule of law.
So, in conclusion, I hope this Committee will do four
things: give ECRA time to work, and I think it will work well;
continue your valuable oversight of the export control process;
ensure that existing control categories are reviewed regularly
and revised to reflect changing threats as well as evolving
technology; and, finally, give BIS the resources it needs to do
the job that you have given it.
Thank you very much. I will be happy to hear your
questions.
Senator Brown. Thank you, Mr. Hirschhorn.
Mr. Daly, thank you for joining us.
STATEMENT OF NOVA J. DALY, FORMER DEPUTY ASSISTANT SECRETARY OF
TREASURY FOR INVESTMENT SECURITY (2006-2009) AND SENIOR PUBLIC
POLICY ADVISOR, WILEY REIN LLP
Mr. Daly. Excellent. Well, I want to thank Chairman Crapo
and Ranking Member Brown for having me here today, Members of
Committee. I am deeply honored to appear before you today and
thank you for the opportunity to testify. The views I express
today are my own. They do not represent my firm or any clients.
And before I get into sort of the heart of the matters that
this Committee is reviewing today, I wanted to applaud this
Committee for passing ECRA and FIRRMA, excellent bills that
will help this country better hone in and address our
adversaries where they try to acquire U.S. critical technology
through the means of going through a CFIUS process or
otherwise. These pieces of legislation are seminal course
corrections.
In terms of implementation and enforcement of ECRA, I want
to applaud first off this Administration, especially Secretary
Ross and Acting Under Secretary of BIS Nazak Nikakhtar, for
their outstanding work and dedication to the efforts to enforce
U.S. laws, protect U.S. technology, and also grow the U.S.
economy. So good work has been done on implementing ECRA and
BIS. BIS has issued an Advance Notice of Proposed Rulemaking
identifying 14 categories of emerging technologies and has
received and is evaluating over 200 comments to that.
BIS also recently announced that it is going to issue a
Federal registrar on foundational technologies and will issue
very soon a proposed rulemaking identifying the first subset of
emerging technologies.
Since the start of 2017, BIS itself has initiated over
2,000 export control investigations, a 21-percent increase; has
had 89 civil adjudications and 70 criminal prosecutions; and
conducted more than 2,000 end-use checks on technology sales in
more than 65 countries. So it is doing the good work.
So how to establish controls for emerging technology and
foundational technologies while preserving the domestic
innovation? Obviously, this is an important and surgical
exercise that must be done with thorough assessments of U.S.
innovation, their level of maturity in the United States and in
allied nations, and also with foreign adversaries. Assessing
controls requires the engagement of U.S. companies large and
small and the focus of Congress for oversight.
And our U.S. allies and members of multilateral export
control regimes should be willing partners. Ensuring the
protection of intellectual property, broader global security,
and the rule of law creates a platform of trust where
innovation can flourish.
Now, while we must seek and use all our means for
multilateral controls, that does not mean the United States
should not take unilateral action where appropriate. However,
we must preserve a system in the United States where R&D
flourishes. It is critical to our innovation and our future.
Last, I also want to say that identifying emerging and
foundational technologies also has effects, as the Senators
have noted, to CFIUS and foreign investment reviews. Once these
technologies are identified, they are going to be critical
technologies for which for certain investments will require a
mandatory declaration.
So to talk a little bit about the designations of ZTE and
Huawei, as you know, Huawei was designated in May of 2019. The
U.S. Government did so after determining that there was
reasonable cause to believe that it had been involved in
activities contrary to the United States national security and
foreign policy interests. I have known Huawei since my time
running the Committee on Foreign Investment in the United
States since 2007, and so knowing Huawei then and seeing the
actions it has taken since that time, I think BIS'
determination was wholly appropriate.
That said, the President per his recent announcement, BIS
will promptly be taking action to issue certain licenses to
companies that apply, which permit transactions that pose no
national security risk and are not contrary to the United
States foreign policy interests.
The effort to closely scrutinize and restrict transactions
with Chinese entities that pose national security risks is not
limited to this Administration. This Congress has taken
significant action, as noted in the National Defense Authorize
Act Section 889.
So what about the effectiveness of ECRA in addressing China
challenges? I believe that the ECRA-related controls will go a
long way toward improving U.S. transparency and effectiveness
in addressing the challenges related to China and its
persistent diversion tactics. We have seen stronger enforcement
have good progress. In FIRRMA itself and the passage, we have
seen a decrease in China's investments in critical technology.
I myself have been to California and seen first-hand Chinese
involvement, government-controlled entities wanting to seek
investment in our critical technology companies. It is
important we address it because those entities are doing it for
state purposes, not for commercial purposes.
Also with the implementation of ECRA, the U.S. policymakers
will be able to better assess our vulnerabilities of our supply
chains. I can tell you firsthand I have particular clients who
are trying to develop and manufacture in the United States, but
the supply chains to do that technology are not here anymore.
We are in an extremely vulnerable position, and doing this
assessment is critical and necessary to knowing where we are
now and how we need to go in terms of being leaders in
innovation and technology.
That said, possible legislative and oversight
recommendations, in my written testimony I offer a few tools to
address intellectual property theft and broader powers to deal
with government-controlled transactions.
Last, and importantly, oversight by this Committee and
Congress is critically important, and resources. We have big
issues in front of us, and we need to put our resources to it,
and having the person-power to do it critically to do it.
Thank you for the opportunity to appear before you today. I
look forward to your questions.
Senator Brown. Thank you, Mr. Daly.
Mr. Buchanan, please proceed.
STATEMENT OF BEN BUCHANAN, Ph.D., ASSISTANT TEACHING PROFESSOR,
SCHOOL OF FOREIGN SERVICE SENIOR FACULTY FELLOW, CENTER FOR
SECURITY AND EMERGING TECHNOLOGY, GEORGETOWN UNIVERSITY
Mr. Buchanan. Thank you, Ranking Member Brown, for having
me to testify. It is a pleasure to be here. I am an Assistant
Teaching Professor at the School of Foreign Service and a
Senior Faculty Fellow at the Center for Security and Emerging
Technology, both at Georgetown University. My research
specialty is examining how cybersecurity and artificial
intelligence shape international security. As this Committee is
well aware, export controls are legal tools that are applied to
technology. If either the tool or the technology is not a good
fit, export controls will fail.
Given the expertise of my two fellow witnesses on the
nuances of the tools themselves, I believe I will be of most
use to the Committee by talking about some of the technologies
in play and what makes export controls comparatively more or
less suitable to these technologies. As a way of opening our
discussion, I will focus on artificial intelligence because I
think it is one of the most central technologies in play today.
An analogy can help conceptualize AI. One can imagine two
ways of teaching a child to perform a task. The first is to
give very clear instructions in a language the child
understands about what the task is and how it is to be done.
The second is to show the child, through a series of examples,
how the task works and have the child infer important rules and
patterns necessary to succeed. At various points in children's
education, they learn different tasks through each of these
methods.
Traditional software development, and even some older
versions of AI, work in a way that is similar to the first
method. They rely on software developers understanding the
problem in great depth and then imparting this expertise to the
system. For example, in a program designed to play chess, the
software developers may consult with grandmasters to understand
the optimal strategies for a wide range of situations and then
program those ideas into the code. Modern AI systems, known as
machine learning systems, use the second method, the one
involving inference. In a machine learning system, rather than
receive clear instructions about how to do the task, software
developers create an algorithm that determines how the system
should learn. They then provide the algorithm with lots of
relevant data and computational power.
There are thus three parts to a modern machine learning
system: the algorithm, the data, and the computational power.
Together, they form an essential triad, and it is worth
examining each part of this triad for its suitability to export
controls.
It is in vogue to say that data is the new oil. From data,
machine learning systems infer important patterns and nuances
and determine what success and failure look like. It is thus
vital that the data provided to the machine learning system be
plentiful and representative of the problem to be solved in all
of its complexity.
A large part of the reason that companies like Google,
Amazon, and Facebook are successful with the AI systems they
deploy is because they aggregate gigantic amounts of data. In
essence, the large data sets these companies assemble provide
them with a competitive advantage over others. Large companies
based in other nations, such as China's Baidu, Alibaba, and
Tencent, derive similar advantages from their data sets. Export
controls are less valuable in managing this flow of data. This
is both because companies already have an incentive and tools
to secure and not share their assembled data and because export
controls are comparatively ill-equipped for the task relative
to other tools like classification or contracts.
Algorithms are the second part of the AI triad. These
software instructions dictate how the machine learning system
will learn. There are a wide variety of algorithms, each suited
to different kinds of tasks, from classifying images to making
predictions about housing prices, to generation new pictures of
people who look real but do not actually exist. The algorithmic
frontier is rich, and a great deal of progress has been made in
the last 7 years.
The prevailing ethos is that, once an advance is made,
researchers post it online and share it with others. In this
sense, AI research is remarkably open, far more so than the
fierce competition of the technology industry would normally
suggest.
The experience of several decades has shown that Government
efforts to control the export of computer code are usually
futile. More generally, I have doubts about the suitability of
our current list-based export controls, given the changing pace
of technology and the movement of the algorithmic frontier.
This brings us to the last part of the triad: computing. It
is easy to ignore, but it remains vitally important, perhaps
prohibitively so. In the last 7 years, we have witnessed a
revolution in computing power applied to machine learning. One
study by the leading research lab OpenAI indicated that between
2012 and 2018, the computing power applied to top machine
learning systems increased by a factor of 300,000 times; if a
cell phone battery lasted 1 day in 2012 and increased at the
same rate, that battery would now last 800 years.
There is much to discuss about why this increase in
computing power has occurred, but the most salient factor for
our purposes today is that, unlike algorithms and data,
computing power is a function of hardware and not software.
That is, computers are tangible products that are easier to
manage, including with export controls. My judgment is that, to
the degree that export controls are relevant to the problem of
managing AI and other technologies such as 5G, it will be
controls on this hardware component and likely on the hardware
that manufactures specialized computer chips for AI.
To be clear, in order for any such controls to work, they
must be conducted in many cases in a multilateral fashion with
allies, given that a great deal of hardware engineering
expertise is outside the United States.
I thank you again for holding this hearing, and I look
forward to your questions.
Senator Brown. Thank you, Mr. Buchanan.
We will start the questioning with Senator Toomey.
Senator Toomey. Thank you, Senator Brown, and thanks to the
witnesses for joining us.
I think Senator Tester made the point during the course of
his comments that part of the goal here must be to strike the
right balance between limiting exports that would have, you
know, adverse consequences for our country and maintaining our
ability to sell other products around the world. And I am not
sure we are getting that balance right in all cases, so I want
to give you an example of a case that concerns me a bit and get
the reaction of our witnesses.
Lycoming Engines is based in Pennsylvania. They are a
constituent of mine, and they manufacture piston aircraft
engines. They are one of America's leading manufacturers of
piston aircraft engines for general aviation aircraft. And it
is not a great secret to reveal that the technology at the
heart of these piston engines is very old. It is many decades
old. These engines and variations on them have been around for
many, many decades. And it is equally unsurprising that they
are shipped all around the world. Every country has some volume
of general aviation aircraft in the world, and a huge
percentage of these aircraft operate with Lycoming engines.
So it was interesting when folks at Lycoming sought to bid
to provide these very engines on a specific project in China
that involved unmanned vehicles. They determined that they had
a legal obligation to get a license. They applied for the
license to bid on this project, and they were rejected.
Now, it seems to me that what really makes UAVs interesting
and special and dangerous potentially are things like the
software and the sensors and the controls that allow them to be
manipulated remotely. It is also interesting that Air China
operates a fleet of Boeing jets that have vastly more
sophisticated technology than any piston engine for a general
aviation plane. And not only that, there are hundreds of
Lycoming engines that are operating in China in manned
aircraft. These very same engines, the exact same engines, they
are being flown around in China. For instance, they operate the
popular Cirrus SR20 aircraft, which is owned by a Chinese
company that is ultimately owned by the Chinese government. All
right? So a Chinese aircraft company buys these Lycoming
engines every day to fly their planes.
So the idea that this very same engine cannot be sold to a
Chinese company that is involved in developing UAVs, which are
not--on the surface, these UAVs are described s intended to
deliver packages. Anyway, it strikes me that maybe we do not
have this balance right in terms of restricting technology
rather than looking at application. When asked about an
American company providing these engines, the folks at DOD or--
I am not sure if it was DOD or Commerce that suggested an
American company could not sell a screwdriver to the Chinese
effort to build these UAVs.
So I just want to pose this question to our panel, starting
with Mr. Hirschhorn. Are we getting the balance right when we
take such a common, universally operated commodity product like
this old technology piston engine and say, ah, but you cannot
sell it for this purpose?
Mr. Hirschhorn. Well, I am not a bureaucrat anymore, so I
want to try to refrain from giving you a bureaucratic answer.
We are party to a 40-nation agreement called the ``Missile
Technology Control Regime'' that at this point, although
probably it should not any longer, includes unmanned aerial
vehicles. And there are two categories. If they have a certain
payload weight, and can travel a certain distance, they are
very tightly controlled. We also have knowledge of what is
called the ``civil-military joinder,'' namely, that China uses
a lot of technology for both civil and military purposes. And
since Tiananmen in 1989, we have a statutory prohibition,
enacted by the Congress, on any sales of military items or
items for the Chinese military. So when you put all of that
together, I suspect that is Lycoming's problem.
Nevertheless, it may be that in this particular case and in
one-off cases, it would not endanger our national security, but
when you see the web of policies that have to be observed here,
you can see why it is a problem.
Senator Toomey. My question is not really so much whether
the decision was consistent with laws and regulation. It was
more of sort of a theoretical question. Do we have it all right
if this is the outcome that we get?
Mr. Hirschhorn. I think we largely have it right. Whether
we have it right in every case I could not say. Whether we have
it right in this case I could not say. I think if there is a
belief that this is going to ultimately assist the Chinese
military, it is our policy--and maybe it should not be our
policy; that is what you all are here for--not to do anything
that will assist the Chinese military or modernization of the
Chinese military. That is what it is. If it is to change, I
think it would be up to Congress to change it. I doubt that the
Administration, this one or any one, would change it.
Senator Toomey. I am out of time. Thank you.
Senator Brown. Thank you, Senator Toomey.
Mr. Hirschhorn, ECRA, as you know, requires an interagency
process to include giving outside stakeholders an opportunity
to comment as they further define emerging and foundational
technologies. The categories of technologies listed in BIS'
ANPRM are complex technical categories. They will drive global
economies and national security in the coming decades.
What are the most important things BIS should consider when
evaluating controls on these categories of technology?
Mr. Hirschhorn. Well, I think Senator Toomey's point is the
best, which is getting it right, because it is very easy, even
with existing technologies, to over- or under-control them. And
you do not want to do either for the reasons that have been
expressed by Senators and witnesses this morning.
In the work I did for 7 years on export control reform, I
found the input from industry extraordinarily valuable. The
Commerce Department and its sister agencies put forth proposed
regulations and said to industry, How does this work? Does this
work for you? Is it too broad? Are we catching things that are
sold every day all over the world? Are we leaving things out
that we ought to control? We did not get too much industry
input on the last one, but plenty of input where we had it
wrong.
I always used to say when I would speak with industry
groups that, believe it or not, the Government does not always
get it right on its own. So having that input, which I
considered valuable, free, and highly professional, did a great
deal toward making export control reform the success I think it
was. If we cannot get the input of the people who are making
this stuff, who are developing this stuff, we cannot simply
assume that the Government knows enough.
Senator Brown. Does BIS have the resources it needs to
address applications and enforcement and controls?
Mr. Hirschhorn. No, it does not. One of the things I did in
my time there was to beg, borrow, and steal resources wherever
I could get them, from the Congress, from other parts of the
Commerce Department that maybe were a little more flush. It
is----
Senator Brown. And that is a continuing challenge?
Mr. Hirschhorn. BIS is one deep. It is a continuing
challenge. If the engineer who reviews machine tool
applications breaks his leg, you cannot go down the hall and
say to the chemist, ``You are going to do machine tools for the
next 3 weeks.'' It is one deep, and it needs more resources. I
think the budget is around $114 million today. It probably
should be at least $130 million, maybe more.
Senator Brown. Thank you.
Mr. Buchanan, a question for you. I will ask about the best
ways to address China's cybersecurity threats. The
Administration completed a 301 investigation against China in
part because of its government's state-sponsored intellectual
property theft and cyber espionage. As a result of the
investigation, in an effort to bring China to the negotiating
table, the Administration proposed--I am sorry, imposed, not
proposed--imposed tariffs on $260 billion of Chinese imports.
Those tariffs have been in place for a year. Trade talks with
China seem to be at an impasse.
Has any of this gotten the Chinese government to change its
ways with respect to cybersecurity? Do you see signs of that?
Mr. Buchanan. Obviously, it is hard to spot operations that
are meant to be hidden, but I think it is fair to say China
continues to be an aggressive actor in cyberspace and continues
to hack targets in the United States as they perceive suits
their national interest. And this is a pattern that has gone on
for quite a period of time, and I do not see a lot of evidence
that has slowed.
Senator Brown. So answer a bit more broadly. The tariffs on
some areas, it is clear to me tariffs have not changed Chinese
behavior. You sort of speaking expansively make the same claim?
Mr. Buchanan. Yes, I think that is right. I do not claim
expertise beyond cybersecurity, but I have not seen any
indication that in response to tariffs Chinese hacking has
diminished.
Senator Brown. Can you dig down? Is there a way to impose
these tariffs narrowly, or not so narrowly, to change Chinese
behavior in cybersecurity?
Mr. Buchanan. I would be surprised if the Chinese hackers
are responding to tariffs. My sense is those are different
parts of the apparatus. A lot of the Chinese activity we have
seen in the last couple years, at least in public, are
particularly broad operations targeting many millions of
Americans' data in a variety of organizations. That seems to be
removed from the part of the calculus of the Government that
would deal with tariffs.
Senator Brown. OK. Thank you.
Senator Van Hollen.
Senator Van Hollen. Thank you, Senator Brown. I thank all
of you for your testimony today.
Before I ask about export matters, Mr. Buchanan, I am glad
to have you here. You wrote an article in 2016--it was co-
authored--about dealing with Russian interference in our
elections. And you said, and I quote, that ``The United States
should put forth a declaratory policy on the vital importance
of elections, vowing to impose costs on any state that
interferes with the integrity of the process.'' You went on to
say that the United States should ``articulate a policy of
deterrence through cost imposition that would be activated only
if a foreign actor sought to tip an election to one candidate
or introduce significant doubt as to the legitimacy of
democracy.''
I fully agree with that assessment. Senator Rubio and I
have introduced a bipartisan bill that has bipartisan support
that we are trying to get enacted before the 2020 election,
which sets out a very simple proposition, you know: Mr. Putin,
if you get caught interfering in our elections again, you will
face swift, mandatory, and substantial penalties. Is that the
kind of deterrence that you are talking about?
Mr. Buchanan. Yes, Senator. As I think we both agree,
elections are foundational and fundamental to democracy, and my
colleague Michael Sulmeyer and I warned prior to the 2016
election that it was important to make it clear to American
adversaries that this is something we take seriously, something
we seek to protect, and something that, should they decide to
interfere, will be met with consequences. And I think your bill
is a good step in that direction.
Senator Van Hollen. Thank you. I am just going to renew my
request to the Chairman and the Ranking Member, along with
Senator Rubio, that we move forward on the DETER Act in
Committee.
Let me now get to the issue of export controls, and, Mr.
Hirschhorn, and I think listening to all of you, I think I am
on the same page, which is, if we make a determination that
something is in our national security interest, for example, if
we think it is important to put Huawei on the entities list for
the purpose of preventing exports that could strengthen their
5G network, if we make that conclusion as a country, then we
should not then be making tradeoffs with respect to those
national security interests in order to get concessions on
tariffs or other trade-related issues. Would you all agree on
that, starting with Mr. Hirschhorn?
Mr. Hirschhorn. I would agree very emphatically, and I will
add that this is not unique to this Administration. On the
enforcement side, it is unique. But during my service in the
Obama administration, there were always temptations to put
export control issues, national security issues on the table as
part of trade negotiations, and I resisted them successfully.
It looks like the resistance has not been so successful lately.
Senator Van Hollen. Would you all agree with that, though,
that we should not be trading off national security interests
for some kind of concessions on tariffs or a trade issue?
Mr. Daly. National security is also economic security. The
stronger our economy is, the stronger we are as a Nation, the
more we are able to provide for our military and our national
security defenses. So having worked in the National Security
Council, I saw the panoply of issues that come before a
President and come before an Administration. So, broadly
speaking, as long as we are pursuing our goal of national
security, addressing the economic issues is important, too.
Senator Van Hollen. Mr. Buchanan.
Mr. Buchanan. I would agree, Senator. I think we should not
tradeoff between these two goals. I appreciate that there is
overlap, but I think enforcement that seems to vary with the
tenor of trade talks undermines the credibility of that
enforcement.
Senator Van Hollen. Well, let me just give everyone some
examples. The President about a year ago--I should say
Secretary Ross put ZTE--he put a blocking order on ZTE and
stated that this was in the national security interest of the
United States because ZTE had violated Iran sanctions. Within a
short period of time, the President tweeted out, ``I am going
to remove ZTE from the blocking list because my friend
President Xi as me to.'' Example number one.
Number two Huawei, two examples. One, Huawei was also found
to be in violation of sanctions, and as a result, we have asked
the Canadians to arrest the CFO of Huawei. And then the
President says that he would intervene in the arrest of
Huawei's CFO Sabrina Meng Wanzhou if it helps secure a trade
deal with China.
Now, in my view, this is a perfect example, Mr. Hirschhorn,
of what you say is dangerous, because this undermines the rule
of law. If we are going to arrest somebody because they
violated U.S. law, in my view--and I am asking for your
opinion--it is very worrisome, risky, and counterproductive for
the President of the United States to suggest that he is going
to release somebody if he gets a deal or a concession on trade.
Would you agree with that?
Mr. Hirschhorn. I agree with that emphatically. There are
plenty of things that are trade related that can be put on the
table. Law enforcement and national security do not belong
there.
Senator Van Hollen. Let me just say, Mr. Chairman, look, I
agree with a lot of the efforts this Administration is taking
with respect to addressing Chinese theft of technology and the
national security part. I agree with their Huawei policy. But
it is very, very scary to start trading off national security
issues and the rule of law and arresting people with respect to
trade. It is a recipe for getting other countries to grab
Americans and detain and arrest them as part of an effort to
extract trade concessions from the United States. Very
dangerous, and I hope we will all agree that it is a bad idea.
Thank you.
Chairman Crapo. [Presiding.] Thank you.
Senator Jones.
Senator Jones. Thank you, Mr. Chairman.
You know, Mr. Chairman, I am so tempted to follow on this
and talk about how or whether or not all the Mercedes Benz and
BMWs and Nissans and Toyotas are a threat to national security
here, but I am going to resist the urge. I have got plenty of
time to do that in other forums, I think.
So what I do want to talk to the panel about--and thank you
all for being here. I apologize for being late. The higher
education system in my State of Alabama are among the
universities that are leading the way in emerging technologies
and specifically nanotechnology. The University of Alabama in
Birmingham has a Center for Nanoscale Materials and
Biointegration, and they dive into the uses of nanotechnology
and how it can be manipulated for commercial use.
Now, the results of this can transform medical care we
receive but also military flights. So I do not want them to get
in trouble. It is a very sensitive issue. So for each of you,
if you could address a little bit how is the Commerce
Department and the Federal Government working with these
universities and other education systems so that they can
seamlessly and effectively navigate the export control laws? I
will leave that to anybody.
Mr. Hirschhorn. I will give you 30 seconds on it. The
Export Administration Regulations have for many years excluded
from coverage teaching in catalogue courses by universities and
associated labs. Moreover, fundamental research is not covered
by the export control laws. There are a lot of people who
disagree with that. There is some suggestion that there is a
First Amendment need for it. But it is the work of graduate
students on funded projects that really is where there are
problems, and the Federal Government in my experience is quite
willing to work with universities. I think universities are not
always as willing to work with the Federal Government. I think
some of them tend to view export controls as rules for for-
profit businesses and not for universities. I think there
should be closer coordination between universities and the
Federal Government to make sure that, as you say, they do not
get themselves in trouble.
Senator Jones. Right. Anybody want to add anything, either
of you?
Mr. Daly. Yeah, I just think it is critically important to
understand that there are state-led actors who would seek to
get this point of the spear critical technology,
nanotechnology, and bring it back to their own home country. So
being extremely aware of who is involved in what studies and
involved in what projects and who is funding what research is
important. So it is important not only for the continuation of
the great things that are happening in that university and
their innovation capacity, but also making sure they have
control of it for the long term.
Senator Jones. OK. So let me move on to something else that
I have been very involved in with Members of this Committee,
and that is trying to update our anti-money-laundering laws
throughout the systems. In particular, we have been working
with Senators Warner and Cotton and Rounds involving beneficial
ownership, which is a real problem when you are trying to trace
back funds, whether it is in human trafficking, drugs, or
whatever.
So as it pertains to the export control laws, I can also
see where there would be problems with entities who we do not
really know who they are being controlled. How often in your
experiences have you seen firms try to hide their true
ownership in attempts to evade the export control restrictions?
And are there strategies that we can employ? What are they
doing? What can we do better? How can we tighten that up if it
is a problem?
Mr. Daly. Sure. Yeah, thank you, Senator, for that
question. It is an excellent one. In the private sector, I have
seen that occur and also in the course of reviews by the
Committee on Foreign Investment in the United States. Certainly
there are in many attempts to obfuscate ownership and control,
and, thankfully, we have intelligence services here in the
United States, DNI and 15 other intelligence agencies that can
collectively be able to identify who is actually in control and
what levers they are utilizing to either control U.S.
industries or gain information.
Senator Jones. Have you seen the bill that we have got
pending right now and how the data would be collected and
maintained? Have you had a chance to look at that?
Mr. Daly. I have not been able to, but I look forward to
doing it.
Senator Jones. OK. If you would, just take a look at it and
see if there is something that particularly we might need to
tweak a little bit as it pertains to, you know, imports and
exports to try to help better do this. I would appreciate that.
Mr. Daly. Absolutely, Senator.
Senator Jones. Awesome. Thank you, sir.
Thank you, Mr. Chairman.
Chairman Crapo. Thank you, Senator Jones.
And to the witnesses, again, I apologize for having to slip
out earlier. I had to go defend the Banking Committee's
jurisdiction--which we successfully did, by the way.
Let me just conclude the hearing here with a few questions.
My first one is for you, Mr. Hirschhorn. Traditionally, our
export control system has--and I apologize also if this has
already been covered by the other questions you have been
asked. But, traditionally, our export control system has
focused on national security and foreign policy. Should we
expand the focus of our controls to address issues of economic
competitiveness, for example, things like the Made In China
2025 Initiative?
Mr. Hirschhorn. It is tempting. Certainly in the last 75
years since World War II, our export control system has been
focused on three things: national security and foreign policy
in terms of the philosophy, but also multilateralism, which is
essential for effectiveness. And many times our allies as well
as our adversaries have said, ``Oh, you really just want to
impose these controls so you get an economic competitive
advantage.'' We have truthfully denied that. Administrations of
both parties over many decades--and I have been involved in
this area for 40 years--have truthfully denied that. If we are
going to expand export controls to cover economic issues and
economic competitiveness, we are going to have a much harder
time convincing our allies, who are essential to making any
controls work, to go along with us.
So it is a difficult problem. It is a real dilemma.
Chairman Crapo. Thank you.
Mr. Daly, do you have a thought on that?
Mr. Daly. I think we are in a singular period in time where
we have to address China. So if you look at the emerging
technology categories, the 14 categories, in many ways they
mirror China's 2025 strategy of category of industry. So I
agree with the Honorable Mr. Hirschhorn that we do have to
engage and seek multilateral efforts to do it. But we have to
focus on addressing what China is seeking to do and what that
means to our innovation and innovative capacity well in the
future, not only for our companies but for the militaries they
also provide.
Chairman Crapo. All right. Thank you. And, again, Mr. Daly,
China engages in unfair trade practices and it artificially
subsidizes its companies in order to overdevelop and
overproduce in key sectors such as semiconductors in order to
dominate the world marketplace.
In order to protect the economic viability of U.S.
companies, some propose that we should use export control rules
to cutoff the flow of basic commercial technology that the
Chinese need to compete against our companies, even if the
technology has nothing to do with these, as we have been
talking, the foreign national security.
Now, my question is: If we take this approach that we have
been talking about, what is to prevent a non-U.S. company such
as Europe or Japan from simply filling in behind? And we have
seen this issue raised in the semiconductor world recently in
terms of our reactions to China. Again, is the answer simply
that we must work in coalition with our allies before we engage
in this type of export control? Or is there some other aspect
of this that we could utilize?
Mr. Daly. Yes, Senator, that is an excellent question, and
it goes to the whole heart of ECRA and the purpose of it and
getting the balance right in terms of protecting national
security and ensuring economic growth.
Certainly engaging with our multilateral partners to come
up with a combined agreement on what we should and should not
export is critical to those efforts and should be ultimately
fully pursued.
You know, interestingly enough, too, one of the issues to
really focus on here is why is the supply chain being
juggernauted in one particular area? Why does it require us
providing that good in one country? Why aren't there other
opportunities and other places to be able to sell that where we
can have a more balanced equation and less concern about
longer-term national security issues? So I think that is
another consumer we have to take upon is: What have we allowed
our supply chain to be held by?
Chairman Crapo. Thank you.
And, Mr. Hirschhorn or Dr. Buchanan, do either of you have
anything further on this issue you would like to say?
Mr. Hirschhorn. No. I think Mr. Daly has stated it well. I
think that ECRA is right in stressing multilateralism. I once
heard someone say that unilateral controls are like damming
half a river, and I do not think it is a place we should go
except in special circumstances.
Chairman Crapo. Dr. Buchanan?
Mr. Buchanan. I think my colleagues have covered it well.
Chairman Crapo. All right. Thank you.
My next question then is for you, Dr. Buchanan, and I want
to kind of move into big data. I know that this is a little off
topic, but the Banking Committee has been dealing with the big
data issue and on a very broad basis. We have held three
hearings on privacy in that zone, including on how data is used
to segment, score, or otherwise make predictions about
individuals' creditworthiness, employability, or general
reputation.
AI is at the center of this discussion, and I am concerned
with the extent to which individuals' data is collected and
processed without their knowledge, consent, or any real
understanding of its scope. I believe individuals should have
rights over their data similar to those that Europe in the GDPR
has established, including access, control, the ability to
correct, and the ability to delete.
How do AI systems complicate or challenge the ability of
individuals to exercise data rights?
Mr. Buchanan. Well, AI systems excel at processing large
amounts of data, so they increase the incentives for
corporations and other organizations to try to collect that
data because then they can make better use of it with such
systems. They can process it at a scale that otherwise would be
quite difficult.
I think it is fair to say that machine learning technology
is at the core of many of the major tech companies in the world
today as a result of this. So there is a greater incentive to
collect the data if you can do more with it, and AI systems
enable better slicing and dicing of data.
Chairman Crapo. And in order to protect individuals' rights
or essentially enhance individuals' rights to control that I
would like to see us give them over their own personal data,
are there things that we could do or should do legislatively--
that we could do legislatively that would help to mitigate this
ability of AI to overcome those rights?
Mr. Buchanan. Sure. I think it is important that users have
not just some kind of abstract legal consent to something, but
they have a meaningful understanding of how their data is being
used by companies that collect it, which companies are
collecting it, what they do with it.
One example of something that probably deserves--probably
is not on the mind of many Americans is that as data is
collected, even if the data itself is not sold or shared,
inferences from that data can be--so there is a lot of nuance
on the technology there, again, in part enabled by machine
learning systems' ability to parse large amounts of data. And
my sense is that many Americans do not have a good sense of how
that all works. So meaningful consent to data use certainly
would be a good thing.
Chairman Crapo. And that would be including the management
of the data as well as the sale of the data as well.
Mr. Buchanan. That is exactly right, the management of the
data, the security of the data, and also how inferences from
that data are sold or shared for ad targeting, for example.
Chairman Crapo. All right. Thank you. Then one last
question, and I will give--this is also for Mr. Buchanan, but I
will give Mr. Daly and Mr. Hirschhorn an opportunity to comment
on this if you would like. And it is still on the data issue.
Mr. Buchanan, in your testimony you described export
controls as a relatively ineffective tool in stopping the
export of algorithms given the rate of innovation and the fact
that AI is a fairly open resource. You also identified the mass
of personal and behavioral data as the competitive advantage
for large technology companies as opposed to their AI system.
It would seem to me then that the data could also be a real
vulnerability if, for instance, a foreign adversary were to
obtain all of Google's consumer data. These companies are
incentivized to secure their systems, but that may not be
enough. And my question is: What comprehensive privacy controls
or practices could help mitigate the risk of big data being
used in this way?
Mr. Buchanan. I think it is important to disentangle
security and privacy here. So for security, I think you are
quite right to suggest that companies like Google and, indeed,
many American companies that have large data sets are
significant targets of foreign intelligence agencies. We have
seen cases of Chinese hackers targeting Google going back
almost a decade at this point. So the security of that data is
definitely something that is vitally important as you suggest.
Privacy is very important, as your previous question
suggested. I think it is slightly distinct insofar as there are
privacy concerns that do not relate to foreign actors but
relate to the companies themselves. But if your question is
should we be worried about foreign intelligence agencies trying
to seek access to large data sets of Americans held by American
companies, the record unequivocally suggests the answer is,
yes, we should be concerned.
Chairman Crapo. Thank you.
Mr. Daly or Mr. Hirschhorn, do you have anything to add to
those questions?
[Witnesses shaking heads.]
Chairman Crapo. All right. Well, again, I want to thank
you, all of you, for not only coming here today and sharing
your insights and wisdom on this with us, but for the support
and assistance you have given us as we deal with this issue.
These are obviously becoming more and more important and
critical as we move forward to deal with--the obvious example
is China, but to deal with this set of issues across the globe.
So I appreciate you being here today and look forward to
working with you in the future.
And that brings me to this: For Senators wishing to submit
questions for the record, those questions are due in 1 week, on
Thursday, July 25th, and we ask that each of you respond to
these questions if they come in as promptly as you can.
Again, thank you for being here. This hearing is adjourned.
[Whereupon, at 11:08 a.m., the hearing was adjourned.]
[Prepared statements, responses to written questions, and
additional material supplied for the record follow:]
PREPARED STATEMENT OF CHAIRMAN MIKE CRAPO
The hearing will come to order.
No one can dispute that technological advances are of vital
importance to United States progress and development, where progress in
knowledge and innovations undergird the growth of U.S. economic
productivity.
The U.S. China Commission found that about half the U.S. GDP and
two-thirds of its productivity gains is attributable to U.S. technology
improvements.
In August 2018, the President signed the Foreign Investment Review
Modernization Act, called ``FIRRMA,'' and the Export Control Reform
Act, known as ``ECRA'' into law.
FIRRMA is designed to strengthen the existing regulatory
architecture in significant ways to deal with inbound foreign
investments that would have the potential to threaten U.S. national
security interests.
ECRA importantly reauthorizes an otherwise moribund Export
Administration Act, continued only by annual reissuances of
Presidential national security declarations.
It authorizes the Bureau of Industry and Security (BIS) at Commerce
to update controls on exports designed to prevent certain U.S. dual-use
technologies, lower-level military items and other things from ending
up in the wrong hands.
These two important, hugely bipartisan bills were intended, in no
small part, to ensure that with proper controls in place to establish
highly guarded inward and outbound regimes, a productive relationship
between the United States and China is not only possible, but could be
of the highest value in terms of global prosperity and security.
Today's hearing picks up from where the Committee left off when it
last looked at assessing investment controls on technology in its June
4th hearing on ``Confronting Threats from China.''
On June 4th, we examined China's intention to secure global
technological leadership for itself, with a particular emphasis on some
of its inbound foreign direct investment strategies, particularly into
the U.S. semiconductor industry.
Today, the Committee shifts gears slightly to examine control
issues surrounding exports of things outbound from the United States,
and other re-exports or transfers that may occur abroad.
Right now, there is a raft of export control regulation on the
horizon at the Commerce Department.
So far, BIS is actively engaged on two rulemaking fronts covering
``emerging and foundational technologies,'' which include technologies
from such sectors as artificial intelligence, computing, additive
manufacturing, data analytics, robotics, surveillance and a long list
of others.
Importantly, items BIS designates as ``emerging technology'' will
also be deemed to be ``critical technology'' under FIRRMA, and subject
many potential inbound investment deals to CFIUS review notification
requirements.
The current rulemaking under consideration at BIS is not set in
stone.
It is busy pouring over a myriad of industry and government
comments that will inform its application of strict controls over
emerging technologies, which industry will use to understand to whom it
can transfer these technologies, who can otherwise use them and who can
even research them.
The Committee has before it a very accomplished panel of witnesses
assembled to help us pull apart the underlying risks associated with
the United States continuing its robust international economic
relationships, including that with China, against preserving U.S.
technological leadership over these emerging and foundational
technologies and some of the more sensitive items that that would
produce.
In the past, export controls sometimes have not been able to keep
up with innovation, and this problem is exacerbated by today's pace of
advancements, particularly in the `artificial intelligence' sector,
which owing to its nature is itself a difficult sector to control.
Considering that BIS is very unlikely to designate all artificial
intelligence technology, we are fortunate to have Dr. Buchanan here
today to help the Committee better understand what ``artificial
intelligence'' means, how it works, and why or why not certain aspects
are more controllable than others.
Our professional export control experts, Mr. Hirschhorn and Mr.
Daly are expected to offer their assessments on how BIS may establish
controls that address emerging and foundational technologies, while
preserving the innovative capacity of the United States.
______
PREPARED STATEMENT OF SENATOR SHERROD BROWN
Thank you, Chairman Crapo, for holding this hearing, and thank you
to our witnesses for being here today.
Last year, Congress passed ECRA, the Export Control Reform Act,
which provided a permanent statutory basis for U.S. export controls,
alongside and in tandem with FIRRMA, the Foreign Investment Risk Review
Modernization Act, to broaden the range of transactions that the CFIUS
process would assess. Both of these measures exist to serve key U.S.
national security and foreign policy objectives.
Today, nearly 1 year later, this hearing will help us to assess our
current export control regime and whether ECRA is being implemented and
enforced in a system that is resourced to get the job done.
In ECRA, Congress included provisions designed to address emerging
and foundational technologies. In crafting these provisions, Congress
recognized the dynamic nature of technological innovation and the
importance of control and enforcement processes that would evolve with
those changes.
Congress also sought to ensure that identification of these
technologies remains an ongoing and organic process, and that new
controls be limited to technologies that are considered essential to
U.S. national security.
It also directed Federal agencies to take into account foreign
development and availability of those technologies, and the effect
controls would have on the development of the technologies within the
United States.
We want to protect U.S. national security priorities through tough,
appropriate export controls. Ultimately, important national security
and law enforcement considerations should be paramount, but kept
separate from routine trade and economic considerations. Unfortunately,
as with its treatment of ZTE and Huawei, this Administration seems to
be failing that crucial test.
Although export control decisions can appear to be simple, each one
requires complex policy and legal analyses involving statutes,
regulations, international commitments, intelligence and law
enforcement equities, industrial base implications, license
administration, foreign availability, and multilateral and bilateral
foreign policy issues.
The technologies are often complex and evolving. Technologies that
were once sensitive become ubiquitous. Generally nonsensitive
commercial technologies can be applied to new uses or by end users of
concern in ways that can harm our interests. Concerns about
destinations, end users, and end uses vary widely and change
constantly. This is, in other words, complicated stuff. And we must get
it right.
As Commerce proceeds with its rulemaking process on emerging and
foundational technologies, this Committee must ensure that Commerce
hews to the standards established in ECRA.
It's hard to have a conversation about export controls and emerging
technologies without addressing the role China plays in these areas.
Through its Belt and Road Initiative and Made in China 2025
initiative, China is executing ambitious plans to develop new
technology and manufacturing capabilities. It is investing in
artificial intelligence and 5G infrastructure. It is reported to be
investing $10 billion in a national quantum information lab. And it is
2 years into an additive manufacturing plan to create a $3 billion
industry by next year.
China is laser-focused on dominating technology and manufacturing
sectors in the decades to come. China's history of diversion of dual-
use items to help modernize its military and its civil-military fusion
policies were a key driver of our efforts to update CFIUS and export
controls last year. They should remain a focus of our executive
agencies as they set controls and issue licenses under new export
control laws and regulations.
China's sometimes illegal acquisition strategies require a forceful
response from the U.S. Government and our international allies. In that
sense, the United States is not alone in the issues it faces from
China.
That's why, as Commerce and other agencies identify and consider
controls on emerging and foundational technologies, it's important that
any new unilateral controls be implemented with an eye toward
multilateral agreements.
Multilateral controls--like multilateral sanctions--are much more
effective if they are imposed by and with our allies, and if control
standards are harmonized to the degree they can be.
Thank you. I look forward to hearing from our witnesses.
______
PREPARED STATEMENT OF ERIC L. HIRSCHHORN
Former Under Secretary for Industry and Security
Department of Commerce
July 18, 2019
Chairman Crapo, Ranking Member Brown, and Members of the Committee,
I am honored to be asked to share my thoughts on a number of critical
current issues in U.S. export controls. My involvement in the field
spans more than 40 years and includes service in the Administrations of
Jimmy Carter and Barack Obama, with three decades of private law
practice in between. Although I provide some export control assistance
to private clients these days, my comments here reflect my personal
opinions only.
The Chairman's invitation requests my ``assessment of current
implementation and enforcement of ECRA, including related regulations,
and how the United States may establish controls that address emerging
and foundational technologies while preserving domestic innovation,''
as well as my thoughts about ``recent designations of ZTE, Huawei, and
other Chinese technology companies'' and, ``[w]ith respect to China,
including the persistent diversion challenges it poses, [my]
perspectives on whether . . . emerging ECRA-related control structures
in the United States will be effective in confronting these
challenges.'' Finally, you ask for ``any other legislative or oversight
recommendations'' I might have to offer. I will do my best to respond
to each of these requests.
When I had the honor of serving as head of the Bureau of Industry
and Security, I often described BIS' job as being the other side of the
coin from that of the Department of Defense. DOD's job is to ensure
that if our soldiers have to go onto the battlefield, they carry the
most advanced, most reliable weapons and other equipment that we can
give them. The job of BIS and its sister agencies is to ensure that our
adversaries on that battlefield do not have the very best. That long
has been the central aim of our export control system.
We seek this objective by controlling the transfer of sensitive
technology to those who might employ it against our interests. The
Export Control Reform Act--ECRA--wisely points out, though, that the
imposition of controls should come ``only after full consideration of
the impact on the economy''\1\ and on U.S. competitiveness in global
markets,\2\ as well as consideration of whether the technology in
question is ``widely available from foreign sources.''\3\
---------------------------------------------------------------------------
\1\ Export Control Reform Act of 2018 (ECRA), Pub. L. No. 115-232,
Sec. 1752(1), 132 Stat. 2210 (codified at 50 U.S.C. Sec. 4811(1)).
\2\ ECRA Sec. 1752(3) (codified at 50 U.S.C. Sec. 4811(3)).
\3\ ECRA Sec. 1752(6) (codified at 50 U.S.C. Sec. 4811(6)).
---------------------------------------------------------------------------
Let me note parenthetically that in my 40 years of involvement with
export controls, I have observed that although there can be vigorous
disagreements about control policies, individual licenses, and the
like, the disputes are decidedly not partisan. The Obama
administration's Export Control Reform initiative offers a good
example. Some Democrats criticized what we were doing and many
Republicans were supportive. Indeed, the chairman of the House Foreign
Affairs Committee, a long serving Republican Member, complained to me
at one point that we were not moving quickly enough.
ECRA Implementation and Enforcement
ECRA was enacted last August. Like most statutes that address
ongoing issues, it does not have an expiration date. This means that
its passage ended a decades-long pattern in which the Export
Administration Act of 1979 would expire, the President would continue
the Commerce Department's export control authorities under the
International Emergency Economic Powers Act, renewing the authorities
annually, until Congress revived the Export Administration Act, the
export act would expire again, and the pattern would repeat itself.
ECRA relates to exports from the United States, as well as to
subsequent reexports and transfers abroad. It establishes a control
system for so-called dual use items--those having recognized civilian
as well as military applications--and low-level military items. That
system is administered by the Department of Commerce in consultation
with the Departments of Defense, State, and Energy.
ECRA was enacted with a companion statute called the Foreign
Investment Risk Review Modernization Act, or FIRRMA, which amends the
process for reviewing
foreign investments that are inbound into the United States.\4\ The
inbound investment review is conducted by CFIUS--the Committee on
Foreign Investment in the United States. One goal of the CFIUS process,
which also is a long time goal of the export control system, is to
ensure that a foreign person who invests in the United States will not
thereby gain access to technology that we would not allow to be
exported directly to his or her home country.
---------------------------------------------------------------------------
\4\ Foreign Investment Risk Review Modernization Act of 2018
(FIRRMA), Pub. L. No. 115-232, Sec. Sec. 1701-1728, 132 Stat. 2174.
---------------------------------------------------------------------------
The original FIRRMA legislation would have directed CFIUS to draw
up a sensitive technologies list that would have been similar, but not
identical, to the lists that already are part of the existing export
control system. I and others ultimately convinced the sponsors of the
FIRRMA bill that rather than have a body without export control
expertise set up a potentially duplicative list, the measure should
strengthen the existing export control system. Given the already-felt
need of many in Congress to enact permanent export control legislation,
Congress sensibly came up with ECRA as the solution.
What does ECRA do? To a considerable degree, it codifies the
existing Commerce Department control mechanism, including the changes
made by the Export Control Reform initiative. For that reason, ECRA
requires few substantive regulatory changes aside from those involving
emerging and foundational technologies, which I'll address in a moment.
ECRA sets out a statement of policy that continues the traditional
emphasis on military security and foreign policy, including prevention
of the proliferation of weapons of mass destruction, strengthening our
defense industrial base, and focusing controls ``on those core
technologies and other items that are capable of being used to pose a
serious national security threat to the United States.''\5\ It also
expresses a preference for multilateral controls over unilateral
controls, cautions against control measures that will adversely affect
the U.S. competitive position in global markets, calls for regular
updates of U.S. controls, encourages strong enforcement, and notes the
complementarity of the export control and CFIUS processes in
``controlling the transfer of critical technologies to . . . foreign
persons.''\6\
---------------------------------------------------------------------------
\5\ ECRA Sec. 1752(1)-(2) (codified at 50 U.S.C. Sec. 4811(1)-(2)).
\6\ ECRA Sec. 1752(3)-(10) (codified at 50 U.S.C. Sec. 4811(3)-
(10)).
---------------------------------------------------------------------------
Substantively, ECRA continues in force the broad existing powers of
the Commerce Department to administer and enforce controls on exports
of dual-use and lower-level military items, as well as restrictions on
activities of U.S. persons in support of foreign military and
intelligence activities.\7\ ECRA also clarifies and expands
considerably the tools available to BIS' Office of Export
Enforcement.\8\
---------------------------------------------------------------------------
\7\ ECRA Sec. 1753 (codified at 50 U.S.C. Sec. 4812).
\8\ ECRA Sec. Sec. 1754(a)(10), 1761 (codified at 50 U.S.C. Sec.
Sec. 4813(a)(10), 4820).
---------------------------------------------------------------------------
ECRA requires that licensing decisions take into account whether
denial of a proposed export will have a significant negative effect on
the U.S. defense industrial base, as well as whether approval would
engender ``significant production of items relevant for the defense
industrial base outside the United States.''\9\
---------------------------------------------------------------------------
\9\ ECRA Sec. 1756(d) (codified at 50 U.S.C. Sec. 4815(d)).
---------------------------------------------------------------------------
Emerging and Foundational Technologies
ECRA requires the executive branch to identify, and the Commerce
Department to control exports of, ``emerging and foundational
technologies that . . . are essential to the national security of the
United States'' and are not already controlled under one of our
existing export control programs.\10\ The statute directs that this
effort take into account such criteria as national security, foreign
availability, whether a unilateral control would harm domestic research
and innovation, the effect on our defense industrial base, and the
willingness of our allies to impose similar restrictions. For a host of
reasons, I am uncertain whether this exercise will yield significant
results.
---------------------------------------------------------------------------
\10\ ECRA Sec. 1758 (codified at 50 U.S.C. Sec. 4817).
---------------------------------------------------------------------------
The Commerce Department has thus far taken two initial regulatory
steps in carrying out this mission. First, an Advance Notice of
Proposed Rulemaking (ANPRM), seeking comments on possible emerging
technology controls, was published in November 2018.\11\ The comment
period closed in January and a substantial number of comments were
received. I'm told that further action on that rulemaking, as well as
on a companion ANPRM on foundational technologies, was delayed
substantially by the Government shutdown earlier this year but that
progress is being made on both fronts.
---------------------------------------------------------------------------
\11\ 83 Fed. Reg. 58201 (Nov. 19, 2018).
---------------------------------------------------------------------------
Second, BIS promulgated a number of new and revised export controls
on emerging technology items in May.\12\ These had been agreed to in
the Wassenaar Arrangement, which is a group of about 40 countries that
agree upon and then implement ``national security'' controls. Strictly
speaking, the controls promulgated in May aren't within the new
procedural framework established by ECRA but I suspect they are
indicative of the kinds of controls we will see on emerging
technologies.
---------------------------------------------------------------------------
\12\ 84 Fed. Reg. 23886 (May 23, 2019).
---------------------------------------------------------------------------
Emerging technologies. I agree wholeheartedly that we should impose
appropriate controls on emerging technologies with national security
implications and should do so as early in their development as
practicable. Indeed, that is what the executive branch has been doing
for decades. For controls to be truly effective, they should be adopted
by our allies in the four multilateral export control regimes as well
as unilaterally by the United States.\13\
---------------------------------------------------------------------------
\13\ In addition to the Wassenaar Arrangement, these are the
Nuclear Suppliers Group, the Australia Group (chemical and biological
weapons), and the Missile Technology Control Regime.
---------------------------------------------------------------------------
The principal problem with regulating an emerging technology is
that until it is being applied in fairly specific ways, it's difficult
to write regulations that are sufficiently precise to be meaningful to
regulators and exporters. By way of example, the Commerce Department
can't very well promulgate a regulation that just says, ``Don't send
advanced materials technology to China'' unless that regulation sets
out particular applications and technical parameters. A general or
generic prohibition isn't specific enough to inform exporters what can
and cannot be sent to China, or to tell enforcement agents,
prosecutors, judges, or juries when an exporter has broken the law.
Less than a month ago, the Supreme Court reminded us that ``[i]n
our constitutional order, a vague law is no law at all''\14\ and that
``[v]ague laws contravene the first essential of due process of law
that statutes must give people of common intelligence fair notice of
what the law demands of them.''\15\ In short, due process requires that
a regulation set out clearly and specifically the boundary between what
is lawful and what is not. That in turn requires the kind of
specificity that one sees in entries on the Commerce Control List and
the U.S. Munitions List.
---------------------------------------------------------------------------
\14\ United States v. Davis, 139 S. Ct. 2319, 2323 (2019) (Gorsuch,
J.).
\15\ Id. at 2325 (interior quotation marks omitted).
---------------------------------------------------------------------------
And beyond due process considerations, if we unilaterally control
any technology too tightly, whether it's emerging or well on the way to
being in common use, there's a good chance that we will drive research
and development, and ultimately production as well, offshore. This is
not idle speculation, as we have seen very tight U.S. export controls
engender the development of foreign competition in such sectors as
machine tools, commercial space, and commercial thermal imaging.
Further, we saw in the course of Export Control Reform how
important it is to seek private sector input on proposed controls. The
Government's technical experts are knowledgeable but they don't always
have full information on what currently is available in the global
marketplace. Input from industry helped ensure that our rules, when
published in final form, neither over-controlled nor under-controlled
the technologies in question.
So when it comes to controlling emerging technologies, the sensible
approach is for the Government to do what it already has been doing for
decades and what ECRA is telling it to do now: Follow emerging
technologies, with a particular eye toward applications that would give
an adversary a military or intelligence advantage. If and when those
potential applications begin to become concrete (and hence to be
suitable subjects for legally enforceable regulation), control those--
if at all possible, in the context of the multilateral export control
groups rather than unilaterally. Securing agreement for multilateral
control is difficult, time-consuming work but it is the most promising
route to success.
Foundational technologies. In a sense, foundational technologies
are at the opposite end of the developmental spectrum from emerging
technologies. The problem with an emerging technology is that it can be
too soon to control it if specifics are not available. The problem with
foundational technologies, by contrast, is that it may be too late to
control them effectively. By definition, their uses are widespread--so
much so that they're well known and typically available from numerous
sources outside the United States. In many instances, most or all
export restrictions on them--unilateral as well as multilateral--have
been lifted or sharply curtailed.
A frequently cited example is that of semiconductors being exported
to China. Yes, China would love to get its hands on cutting-edge chips
and use them for military purposes. Those high-end chips are subject to
tight, multilateral export controls, however, and China cannot obtain
them legally.
But China also is very happy to buy large volumes of chips and
other commodities whose technology is several generations old, for use
in consumer products in furtherance of its Made in China 2025 effort.
These items, and the technology needed for their production, no longer
are viewed as having significant military utility and so are subject to
reduced controls, or even de facto decontrol, by the multilateral
groups to which the United States belongs. The United States presumably
can recontrol the U.S.-origin technologies and cutoff the sale of the
resulting commodities to China but it's far from certain that our
allies would agree to do the same. China would prefer to purchase the
products that use U.S. technology because they know that our goods are
the most reliable, but if U.S.-based supplies were to become
unavailable, China would shift its purchases to other sources.
The problem with controlling foundational technologies, then, is
their ubiquity. Simply put, the United States ordinarily isn't the only
potential source, so preventing China from acquiring these items made
here or based on our technology may hurt U.S. companies, U.S. workers,
and our overall defense industrial base more than it impairs the
Chinese effort to dominate us economically.
Underlying the idea of restricting foundational technology exports
is the longstanding question whether export controls should be used to
address only concerns about military security and foreign policy or
should be expanded to address concerns about economic security or
economic competitiveness. Since the end of World War II, U.S. export
controls have been focused on military and foreign policy concerns.
ECRA continues this approach, stating in section 1752(1)\16\ that
export controls should be focused on contributions to the military
potential of possible adversaries and on furthering the foreign policy
of the United States.
---------------------------------------------------------------------------
\16\ Codified at 50 U.S.C. Sec. 4811(1).
---------------------------------------------------------------------------
Other countries, including not only adversaries but also some of
our closest friends, have voiced suspicions over the years that our
controls are intended to advance U.S. commercial and economic goals.
Successive U.S. Administrations of both parties--truthfully, in my
view--have denied this forcefully. Although the focus of our controls
could be expanded, doing so would represent a sharp break from past
policy, would be inconsistent with the ECRA legislation that Congress
passed less than a year ago, and would make it more difficult to
convince our allies to follow our lead.
China Enforcement Issues
During my time in the Obama administration, I was involved in the
development of the Commerce Department case against ZTE. I also was
aware of the beginnings of Commerce's Huawei investigation. I think it
best to avoid specific comments on these two matters or other
individual cases that were pending during my tenure. I will comment,
though, on the high degree of professionalism among BIS' enforcement
agents and lawyers. I cannot imagine that the cases they developed
against these or any other defendants were politically motivated or
otherwise not strictly ``by the book.'' They may not always be right
but their motivations are bona fide.
As a policy matter, I don't think it's a sound idea to treat export
controls--which are imposed for military security and foreign policy
reasons--as an element of our commercial trade policy, to be bargained
over along with sales of beef, chicken, soybeans, and the like. It is
even worse to treat the enforcement of export controls in that manner.
Public horse trading of national security and law enforcement for
sales of agricultural commodities sends the wrong message to those who
would violate our laws and put our country at risk. Such a course of
action places the lives of our uniformed men and women in jeopardy as
well as undercutting the mission of our law enforcement agents and
public respect for the rule of law.
Other Issues
Like my friend and former Commerce colleague, Kevin Wolf, who
testified here about 6 weeks ago, I think that your best course of
action is fourfold.
First, give ECRA time to work--and I expect that it will
work well.
Second, continue the Committee's valuable oversight of the
export control process, including ECRA implementation.
Third, ensure that existing control categories are reviewed
regularly and, with industry input, revised to reflect changing
threats as well as evolving technology development and
applications.
Finally, give BIS the resources it needs to do the job that
Congress has assigned to it. This final point is important.
BIS' talented and dedicated staff cannot carry out their
responsibilities without adequate resources. The budget was too
small when I was there and the substantial workload increase
since then has greatly outstripped the modest resource increase
that has accompanied it. Do not starve this valuable operation,
which punches far above its weight.
Thank you again for your interest in this important topic. I'd be
glad to respond to any questions the Committee may have.
______
PREPARED STATEMENT OF NOVA J. DALY
Former Deputy Assistant Secretary of Treasury for Investment
Security (2006-2009)
Senior Public Policy Advisor, Wiley Rein LLP
July 18, 2019
Chairman Crapo, Ranking Member Brown, and Members of the Committee,
I am honored to appear before you today and thank you for the
opportunity to testify. The views I express today are my own and do not
reflect those of my firm, Wiley Rein LLP, nor any client. My views are
based on my over 20 years of experience in and outside of Government.
They include service at the U.S. Treasury Department administering the
Committee on Foreign Investment in the United States (CFIUS), at the
National Security Council, on the Senate Finance Committee, and in
other positions at the U.S. Department of Commerce, as well as work in
the private sector addressing trade, export control, sanctions, foreign
investment and multiple national security matters. Again, thank you for
the opportunity to testify.
My testimony today will address five matters that this Committee is
exploring regarding the implementation of U.S. export control reforms,
notably those under the Export Control Reform Act of 2018 (ECRA).\1\ My
presentation:
---------------------------------------------------------------------------
\1\ Subtitle B, Part 1, P.L. 115-232.
I. Provides an assessment of the current implementation and
---------------------------------------------------------------------------
enforcement of ECRA, including related regulations;
II. Describes how the United States may establish controls that
address emerging and foundational technologies while preserving
domestic innovation;
III. Addresses recent designations of Zhongxing Telecommunications
Equipment Corporation (ZTE), Huawei Technologies Co. Ltd.
(Huawei), and other Chinese technology companies;
IV. Discusses whether ECRA-related control structures in the United
States will be effective in confronting the challenges raised
with respect to China, including the persistent diversion
challenges China evokes; and
V. Proposes possible legislative or oversight recommendations
regarding the topics covered today.
Before addressing these matters, I want to applaud this Committee
for its work in passing ECRA as well as the Foreign Investment Risk
Review Modernization Act of 2018 (FIRRMA).\2\ These two pieces of
legislation are historic and seminal ``course corrections,'' providing
the United States with the ability to address the actions of
adversarial powers and persons more adroitly and comprehensively in a
world where economic and cyber security and technological leadership
are pivotal to core and peripheral U.S. national and economic security
considerations as well as global peace and order.
---------------------------------------------------------------------------
\2\ Title XVII, P.L. 115-232.
---------------------------------------------------------------------------
I. Assessment of Current Implementation and Enforcement of ECRA,
including Related Regulations
In order to appropriately frame this topic, it's important to take
account of the accomplishments of this Administration and Congress that
have been undertaken to address U.S. economic and national security
vulnerabilities. These include the development and passage of ECRA,
FIRRMA, provisions within the National Defense Authorization Act for
Fiscal Year 2019 (NDAA),\3\ addressing telecommunication and video
surveillance vulnerabilities, Section 232 investigations under the
authority of the Trade Expansion Act of 1962, increased enforcement
activities by BIS, and executive orders (E.O.) on supply chain security
\4\ as well as those that seek to stimulate U.S. manufacturing and job
growth. I applaud the leadership of Senator Crapo in this Committee in
passing multiple national security legislative actions and oversight,
as well as that of Senator Brown, including his proposed bill to
safeguard matters impacting economic and national security.
---------------------------------------------------------------------------
\3\ P.L No: 115-232.
\4\ E.O. 13873, ``Securing the Information and Communications
Technology and Services Supply Chain.''
---------------------------------------------------------------------------
As Commerce Secretary Wilber Ross recently noted, ``[e]conomic
security is essential to national security'' and safeguarding our
technology ``is not easy, since the boundaries between civilian and
military technologies become ever more narrow as technologies are
increasingly omnipresent.''\5\
---------------------------------------------------------------------------
\5\ Remarks by U.S. Commerce Secretary Wilbur L. Ross at the Bureau
of Industry and Security Annual Conference on Export Controls and
Security, July 9, 2019.
---------------------------------------------------------------------------
The efforts of this Administration, and specifically Secretary Ross
and acting Under Secretary for the Bureau of Industry and Security
(BIS) Nazak Nikakhtar are to be greatly lauded and supported. Given the
tasks before them \6\ and the degree of increased vulnerabilities to
U.S. technology, infrastructure and innovation, it is critical that
additional resources and support be provided to safeguard U.S. national
security and ensure the rapid implementation of new programs.
---------------------------------------------------------------------------
\6\ Since the start of 2017, BIS has initiated 2,284 export control
investigations, a 21 percent increase in the number of cases opened
from the previous two-and-a-half years.
---------------------------------------------------------------------------
Focusing on the implementation of ECRA, on November 19, 2018, BIS
issued an Advance Notice of Proposed Rulemaking (ANPRM) requesting
public comment on identifying 14 categories of ``emerging technology.''
The full list of emerging technologies that BIS identified is available
at: https://www.federalregister.gov/documents/2018/11/19/2018-25221/
review-of-controls-for-certain-emerging-technologies.\7\
---------------------------------------------------------------------------
\7\ The list includes: 1. Biotechnology; 2. Artificial intelligence
(AI); 3. Position, Navigation and Timing (PNT) technology; 4.
Microprocessor technology; 5. Advanced computing technology; 6. Data
analytics technology; 7. Quantum information and sensing technology; 8.
Logistics technology; 9. Additive manufacturing (e.g., 3D printing);
10. Robotics; 11. Brain-computer interfaces; 12. Hypersonics; 13.
Advanced Materials; and 14. Advanced surveillance technologies.
---------------------------------------------------------------------------
As BIS relayed, the categories of emerging technology were provided
for illustrative purposes and comments to them were not restricted just
to those categories. BIS noted further that any controls on identified
emerging technologies would not apply broadly to the general categories
listed in the ANPRM, but rather on a narrow and meaningful subset of
those categories.
The ANPRM summarized BIS' objective as follows:
As controls on exports of technology are a key component of the
effort to protect sensitive U.S. technology, many sensitive
technologies are listed on the Commerce Control List (CCL),
often consistent with the lists maintained by the multilateral
export control regimes of which the United States is a member.
Certain technologies, however, may not yet be listed on the CCL
or controlled multilaterally because they are emerging
technologies. As such, they have not yet been evaluated for
their national security impacts. This advance notice of
proposed rulemaking (ANPRM) seeks public comment on criteria
for identifying emerging technologies that are essential to
U.S. national security, for example because they have potential
conventional weapons, intelligence collection, weapons of mass
destruction, or terrorist applications or could provide the
United States with a qualitative military or intelligence
advantage. Comment on this ANPRM will help inform the
interagency process to identify and describe such emerging
technologies. This interagency process is anticipated to result
in proposed rules for new Export Control Classification Numbers
(ECCNs) on the CCL.
Commerce does not seek to expand jurisdiction over technologies
that are not currently subject to the Export Administration
Regulations (EAR), such as ``fundamental research'' described
in 734.8 of the EAR. For purposes of this ANPRM, Commerce
does not seek to alter existing controls on technology already
specifically described in the CCL. Such controls would
generally continue to be addressed through multilateral regimes
or interagency reviews.
Following the issuance of the ANPRM, I understand that BIS received
just over 230 comments and is currently evaluating them and working
through an interagency process to identify controls, where warranted.
BIS recently announced that an ANPRM for ``foundational''
technologies will be issued very soon, and that a proposed rule
identifying a first subset of controls on ``emerging'' technologies
will be forthcoming as well. Further, BIS has emphasized throughout
this regulatory process that the controls that will be implemented will
be thoughtful, targeted, and focused on ``choke points,'' as opposed to
broad, blanket controls on technologies initially identified in the
ANPRM process. BIS has emphasized the critical importance of industry
input, and that it is taking into account all of the comments that have
been submitted on emerging technologies.
BIS has additionally made clear that achieving multilateral
controls on these technologies would make the most sense and that the
process of identifying and implementing controls on emerging and
foundational technologies will be ongoing, consistent with BIS' normal
rulemaking approach. Toward that end, it should be noted that as a
result of a Wassenaar Plenary in 2018, in May 2019, BIS published a
final rule that revises the CCL to implement certain changes made to
the Wassenaar Arrangement List of Dual-Use Goods and Technologies
maintained and agreed to by governments participating in the Wassenaar
Arrangement on Export Controls for Conventional Arms and Dual-Use Goods
and Technologies (Wassenaar Arrangement).\8\
---------------------------------------------------------------------------
\8\ This rule added five recently developed or developing
technologies (i.e., emerging technologies) that are essential to the
national security of the United States to the EAR's CCL, including
discrete microwave transistors (a major component of wideband
semiconductors), continuity of operation software, post-quantum
cryptography, underwater transducers designed to operate as
hydrophones, and air-launch platforms.
---------------------------------------------------------------------------
Taking on this mandate under the ECRA is no small task. In my view,
addressing these matters is one of the most critical actions this
Administration will undertake, and I believe that good progress is
being made given the critical nature of the efforts, the extent of
industry input, and the domestic and global impact of BIS'
determinations.
II. Establishing Controls that Address Emerging and Foundational
Technologies While Preserving Domestic Innovation
The establishment of export controls that address emerging and
foundational technologies should be a surgical exercise, and as I
alluded to earlier, in my view probably one of the most important
undertakings affecting U.S. technology innovation and leadership now
and well into the future. The task of identifying emerging technologies
is necessarily complex because these technologies are currently being
developed (hence ``emerging'') as opposed to more mature technology
(e.g., foundational).
The process that BIS should and is undertaking to identify emerging
technologies includes an assessment of U.S. innovation in various
categories of emerging technologies, the level of maturity of these
technologies in the United States and in allied nations, and foreign
adversarial uses of these emerging technologies. Once BIS and its
interagency partners develop a good understanding of these facts, they
can better assess what types of controls, if any, make sense for
particular emerging technologies with the goal of ultimately also
gaining agreement on multilateral controls.
While this task requires an understanding of which technologies are
broadly disseminated and which are not, it does not mean that
technology which is available outside of the United States should
automatically be excluded from targeted unilateral actions to control
it, where appropriate.
U.S. allies and members of multilateral export control regimes
should be willing partners. Ensuring the protection of intellectual
property, broader global security and the rule of law creates a
platform of trust where innovation can flourish. Without such a
platform and without such unity, clearly the United States will need to
take certain unilateral actions. It is my hope that where the United
States sees a necessity to protect particular emerging and foundational
technologies, our allies will step up and work with us. We should all
encourage active participation and support by our allies.
Currently, the United States has four multilateral regimes for
export controls: the Wassenaar Arrangement; the Australia Group; the
Nuclear Suppliers Group; and the Missile Technology Control Regime.
Through each of these regimes, countries identify the items to control
(i.e., products, software, and technology), but the controls must be
implemented in national legislation. More specifically, while countries
multilaterally agree on controls of specific items, all countries have
divergent licensing policies on their exports, some with stringent
policies restricting exports, and some with more relaxed policies. This
issue can frustrate the purpose of a multilateral regime because
companies facing more stringent policies in certain countries cede
global market share of the controlled items to companies in countries
with more relaxed policies. The resultant pressure on countries to
protect market share often leads to an underutilization of export
control authority. This is not to mention that the controls themselves
are not effective when countries have different licensing policies.
I understand that BIS is, however, actively engaging with like-
minded partners to establish a working group, at the leadership level,
to discuss coordinating policies on emerging technologies so that U.S.
policies of control--licensing review--are consistent across countries,
and that there is better information-sharing among countries as to what
items are being exported to what countries, and what items are facing
broader export restrictions. Export controls need to be harmonized if
they are to be effective.
Addressing controls on emerging and foundational technologies also
requires engagement with U.S. companies large and small, the focus of
Congress to provide resources and oversight, and frankly a certain
degree of patriotism. U.S. companies must be clear eyed in knowing that
certain potential ``business'' partners actually represent the
interests of foreign governments who will use their technology and
know-how to the economic and military detriment of the United States
and our allies.
That said, it is important that we have a system where R&D works
here in the United States, but also that key technology does not leave
our shores, especially where there is a national security/military
nexus. Further, placing appropriate controls on emerging and
foundational technologies should be undertaken to address China's
``Made in China 2025'' initiative. This initiative/plan emphasizes
China's priorities for high-tech industries as relayed in the 13th Five
Year Plan. The industries that China has identified include: 1) new
advanced information technology; 2) automated machine tools & robotics;
3) aerospace and aeronautical equipment; 4) maritime equipment and
high-tech shipping; 5) modern rail transport equipment; 6) new-energy
vehicles and equipment; 7) power equipment; 8) agricultural equipment;
9) new materials; and 10) biopharma and advanced medical products.
Last, the identification of emerging and foundational technologies
will also impact the work of CFIUS and its Pilot Program mandatory
declarations. Prior to the enactment of FIRRMA, CFIUS was essentially a
voluntary process, and CFIUS was authorized to review only transactions
that could result in foreign control of a U.S. business. However, under
FIRRMA and the Pilot Program, CFIUS is now able to review certain
noncontrolling investments in U.S. critical technology companies,
including any acquisition of an equity interest that affords a foreign
person with access to specified information or governance rights.
Transactions covered under the Pilot Program include any investment in
a U.S business engaged in critical technology that operates in 1 of 27
specifically identified protected industries (Pilot Program
Industries). If a transaction is covered by the Pilot Program, failure
to file a ``declaration'' or a full CFIUS notice 45 days prior to
completion of the transaction could result in civil penalties up to the
value of the transaction. Once identified by BIS, emerging and
foundational technologies will also be considered critical technologies
for the purpose of mandatory CFIUS declarations.
While at Treasury running the CFIUS process, I saw first-hand the
limitations of the voluntary process where actors acquired new and
critical technologies outside of CFIUS' purview. Now with ECRA and
FIRRMA, we can better safeguard the loss of our critical technologies
(including emerging and foundational technologies) to those who would
do harm to our economic and national security. Since my service in
Government, I have observed an increasing number of transactions
involving Chinese parties where the technology at issue could be viewed
to present a lower threat, but the actual threat posed by the
transaction related to vulnerabilities in the U.S. supply chain. Such
vulnerabilities augment the ability of rogue actors to leverage the
U.S. supply chain thereby raising national security concerns,
including: undercutting direct competitors; eroding the existing U.S.
technology of acquired companies; impacting the availability of
upstream inputs; and undermining the ability of downstream purchasers
and producers to compete.
Thus, CFIUS is under increased pressure to evaluate supply chain
factors in its analysis and must also account for China's strategy of
civil-military integration.\9\
---------------------------------------------------------------------------
\9\ See, ``Washington unnerved by China's `military-civil fusion,'
'' Kathrin Hille, Financial Times, November 8, 2019.
---------------------------------------------------------------------------
III. The Recent Designations of ZTE, Huawei, and Other Chinese
Technology Companies
On May 16, 2019, BIS added Huawei and 68 of its non-U.S. affiliates
to the Entity List.\10\ The U.S. Government did so after determining
that there was reasonable cause to believe that Huawei had been
involved in activities contrary to the national security or foreign
policy interests of the United States. The specific activities contrary
to the national security or foreign policy interests of the United
States include those activities alleged in the Department of Justice's
public superseding indictment of Huawei, including alleged violations
of the International Emergency Economic Powers Act (IEEPA) and
conspiracy to violate IEEPA (by providing prohibited financial services
to Iran), and obstruction of justice in connection with the
investigation of those alleged sanctions violations. As a result of its
placement on the Entity List, the sale or transfer of American
commodities, software, or technology to Huawei or its affiliates on the
Entity List requires a license issued by BIS, and a license will be
presumptively denied. By publicly listing such persons, the Entity List
is an important tool to protect U.S. national security and foreign
policy interests.
---------------------------------------------------------------------------
\10\ 15 C.F.R. Pt. 744, Supp.4.
---------------------------------------------------------------------------
Under the President's recent announcement, BIS will be promptly
taking action to issue certain additional licenses, to companies that
apply, which permit transactions that pose no national security risk,
are not contrary to U.S. foreign policy interests, and are used to
maintain, service and support: (A) widely available commodity chipsets
and certain electronic integrated circuits; (B) software and tools that
are generally available to the public; or (C) operating system software
and applications and system services for mobile devices, as well as
technology and software necessary to support the operating systems.
Other license applications that pose no national security threat and
are not contrary to U.S. foreign policy interests will also be promptly
considered.
Prior to Huawei's designation, BIS also targeted ZTE. In March
2016, ZTE and several of its affiliates were added to the Entity List
for their involvement in a scheme to reexport U.S.-controlled items to
Iran. ZTE reached a settlement with BIS in March 2017, paying a total
of US$1.19 billion in fines, and was subject to a suspended denial
order. Having not complied with certain conditions of that settlement,
BIS activated the Denial Order on ZTE in April 2018. The import ban has
since been lifted as ZTE agreed to a settlement with BIS with
significant conditions, including a US$1 billion fine. BIS rightfully
took a strong stance against ZTE, imposing unprecedented compliance
measures as part of the settlement. These actions demonstrate a robust
commitment on the part of the Administration to combat technology-
related national security issues. Such efforts were necessary and long
overdue.
The effort to closely scrutinize and restrict transactions with
Chinese entities that pose potential national security risks is not
limited to the Administration. Congress, through Section 889 of the
NDAA, has effectively banned the Federal Government from purchasing
equipment from Huawei and ZTE, citing them as national security risks.
Specifically, Section 889 prohibits Federal agencies, Federal
contractors, and grant or loan recipients from procuring certain
``covered telecommunications equipment or services,'' (equipment and
services produced by Huawei and ZTE, and with respect to certain public
safety or surveillance applications, Hytera Communications Corporation,
Dahua Technology Company, and Hangzhou Hikvision Digital Technology
Company) as a ``substantial or essential component of any system, or as
critical technology as part of any system.'' Congress clearly believes
that taking a strong stance against national security threats is
warranted and necessary. We have recently seen a concerted effort from
Congress and the Administration to protect U.S. national security
against threat actors in the technology and telecommunications sectors.
Continued diligence in this area is crucial to protecting U.S. national
security moving forward.
IV. Effectiveness of ECRA-Related Control Structures in the United
States in Confronting the Challenges Raised with Respect to
China, including Persistent Diversion Challenges
I believe that the ECRA-related controls will go a long way toward
improving U.S. transparency and effectiveness in addressing the
challenges related to China and its persistent diversion tactics. We
have seen that stronger enforcement and broader application of law
under FIRRMA has had an effect. As reported by a number of sources,
including the Rhodium group, Chinese investments into the United States
have been significantly curtailed. This was important given the
statistics on Chinese government backed investment happening in our
most advanced and innovative companies. The Rhodium group had
calculated that, on average, 21 percent of Chinese venture investment
in the United States from 2000 through 2017 came from state-owned
funds, which are controlled at least in part by the Chinese government.
In 2018, that figure surged to 41 percent.\11\
---------------------------------------------------------------------------
\11\ See Reuters ``Chinese tech investors flee Silicon Valley as
Trump tightens scrutiny,'' by Heather Somerville, January 7, 2019.
---------------------------------------------------------------------------
Also, with the implementation of ECRA and designation of emerging
and foundational technologies, U.S. policymakers will be able to better
assess the vulnerability of our supply chains and where the United
States stands in terms of critical technology leadership, including
where that leadership has been eroded.
However, clearly, we need a ``whole of Government'' defensive
strategy where it concerns these national security threats. When China
utilizes government actors to hack into U.S. private companies to take
proprietary technology and give such information to Chinese companies,
the United States must address the issue broadly. Pulling from a recent
speech by the U.S. Justice Department, I note: ``since 2011, more than
90 percent of the Department's economic espionage prosecutions (i.e.,
cases alleging trade secret theft by or to benefit a foreign state)
involve China, and more than two-thirds of all Federal trade secret
theft cases during that period have had at least a geographical nexus
to China. Some of those cases demonstrate that China is using its
intelligence services and their tradecraft to target our private
sector's intellectual property.''\12\ Clearly, we must continue to
improve our ability to protect U.S. private companies from Chinese
nation-state threat actors.
---------------------------------------------------------------------------
\12\ Remarks of Deputy Assistant Attorney General Adam S. Hickey of
the National Security Division at the Fifth National Conference on
CFIUS and Team Telecom, Washington, D.C., Wednesday, April 24, 2019.
---------------------------------------------------------------------------
V. Possible Legislative or Oversight Recommendations
In closing, I applaud this Committee and this Administration for
the hard work to create new and stronger mechanisms to address national
security vulnerabilities arising from the loss of critical
technologies, military, emerging and foundational. While implementation
of ECRA and FIRRMA are underway, there is even more that could be done.
We should create additional enforcement tools to better address
cyber and intellectual property (``IP) theft. Perhaps an IP ``Entities
List,'' similar to USTR's Notorious Markets List. Further we should
consider taking additional actions in response to cyber attacks using
executive powers. With the full implementation of FIRRMA, foreign
government-controlled transactions and transactions involving critical
infrastructure should be subject to mandatory filing requirements. We
also need additional tools to address overcapacity by foreign state-
owned enterprises that are able to enter the U.S. market unimpeded or
create global market distortions to the detriment of our producers and
U.S. innovation and jobs.
Last and importantly, the key to ensuring that BIS and other export
control agencies are able to carry out their missions and the new
responsibilities under ECRA is additional funding and resources. If we
are serious about addressing the current and future loss of U.S.
emerging and foundational technology, if we want to ensure that the
United States continues to be a global leader for innovation, security
and freedom, it is critical that such funding and resources is
provided.
As Secretary Ross said: ``We can no longer accept the decline of
U.S. industries due to state-supported overcapacity, and the
strategic--often clandestine--foreign purchases and investments in our
most important technology enterprises.''\13\
---------------------------------------------------------------------------
\13\ Remarks by U.S. Commerce Secretary Wilbur L. Ross at the
Bureau of Industry and Security Annual Conference on Export Controls
and Security, July 9, 2019.
---------------------------------------------------------------------------
Thank you for the opportunity to appear before you today. I look
forward to your questions.
______
PREPARED STATEMENT OF BEN BUCHANAN, Ph.D.
Assistant Teaching Professor, School of Foreign Service
Senior Faculty Fellow, Center for Security and Emerging Technology
Georgetown University
July 18, 2019
Thank you, Chairman Crapo and Ranking Member Brown, for holding
this important hearing and for inviting me to testify.
My name is Ben Buchanan. I am an Assistant Teaching Professor at
the School of Foreign Service and a Senior Faculty Fellow at the Center
for Security and Emerging Technology, both at Georgetown University. I
am also a Global Fellow at the Woodrow Wilson International Center for
Scholars, where I teach introductory classes on Artificial Intelligence
and cybersecurity for congressional staff. My research specialty is
examining how cybersecurity and AI shape international security. I co-
authored a paper entitled ``Machine Learning for Policymakers.''\1\
---------------------------------------------------------------------------
\1\ Buchanan, Ben and Taylor Miller. ``Machine Learning for
Policymakers.'' Belfer Center for Science and International Affairs
(2017), https://www.belfercenter.org/sites/default/files/files/
publication/MachineLearningforPolicymakers.pdf.
---------------------------------------------------------------------------
As this Committee is well aware, export controls are legal tools
that are applied to technology. If either the tool or the technology is
not a good fit, export controls will fail. Given the expertise of my
two fellow witnesses on the legal nuances of the tools themselves, I
believe I will be of most value to the Committee in talking about some
of the technologies in play and what makes export controls
comparatively more or less suitable with these technologies. As a way
of opening our discussion, I will focus on one particular suite of
technologies that is particularly notable, artificial intelligence, but
I believe this discussion will also apply to other relevant
technologies.
Conceptualizing AI
Nobody has a crystal ball, but there are other ways to consider our
modern and near-future era of AI that will be useful for this
discussion. To do so, it is important to understand how AI differs from
so much of what came before it. An analogy will help.
One can imagine two ways of teaching a child to perform a task. The
first is to give very clear instructions in a language the child
understands about what the task is and how it is to be performed. The
second is to show the child, through a series of examples, how the task
works, and have the child infer the important rules and patterns
necessary to get the job done. At various points in a child's
education, they learn different tasks through each of these methods.
Traditional software development, and even some older versions of
AI, work in a way that is similar to the first method. They rely on
software developers understanding the problem to be solved in great
depth, and then imparting this expertise to the system. For example, in
a program designed to play chess, the software developers may consult
with grandmasters to understand the optimal strategies for a wide range
of situations, and then program those ideas into the code.
Modern AI systems, known as machine learning systems, use the
second method, the one involving inference. In a machine learning
system, rather than receive clear instructions about how to do the
task, software developers create an algorithm that determines how the
system should learn. They then provide that algorithm with lots of
relevant data and computational power (the processing hardware that
makes machine learning algorithms function).
There are thus three parts to this system: the algorithm, the data,
and the computational power. Together, they form an essential triad.
Each is more or less important in various versions of machine learning,
but at the same time, each in its own way is critical. To understand
why, it is worth examining the triad in a little more detail.
Data
It is in vogue to say that data is the new oil. This is because, to
use the second kind of program I described above--the machine learning
method--a lot of relevant data is often required. From this data the
machine learning system will infer important patterns and nuances, and
will determine what success and failure look like. It is thus vital
that the data provided to the machine be representative of the problem
in all its complexity and plentiful.
A large part of the reason that companies like Google, Amazon, and
Facebook are successful with the AI systems they deploy is because they
aggregate gigantic amounts of data. In essence, the large datasets
these companies assemble provide them with a competitive advantage over
others. Large companies based in other nations, such as China's Baidu,
Alibaba, and Tencent, derive similar advantages from their datasets. It
seems to me that export controls are unlikely to be of much use in
managing this competition or guarding against potential threats from
data, both because companies already have an incentive and tools to
secure and not share their assembled data and because export controls
are comparatively ill-equipped to stop the transfer of sensitive data
relative to other tools like classification (for government data), and
licensing or contractual restrictions regardless of export.
Algorithms
Algorithms are the second component of the AI triad. These software
instructions dictate how the machine learning system will learn. They
stipulate how it will interpret the data, what sort of capabilities it
will develop, and what inferences it will learn to draw that can be
applied to future tasks. There are a wide variety of algorithms, each
suited to different kinds of tasks, from classifying images to making
predictions about housing prices based on historical trends, to
generating new pictures of people who look real but do not actually
exist. The algorithmic frontier is rich, and a great deal of progress
has been made in the last 7 years.
The prevailing ethos is that, once an algorithmic advance is made,
researchers post it online and share it with others. In this sense, AI
research is remarkably open, far more so than the fierce competition of
the technology industry would normally suggest. There are exceptions to
this practice, instances in which algorithms have not been published
due to national security concerns--most notably a decision by OpenAI, a
leading research lab, not to publish a powerful algorithm that could be
used to generate realistic-fake text.
That said, the experience of several decades has shown that
government efforts to control the export of computer code are usually
futile, and I think it is fair to say that export controls are unlikely
to be useful in stopping all but the most powerful of algorithms. And
even with those most powerful algorithms, I have doubts about the
suitability of our current list-based export control systems, given the
changing pace of technology and the movement of the technological
frontier.
Computing Power
This brings us to the last part of the triad: computing power, or
what AI researchers simply call ``compute.'' It is easy to ignore, but
it remains vitally important, perhaps prohibitively so. In the last 7
years, we have witnessed a revolution in computing power applied to
machine learning. One study by OpenAI indicated that between 2012 and
2018, the computing power applied to top machine learning systems
increased by a factor of 300,000; if a cell phone battery lasted 1 day
in 2012 and increased by the same factor, that battery would now last
800 years.\2\
---------------------------------------------------------------------------
\2\ ``AI and Compute,'' OpenAI, (2018), https://openai.com/blog/ai-
and-compute/.
---------------------------------------------------------------------------
There is much to discuss about why this increase in computing power
has occurred, but the most salient factor for our purposes today is
that, unlike algorithms and data, computing power is a function of
hardware, not software. That is, computers are tangible products that
are easier to manage, including with export controls. My judgment is
that, to the degree that export controls are relevant to the problem of
managing AI and other technologies such as 5G, it will controls on this
hardware component, and likely on the hardware that manufactures
specialized computer chips for AI. This statement is both a commentary
on the limitations of export controls to the problem but also on the
more narrow areas where they might be suitable for protecting national
security.
To be clear, in order for any such controls to work--whether on AI
hardware or something else--they must be conducted in a multilateral
fashion with allies, given that a great deal of hardware engineering
expertise is outside the United States.
I thank you again for holding this hearing and the opportunity to
lay out the basics of this complicated, fast-changing field for your
consideration as you review the implications of export control for AI
and other technologies. As you know, it is vital that we both protect
national security and not squash innovation. This is an area that the
Center for Security and Emerging Technology has been studying, and we
expect to publish our analysis on it in the weeks to come. In the
meantime, I look forward to your questions.
RESPONSES TO WRITTEN QUESTIONS OF CHAIRMAN CRAPO
FROM ERIC L. HIRSCHHORN
Q.1. Expansion of Scope--Traditionally our export control
system has focused on national security and foreign policy.
Should we expand the focus of our controls to address
issues of economic competitiveness like Made in China 2025?
A.1. No one can deny the serious policy implications for the
United States of the Made in China 2025 plan. Artificial state
subsidies of particular technology sectors are market
distorting and put U.S. companies at a competitive
disadvantage. ECRA correctly recognizes, though, that in
today's interdependent world, multilateral controls are far
likely than unilateral controls to be successful.\1\ For that
reason, expanding U.S. controls to address issues of economic
competitiveness, though tempting, might be self-defeating. We
generally have been successful in recent decades in convincing
our allies to join us in controlling exports of items that
could put our collective national security at risk or engender
the spread to undesirable end users of weapons of mass
destruction. This has been so despite our allies' often-
expressed suspicions--and our truthful denials--that the U.S.
is seeking economic as well as national security advantage. To
expand our export controls to expressly address economic
competitiveness concerns could lead our allies to think twice
about supporting our efforts.
---------------------------------------------------------------------------
\1\ ECRA 1752(4)-(6) (codified at 50 U.S.C. 4811(4)-(6)).
---------------------------------------------------------------------------
Some observers contend that if the U.S. were to take the
lead in imposing controls on technology exports of technology
to China for economic reasons, our allies would follow the
example. I doubt it. Our allies have traditional views of
export controls and probably would not agree to control the
flow of technology for other than the traditional national
security and foreign policy objectives that are set out in
ECRA.
Q.2. Foundational Technologies--Are there any ``foundational''
technologies that are not, by definition, already widely
available?
A.2. Neither ECRA nor FIRRMA nor--so far, at least--the
Administration has defined the term ``foundational.'' I have
assumed that it refers to technologies that are widely
available--i.e., technologies that are export controlled, if at
all, only to the handful of countries that are designated as
supporting terrorism. I have been told that the executive
branch shares that view of what the term means. Unilaterally
controlling technologies that are widely available from other
countries would harm our domestic economy without preventing
China and other countries of concern from acquiring such
technologies. The Administration's task, then, is to ascertain
whether there are ``foundational'' technologies that (1) are
useful to China and (2) ``essential to the national security of
the United States''\2\ but (3) not widely available elsewhere,
and then--as required by ECRA\3\--seek multilateral control of
such technologies.
---------------------------------------------------------------------------
\2\ ECRA 1758(a)(1)(A) (codified at 50 U.S.C. 4817(a)(1)(A).
\3\ ECRA 1758(c) (codified at 50 U.S.C. 4817(c)).
Q.3. Evidence of Controls Driving Offshore Activity--Is there
any evidence that tight export controls drive research and
---------------------------------------------------------------------------
development, or manufacturing, offshore?
A.3. In recent decades we have seen this phenomenon in the
areas of machine tools, commercial space items, and thermal
imaging items. All were subject to very tight U.S. export
controls. The development and manufacturing competition that
grew up abroad in each of those sectors was, I believe, largely
a result of that action on our part. As long ago as the 1990s,
I had clients tell me that they were shifting their R&D
offshore because of the extreme tightness of U.S. export
controls on their types of technology. Our country narrowly
avoided this fate in two other sectors, namely encryption and
computers, because we relaxed our controls somewhat once we
realized that the horse already was out of the barn.
In a world where advanced technologies can be developed and
produced in many countries, the logic of unilateral controls
versus multilateral controls is clear. If a company cannot
legally export a technology from one country, it will likely
seek to develop and that technology in a country that does not
prohibit such exports. Moreover, investors will make their
investments in such countries rather than in the United States.
Unilateral controls thus can harm the U.S. industrial base and
enhance the industrial base of foreign competitors without
preventing the proliferation of the technology to countries of
concern. For this very basic, logical, economics-driven reason,
Congress wisely provided in ECRA that unilateral controls are
disfavored.
Q.4. Control Rulings--ECRA essentially requires an interagency
review of decisions to add or remove items from the control
lists and to approve or deny individual license applications.
Is Commerce the best department to lead the dual-use export
control system?
What would be the harm if we transferred the export control
system to the Defense Department for it to decide what should
or should not be exported?
A.4. Commerce has administered controls on dual-use items since
the late 1940s, with lower-level military items being added in
recent years as part of the Export Control reform initiative.
For about 25 years, the Defense, State, and Energy departments
have been empowered to review any Commerce license
application--and in fact do review almost all such
applications. Moreover, any changes to the regulations,
including additions to and subtractions from the control lists,
essentially require consensus of the four agencies before they
may be implemented. Contrary to the false statements of some
that Commerce somehow routinely ``overrules'' Defense on
national security judgments and State on foreign policy
judgments, the current system does an excellent job of
accounting for the expertise and equities of different parts of
the Government. To quote a sage Washington observer from the
past--``If it ain't broke, don't fix it.''
Moreover, such controls long have taken into account not
only potential military applications, as to which the Defense
Department has special expertise, but also such salient issues
as foreign availability, foreign policy, and the like, where
Commerce and State bring their expertise to the table. ECRA has
it right: Commerce controls should take into account their
``impact on the economy of the United States'' and should be
imposed ``only to the extent necessary'' to achieve U.S.
national security and foreign policy aims.\4\ The
considerations set out in ECRA include the strength of the U.S.
defense industrial base--a role shared by Commerce and Defense
under the Defense Production Act of 1950--the maintenance of
U.S. leadership in the ``science, technology, engineering, and
manufacturing sectors,'' and foreign availability.\5\ Defense
already--and appropriately--has a full voice when it comes to
potential military application of technology but the voices of
Commerce, State, and, where nuclear issues are concerned,
Energy also are essential to the proper functioning of the
system. Commerce has technical and policy expertise in all
these areas but has demonstrated particular skill in
administering and enforcing a reliable, predictable regulatory
regime that pursues all these objectives.
---------------------------------------------------------------------------
\4\ ECRA 1752(1) (codified at 50 U.S.C. 4811(1)).
\5\ ECRA 1752(2)(C), (3), (6) (codified at 50 U.S.C. 4811(2)(C),
(3), (6)).
Q.5. Huawei Delisting--The addition of the Huawei to the
Commerce Department's Entity List is one of the most public and
significant export control topics in today's headlines.
Without commenting on Huawei, could you tell us what the
Entity List is, its purpose, and whether or not it has been
historically effective and how is it different from a civil or
a criminal penalty, a denial order, a Treasury Department
sanction, or other actions the U.S. Government can take against
a foreign company?
A.5. In a strictly legal sense, the Entity List is fairly low
on the totem pole of actions the United States can take against
a company. It imposes no criminal, civil, or administrative
penalty against a named party but merely requires that all
items ``subject to the Export Administration Regulations''
require a license to be exported to that party. Of course, this
means that many items that don't need a license to go to anyone
else will have to wait while a license application is
submitted, considered, and possibly denied. Importantly, items
that are not ``subject to the EAR'' are not caught by a foreign
importer's appearance on the Entity List. That means that
unlike, say, sanctions administered by the Treasury
Department's Office of Foreign Assets Control, the Entity List
doesn't reach foreign-made items located outside the U.S., even
if they're sold from such locations to Entity List companies by
U.S. companies.
The most common reason for placing a company on the Entity
List is to encourage it to clean up its act in terms of
respecting U.S. export controls. Once it has demonstrated its
compliance, and provided relevant information to the Office of
Export Enforcement at Commerce, the company can seek removal
from the List. Absent the possibility of removal in exchange
for cooperation, there is little incentive for a listed company
to cooperate with the U.S. Government.
As difficult as the formal effect of being on the Entity
List may make a company's life, the secondary and unofficial
effect can be worse. This is because financial institutions and
large companies throughout the developed world use software to
screen for ``bad'' customers. That software includes everyone
who's been fined, indicted, listed as a denied party, debarred,
or placed on the Entity List. So although the Entity List
technically isn't a penalty, companies listed there get added
to that software, too. That in turn means that many such
financial institutions and companies simply will refuse to do
business with them.
The only less draconian action than the Entity List is the
Unverified List. Placement there usually means that U.S.
officials have not been permitted to make post-shipment visits
to the entity in question to check whether the U.S.-origin
items supposedly sent there actually are there. Often the
reason is that the host government, rather than the consignee
company, is the problem. Exporters may ship to parties on the
Unverified List without obtaining additional licenses but they
are on notice that such parties' bona fides are uncertain and
that accordingly they should take care to satisfy themselves
that the orders are legitimate.
Q.6. The United States has a special treatment arrangement with
Hong Kong with regards to export controls. While it is in the
United States interests to have a strong economic relationship
with Hong Kong, there is a lot of concern about growing Chinese
encroachment on Hong Kong's autonomy and the potential
implications for safeguarding technology.
Is our current export control policy equipped to deal with
risk of diversion from Hong Kong to China?
What are some ways in which China is using or could use
Hong Kong as a vector for acquisition of technology that we do
not export to the Mainland?
What are your specific recommendations for strengthening
our export control regime in relation to these challenges?
A.6. The concerns expressed in these questions are legitimate
and not new, though recent efforts of the Chinese government to
narrow Hong Kong's autonomous status of course bring them to
the fore. In my experience, cooperation between the United
States and Hong Kong customs and export control authorities has
been good but Hong Kong is a very busy place with close
political and commercial ties to China. The Obama
administration strengthened protections against unauthorized
diversions to China via Hong Kong by requiring ``persons
intending to export or reexport to Hong Kong any item subject
to the Export Administration Regulations (EAR) and controlled
on the Commerce Control List (CCL) for national security (NS),
missile technology (MT), nuclear nonproliferation (NP column
1), or chemical and biological weapons (CB) reasons to obtain,
prior to such export or reexport, a copy of a Hong Kong import
license or a written statement from the Hong Kong government
that such a license is not required.''
That rule ``also requires persons intending to reexport
from Hong Kong any item subject to the EAR and controlled for
NS, MT, NP column 1, or CB reasons to obtain a Hong Kong export
license or a statement from the Hong Kong government that such
a license is not required.''\6\
---------------------------------------------------------------------------
\6\ 82 FR 6216 (Jan. 19, 2017).
---------------------------------------------------------------------------
I don't know enough about our experience under the 2017
rule to have a view on whether it is working well. I would
encourage continued vigilance over reexports of U.S.-origin
items from Hong Kong, including compliance with that
regulation. Such vigilance is possible, though only if BIS has
the resources to carry it out.
Congress should appropriate additional funds to BIS so that it
can do this work, as well as its other work, thoroughly and
effectively.
------
RESPONSE TO WRITTEN QUESTION OF SENATOR BROWN
FROM ERIC L. HIRSCHHORN
Q.1. The press have reported widely on China's surveillance
state and their gross human rights violations of the Uyghur
people. Whether the technology is organically developed or
stolen IP from American companies, we should all be concerned
with how technology can be perverted to violate civil liberties
and basic human rights.
Mr. Hirschhorn, walk me through how the interagency takes
issues like human rights violations into consideration when
discussing emerging technologies. What are the mechanisms to
mitigate the unintended consequences of bad actors or countries
misusing these technologies, and can that process be improved?
A.1. ECRA sensibly requires that the consideration of export
control policies, as well as individual licensing decisions,
take human rights into account.\1\ This applies not only to
emerging technologies but all export control policies and
license reviews. Applications for Commerce Department export
licenses are shared with the Department of State, whose Bureau
of Democracy, Human Rights, and Labor (DRL) is among the
bureaus consulted internally at that department. The State
Department has an equal vote with Commerce, Defense, and Energy
at all levels of the process for reviewing applications for
Commerce export licenses. Moreover, human rights are not the
exclusive province of State. Commerce, Defense, and Energy can
and do raise human rights concerns about BIS license
applications.
---------------------------------------------------------------------------
\1\ ECRA 1752(2)(D) (codified at 50 U.S.C. 4811(2)(D)).
---------------------------------------------------------------------------
Part of the difficulty in ensuring protection of human
rights is that sometimes the technologies employed to violate
human rights are controlled to only a handful of countries
because they are fairly basic or general purpose, they're
widely available from sources other than the United States, or
both. Where that is the case, refusal to allow the export of
U.S. products may make an appropriate statement from a foreign
policy standpoint but have little or no practical effect on the
target government.
------
RESPONSES TO WRITTEN QUESTIONS OF SENATOR WARREN
FROM ERIC L. HIRSCHHORN
Q.1. At least one U.S. company has been found to have provided
the Chinese government with a tool enabling it to monitor
Uyghur and Central Asian minorities, as part of what one Uyghur
activist described in April 9, 2019, testimony to the Senate
Foreign Relations Subcommittee on East Asia, the Pacific, and
International Cybersecurity Policy as ``an Orwellian mass
surveillance state'' where ``more than one million Uyghurs are
arbitrarily detained outside the legal system in concentration
camps.'' A bipartisan group of Senators introduced the Uyghur
Human Rights Policy Act, of which I am a cosponsor, which
states in part, that:
the Secretary of Commerce should review and consider
prohibiting the sale or provision of any United States-
made goods or services to any state agent in Xinjiang,
and adding the Xinjiang branch of the Chinese Communist
Party, the Xinjiang Public Security Bureau, and the
Xinjiang Office of the United Front Work Department, or
any entity acting on their behalf to facilitate the
mass internment or forced labor of Turkic Muslims, to
the ``Entity List'' administered by the Department of
Commerce.
Please explain your view. Do you agree?
A.1. The United States has imposed end-user-specific
restrictions such as this on many occasions. Indeed, ECRA's
provision that U.S. export controls should ``carry out the
foreign policy of the United States, including the protection
of human rights and the promotion of democracy''\1\ provides
ample legislative authority for such an action.
---------------------------------------------------------------------------
\1\ ECRA 1752(2)(D) (codified at 50 U.S.C. 4811(2)(D)).
Q.2. Can you conceive of any circumstances under which it would
be appropriate for the United States to weaken our export
control laws and regulations, or the enforcement of those laws
and regulations, vis-a-vis China or any other foreign
competitor in order to extract concessions or other commitments
from that foreign competitor on matters related to trade or
---------------------------------------------------------------------------
human rights? Please explain your view.
A.2. I'm not prepared to say that such circumstances never
could arise but it would have to be an extraordinary case. The
recent suggestions by the President that extensive export
control violations by a Chinese telecommunications company
should be traded for sales of beef, chicken, soybeans, and the
like do not meet that criterion, and seriously undermine our
military and our law enforcement.
------
RESPONSES TO WRITTEN QUESTIONS OF
SENATOR CORTEZ MASTO FROM ERIC L. HIRSCHHORN
Q.1. You said in your testimony you believed for export
controls to be effective, they should be adopted both by our
allies and unilaterally imposed by us.
What are the challenges you foresee in convincing allies to
align with us on this issue and how can we overcome them?
A.1. When it comes to issues of military security and
preventing the spread of weapons of mass destruction, we have
had great success over the past 70 years in convincing our
allies to go along with controls that are proposed by the
United States. In the area of foreign policy, most notably with
regard to Cuba, we have had relatively support from our allies.
The greatest obstacles to convincing allies to cooperate are
closely related to one another. First is their oft-expressed
view--one that is incorrect--that we somehow use such controls
to further our own economic and commercial interests. Second is
the fear that U.S. exporters don't share equally in the harm
that necessarily flows to domestic parties from any country's
export controls and economic embargoes.
The best way to convince our allies to cooperate is to
continue--as ECRA prescribes \1\--to limit our export controls
to those driven by military, intelligence, and foreign policy
considerations.
---------------------------------------------------------------------------
\1\ ECRA 1752(1), (11) (codified at 50 U.S.C. 4811(1), (11)).
Q.2. You said in your testimony it is hard to draw up export
controls for certain emerging technologies because you run the
risk of being too generic or broad, which would make the
controls difficult to enforce.
What is the best way to address this problem to ensure
regulations are as specific to technology and application as
possible?
A.2. The best way to ensure the requisite specificity is pretty
much to continue doing what the executive branch has been doing
for decades, namely keeping a close eye on emerging
technologies but not imposing controls until it's clear what
uses the technologies are being put to and which of those uses
have demonstrable potential for military or intelligence use by
adversaries.
ECRA sensibly provides that, in deciding whether to
identify a technology that's ``essential to the national
security''\2\ as being ``emerging'' or ``foundational,'' and
impose unilateral controls on its export, the Administration
take into account--
---------------------------------------------------------------------------
\2\ ECRA 175(a)(1)(A) (codified at 50 U.S.C. 4817(a)(1)(A)).
the development of the technologies in foreign
---------------------------------------------------------------------------
countries;
the effect that such export controls may have on
the development of such technologies in the United
States; and
the effectiveness of export controls imposed
pursuant to this section on limiting the proliferation
of emerging or foundational technologies to foreign
countries.\3\
---------------------------------------------------------------------------
\3\ ECRA 1758(a)(2)(B) (codified at 50 U.S.C. 4817(a)(2)(B)).
Moreover--and this was a valuable lesson of the Export
Control Reform initiative--an important way to limit the
possibility that the Government mistakenly will under or over-
control emerging technologies (or any technologies, for that
matter) is to seek public comment on proposed controls before
actually implementing them. ECRA wisely requires this.\4\
---------------------------------------------------------------------------
\4\ ECRA 1758(a)(2)(C) (codified at 50 U.S.C. 4817(a)(2)(C)).
Q.3. To what extent does the Commerce Department collaborate
with developers to understand what technological aspects should
---------------------------------------------------------------------------
be controlled under ECRA?
A.3. I have been out of the Government for nearly three years
and don't know what contacts those administering our export
controls currently have with the private sector. During my
tenure, we found the input from BIS' technical advisory
committees, as well as from general requests for public
comment, to be of great value in ensuring that our controls
would be appropriate. BIS would be wise to continue and expand
this policy.
Q.4. I am also concerned that if we too tightly regulate the
export of some technologies, we will drive our innovation and
production offshore, as you suggested in your testimony.
In your experience, what is the best way to ensure that we
are maintaining our global leadership in technology, while also
not contributing to adversarial countries' best efforts to
surpass us?
A.4. In recent decades we have seen this phenomenon in the
areas of machine tools, commercial space items, and thermal
imaging items. All were subject to very tight U.S. export
controls. The development and manufacturing competition that
grew up abroad in each of those sectors was, I believe, largely
a result of that action on our part. As long ago as the 1990s,
I had clients tell me that they were shifting their R&D
offshore because of the extreme tightness of U.S. export
controls on their types of technology. Our country narrowly
avoided this fate in two other sectors, namely encryption and
computers, because we relaxed our controls somewhat once we
realized that the horse already was out of the barn.
The message here is that these are judgment calls and that
more is not always better. Although potentially dangerous
technologies of course should be controlled, overly tight
controls--especially if they are unilateral--can be as damaging
as overly loose controls. The result in the thermal imaging
area, for example, has been that we have significant foreign
competition and--even more important from a security
standpoint--have no window into, or influence regarding--where
those foreign products end up. Had we been a shade more
reasonable in deciding how tightly we should control that
technology, other countries might have had less of an incentive
to create their own thermal imaging manufacturing capability.
Sixteen years ago I presented a paper on this point,
``Export Issues for Military Sensors: The Fork in the Road,''
at a Military Sensors Symposium sponsored by the U.S. Army's
Night Vision and Electronic Sensors Directorate.\5\ The
symposium was attended by many of our Government's leaders in
the thermal imaging field. My conclusion, which regrettably was
prophetic, was as follows:
---------------------------------------------------------------------------
\5\ Eric L. Hirschhorn, ``Export Issues for Military Sensors: The
Fork in the Road,'' presented at Military Sensors Symposium sponsored
by Night Vision and Electronic Sensors Directorate, Department of the
Army (Oct. 2003) (proceedings classified Secret; this paper
Unclassified and approved for unlimited public release).
The United States long has been the unchallenged leader
in sensor technology but a move offshore of
production--and leadership--is imminent. Commercial
demand for American sensor products is burgeoning but
the bottleneck caused by export restrictions has
created a supply shortage abroad. Foreign technology
and products are rushing to fill this foreign demand.
They may not yet have matched United States
technological standards but they will get there
quickly, especially if we continue, in effect, to cede
foreign markets to foreign suppliers. Reasons of
national security . . . make it crucial that this not
occur. There is no doubt that it will occur, however,
if the current overcontrol of sensors is not adjusted
---------------------------------------------------------------------------
to comport with reality.
Subsequently, we saw the development of robust thermal
imaging industries in France, Israel, and China, among others.
We now have little or no visibility of, and little or no
influence about, where the products of those countries--
particularly those that are not members of the Wassenaar
Arrangement--end up. This might have been different had we been
more nuanced about how we controlled this technology.
------
RESPONSE TO WRITTEN QUESTION OF SENATOR SINEMA
FROM ERIC L. HIRSCHHORN
Q.1. There appears to be consensus that a multilateral approach
to export controls is most effective in mitigating technologies
that threaten U.S. industry and national security. It also
appears there is consensus that multilateral efforts will work
best in restricting divisive Chinese technology and
infrastructure. Given the importance of a multilateral approach
and the serious national security threats China poses, are you
at all concerned that the Administration's policies and
rhetoric on trade could undermine the necessary goodwill to
work collaboratively with our trading partners to hold China
accountable?
A.1. I am of two minds about this issue. Sovereign nations
generally are more able to look out for and pursue their own
best interests than, say, individuals within our country. That
means that if they see an advantage in cooperating with the
United States on export controls and economic sanctions, they
will do so despite the rude treatment that they may have been
receiving from the current administration. That said, not every
aspect of international relations is bloodless and devoid of
personal emotion. There quite possibly will be cases where our
poor treatment of an ally will discourage it from cooperating
with us.
------
RESPONSES TO WRITTEN QUESTIONS OF CHAIRMAN CRAPO
FROM NOVA J. DALY
Q.1. Expansion of Scope--Traditionally our export control
system has focused on national security and foreign policy.
Should we expand the focus of our controls to address
issues of economic competitiveness like Made in China 2025?
A.1. Chairman Crapo, thank you again for the opportunity to
testify before the Senate Committee on Banking, Housing, and
Urban Affairs. It was an honor. The following responses to
questions are based on my own views and do not reflect those of
my firm, Wiley Rein LLP, nor any client.
Under the Foreign Investment Risk Review Modernization Act
of 2018 (FIRRMA), which governs the Committee on Foreign
Investment in the United States (CFIUS), Congress included as a
national security consideration acquisitions of critical
technology as they affect U.S. leadership in areas related to
national security. Such considerations were also included in
the precursor legislation to FIRRMA, the ``Foreign Investment
and National Security Act of 2007.'' Further, one of the
factors included in the first CFIUS bill is ``the potential
effects of the proposed or pending transaction on United States
international technological leadership in areas affecting
United States national security.'' Thus, there is clear
precedent for treating critical technology protection and
leadership as a core national security consideration. Adding to
this precedent, in December 2017, the Trump administration
published its ``National
Security Strategy of the United States of America.'' That
document states that ``economic security is national
security.''
As may you know, and as reflected in my testimony, China's
``Made in China 2025'' initiative emphasizes China's priorities
for high-tech industries. The Chinese strategy lists multiple
technologies where it seeks global dominance. These
technologies include those that are or may be subject to U.S.
export controls and therefore raise national security and
foreign policy considerations. Placing appropriate controls on
emerging and foundational technologies should be undertaken in
a targeted way to address those Made in China 2025 initiatives
that raise national security and foreign policy considerations.
Such undertakings are already underway. The Department of
Commerce's 14 proposed emerging technology categories mirror in
some respects the industries China identified as part of its
Made in China 2025 initiative. Congress itself could also
consider the national and economic security effects of China's
``military-civil fusion,'' which has been a strategic
initiative for some time. The goal of this initiative is the
assimilation of China's technology industry into its defense
industry in order to propel the advancement of dual-use
technologies.
I applaud the U.S.-China Economic and Security Review
Commission for holding a hearing earlier this year on the topic
of ``Technology, Trade, and Military-Civil Fusion,'' thus
raising this issue with Congress. I believe that it is
imperative that we expand the focus of our export controls to
address Made in China 2025 issues that raise national security
and foreign policy considerations.
Q.2. Foundational Technologies--Are there any ``foundational''
technologies that are not, by definition, already widely
available?
A.2. The Bureau of Industry and Security (BIS) announced that
an advance notice of proposed rulemaking (ANPRM) would be
issued seeking public comment on criteria for identifying
``foundational'' technologies. However, the ANPRM has not yet
been issued. Without this list, it is difficult to answer this
question comprehensively. That said, when the list is issued,
it will be important to address ``point of the spear''
technologies within each category of foundational technologies.
For example, semiconductors are likely to be considered a
foundational technology. Within semiconductors, the
Administration could consider targeting technologies involving
gallium nitride (GaN) for any additional export control
authorities.
Q.3. Evidence of Controls Driving Offshore Activity--Is there
any evidence that tight export controls drive research and
development, or manufacturing, offshore?
A.3. I am aware of anecdotal evidence that tight export
controls have caused lost sales that led to increased sales and
manufacturing by foreign competitors. This is a historic
consequence of having export controls for every country that
has them. However, to address this issue, we must always seek
to ensure that we apply export controls appropriately, while
taking account of commercial considerations. The United States
should continue to work with its multilateral partners to
ensure a broader consensus and consistent application of export
controls. However, unilateral export controls may be required
where we know an adversary is or could utilize such technology
for military or nefarious purposes.
Q.4. Control Rulings--ECRA essentially requires an interagency
review of decisions to add or remove items from the control
lists and to approve or deny individual license applications.
Is Commerce the best department to lead the dual-use export
control system?
A.4. The Department of Commerce is the right agency to lead the
dual-use export control system because of its breadth of
experience and history of authority in this area. Commerce has
authority and expertise beyond BIS, including under the
International Trade Administration (ITA) and various bureaus
that focus on different sectors of the U.S. economy. BIS is
therefore able to collaborate with other bureaus, including
within ITA, to build expertise on global economic matters and
commercial considerations where export control policies
requires broader considerations.
Q.5. What would be the harm if we transferred the export
control system to the Defense Department for it to decide what
should or should not be exported?
A.5. While the U.S. Department of Defense (DOD) has significant
resources, it currently does not have the depth of historical
knowledge or sector-specific resources that the Department of
Commerce has in the application of all dual-use technologies.
While DOD has expertise and experience on certain core military
technologies, Commerce has a breadth of economic resources and
industry knowledge important to the assessment and application
of export controls to dual-use technologies.
Q.6. Unfair Trade Practices--China engages in unfair trade
practices and artificially subsidizes its companies in order to
over-develop and over-produce in key sectors, such as
semiconductors, in order to dominate the world marketplace. In
order to protect the economic viability of U.S. companies, some
propose we use export control rules to cut off the flow of
basic commercial technology that the Chinese need to compete
against U.S. companies--even if the technology has nothing to
do with national security or foreign policy objectives.
If we take this approach, what's to prevent a non-U.S.
company, such as in Europe or Japan, from simply filling behind
or ``designing out'' the U.S. company and profiting off U.S.-
only prohibitions being applicable only to U.S. companies?
A.6. Under the Export Control Reform Act of 2018 (ECRA),
Congress stated that it is the policy of the United States
``[t]o use export controls only after full consideration of the
impact on the economy of the United States and only to the
extent necessary'' to restrict exports (1) that would make a
significant contribution to the military potential of any other
country that would prove detrimental to U.S. national security
or (2) if necessary to significantly further U.S. foreign
policy interests or to fulfill international obligations. (Sec.
1752).
Engagement with our multilateral partners is key to
preventing such outcomes. Export controls, where harmonized,
are much more effective. Toward that end, I understand that BIS
is actively engaging with like-minded partners to discuss
coordinating policies on emerging technologies. The U.S. should
fully pursue engagement with multilateral partners to come up
with combined agreement on what should or should not be
exported and also to improve information-sharing between
allies.
Q.7. Unilateral Controls--Under what circumstances should the
United States impose a unilateral control--that is, a control
that only the United States imposes--on the export to China of
a U.S.-origin commodity or technology?
A.7. Where there are clear economic security and national
security concerns, especially where there is a choke point in
the application of controls, or where our allies are unwilling
to move to create multilateral controls, the United States has
a responsibility to do so for its own national security
interests. The U.S. should responsibly implement unilateral
controls where clear economic security and national security
interests arise.
Q.8. If there is such a case, what would prevent either a U.S.-
subsidiary or non-U.S. company from simply selling such items
outside the United States, thus enhancing the foreign company
and harming the U.S. company?
A.8. Addressing export controls, especially those concerning
emerging and foundational technologies, requires engagement
with U.S. companies large and small, Congressional resources
and oversight, and effective compliance regimes at the company
level. Thus, U.S. companies, including their subsidiaries, must
be aware that certain potential ``business'' partners or
activity actually represent the interests of adversarial
foreign governments who will use such technology and know-how
to the economic security and/or military detriment of the
United States.
There has always been a problem arising from the
application of export controls when foreign countries and
companies sell items controlled by U.S. export laws and
regulations to the detriment of U.S. companies. That is why it
is critical to build broad coalitions with our allies on export
controls and, where we apply them unilaterally, to do so
appropriately and with a clear understanding of economic
considerations.
Q.9. The United States has a special treatment arrangement with
Hong Kong with regards to export controls. While it is in the
United States interests to have a strong economic relationship
with Hong Kong, there is a lot of concern about growing Chinese
encroachment on Hong Kong's autonomy and the potential
implications for safeguarding technology.
Is our current export control policy equipped to deal with
risk of diversification from Hong Kong to China?
A.9. No export control system is foolproof, but BIS has sought
to address the risk of diversion from Hong Kong to mainland
China, including with rules that went into effect in April
2017. In particular, these rules impose new requirements and
supporting documentation for exports of specific controlled
items to Hong Kong and build teams that do end-use checks. BIS
has also issued guidance on due diligence factors for exporters
to consider in order to
prevent unauthorized transshipment or reexport of controlled
items through Hong Kong to China.\1\
---------------------------------------------------------------------------
\1\ U.S. Department of Commerce Bureau of Industry and Security,
``Guidance on Due Diligence to Prevent Unauthorized Transshipment/
Reexport of Controlled Items through Hong Kong to China,'' 2016,
https://www.bis.doc.gov/index.php/policy-guidance/hong-kong-due-
diligence-guidance.
Q.10. What are some ways in which China is using or could use
Hong Kong as a vector for acquisition of technology that we do
---------------------------------------------------------------------------
not export to the Mainland?
A.10. While I do not have direct expertise on this matter, it
is not difficult to imagine that Chinese entities currently do
and would use companies in Hong Kong to gain access to U.S.
controlled technologies. They would likely do so through Hong
Kong companies aligned with or having close commercial ties to
mainland companies and/or customers.
Q.11. What are your specific recommendations for strengthening
our export control regime in relation to these challenges?
A.11. I would suggest additional funding for BIS and additional
oversight by Congress. For BIS, funding could be provided for
more end-use checks to be conducted per annum based on past
performance, and BIS could use help on targeting end-use checks
in Hong Kong through upfront research on no license required
shipments prior to post shipment verification requests,
enhanced and continued intelligence sharing within BIS, and the
utilization of intelligence information to help identify
appropriate end-use checks.\2\ The continued attention of
Congress to such matters is paramount to successful U.S.
efforts to counter reexport of export-controlled goods from
Hong Kong to China.
---------------------------------------------------------------------------
\2\ Id. at 27.
---------------------------------------------------------------------------
------
RESPONSES TO WRITTEN QUESTIONS OF SENATOR BROWN
FROM NOVA J. DALY
Q.1. Mr. Daly, I want to hear your views on whether we should
be developing additional tools to combat the Chinese
government's efforts to dominate specific sectors.
Just as with developing emerging technologies, the Chinese
government is strategic about the amount and targets of Chinese
investment abroad. We've seen the results of that strategy in
the U.S. rail car manufacturing sector, and there are recent
reports about Chinese state-owned-enterprises investing in our
energy sector.
Senator Grassley and I have a bill, the Foreign Investment
Review Act, which would authorize the Secretary of Commerce to
review foreign investments--particularly those made by Chinese
state-owned enterprises--to make sure they're in our long-term
economic interests.
Do you agree that there is value in establishing an
investment screen in place to combat China's threats to our
economic security?
A.1. I greatly appreciate the purpose of the Foreign Investment
Review Act (FIRA) and Senator Brown and Senator Grassley for
continuing to raise the issue of addressing Chinese efforts to
make certain detrimental targeted investments in the United
States and through state-owned enterprises. Understanding the
effects of certain transactions on U.S. economic and technology
leadership, especially those emanating from China or those with
a nexus to China, is very important, and such transactions
should be appropriately reviewed.
As you may know, currently, CFIUS legislation, as modified
by FIRRMA, is undergoing a regulatory process of
implementation. FIRRMA establishes processes that require
mandatory filing of certain technology transactions and certain
transactions involving foreign government ownership. The CFIUS
``Pilot Program'' which addresses technology acquisitions, is
well underway, and it will be useful to see how that program
has helped to address threats to our critical technology
leadership, an important element of our national and economic
security. The mandatory filing requirement for transactions
involving government-controlled entities could have been based
on a control standard, as found in FIRA, rather than on
ownership levels. Doing so would have likely increased filings
made to CFIUS. It may be useful to consider legislation that
would more narrowly apply the control standard to certain
investments made by state-owned or controlled entities
emanating from certain foreign investors from certain foreign
countries. Nonetheless, the FIRA bill is helpful legislation
and should have further congressional consideration.
Q.2. The press have reported widely on China's surveillance
state and their gross human rights violations of the Uyghur
people. Whether the technology is organically developed or
stolen IP from American companies, we should all be concerned
with how technology can be perverted to violate civil liberties
and basic human rights.
Please discuss how the interagency process takes issues
like human rights violations into consideration when discussing
emerging technologies. What are the mechanisms to mitigate the
unintended consequences of bad actors or countries misusing
these technologies, and can that process be improved?
A.2. The protection of human rights is specifically mentioned
in the Statement of Policy in the Export Controls Act of
2018.\3\ Export controls are currently in effect for crime
control categories.\4\ The Department of State submits country
Reports on Human Rights Practices to Congress, which is used by
State and the Department of Commerce to deny licenses for
export of crime control items to any country whose government
engages in a consistent pattern of violations of
internationally recognized human rights in accordance with the
Foreign Assistance Act. Further, State is developing guidance
for exporters of items with intended and unintended
surveillance capabilities. The guidance seeks to provide
insight to exporters on considerations to weigh prior to
exporting these items. It also offers businesses greater
understanding of the human rights concerns the U.S. Government
may have with the export. Lastly, recently, using the
interagency process, BIS added the Xinjiang Uyghur Autonomous
Region People's Government Public Security Bureau, 18 of its
subordinate municipal and county public security bureaus, and
another subordinate institute to the Entity List.
---------------------------------------------------------------------------
\3\ National Defense Authorization Act for Fiscal Year 2018, Pub.
L. No. 115-232 (H.R. 5515) sec. 1752, 115th Cong, (2018).
\4\ Department of Commerce Bureau of Industry and Security, ``2018
Report on Foreign Policy-Based Export Controls,'' https://
www.bis.doc.gov/index.php/documents/pdfs/2186-bis-foregin-policy-
report-2018/file.
---------------------------------------------------------------------------
------
RESPONSES TO WRITTEN QUESTIONS OF SENATOR WARREN
FROM NOVA J. DALY
Q.1. At least one U.S. company has been found to have provided
the Chinese government with a tool enabling it to monitor
Uyghur and central Asian minorities, as part of what one Uyghur
activist described in April 9, 2019, testimony to the Senate
Foreign Relations Subcommittee on East Asia, the Pacific, and
International Cybersecurity Policy as ``an Orwellian mass
surveillance state'' where ``more than one million Uyghurs are
arbitrarily detained outside the legal system in concentration
camps.'' A bipartisan group of Senators introduced the Uyghur
Human Rights Policy Act, of which I am a cosponsor, which
states in part, that:
the Secretary of commerce should review and consider
prohibiting the sale or provision of any United States-
made goods or services to any state agent in Xinjiang,
and adding the Xinjiang branch of the Chinese Communist
party, the Xinjiang Public Security Bureau, and the
Xinjiang Office of the United Front Work Department, or
any entity acting on their behalf to facilitate the
mass internment or forced labor of Turkic Muslims, to
the ``Entity List'' administered by the Department of
Commerce.
Do you agree? Please explain your view.
A.1. The continued promotion of human rights and religious
freedom is an important matter in the consideration of export
control policy. I applaud the continued efforts of the Members
of the Committee on Banking, Housing, and Urban Affairs to
promote religious freedom and human rights in China and around
the world and the purpose of the Uyghur Human Rights Policy
Act.
Toward that end, I would note that in October of this year
BIS added eight Chinese tech companies in the video
surveillance, facial/voice recognition, cybersecurity, and
artificial intelligence/machine learning sectors to its Entity
List, effectively banning these companies from receiving U.S.
products and technology without a license. Similar to the
restrictions imposed on Huawei and a number of its affiliates
starting in May of this year, a license now will be required to
export all items subject to the Export Administration
Regulations (EAR)--including commercial U.S. hardware,
software, and technology--to the companies identified.
Additionally, the Xinjiang Uyghur Autonomous Region (XUAR)
People's Government Public Security Bureau, 18 of its
subordinate municipal and county public security bureaus, and
another subordinate institute were added to the Entity List for
what the BIS notice describes as ``human rights violations and
abuses in the implementation of China's campaign of repression,
mass arbitrary detention, and high-technology surveillance''
against minority groups in the XUAR. The same restrictions
described above apply to these government entities.
Q.2. Can you conceive of any circumstances under which it would
be appropriate for the United States to weaken our export
control laws and regulations, or the enforcement of those laws
and regulations, vis-a-vis China or any foreign competitor in
order to extract concessions or other commitments from that
foreign competitor on matters related to trade or human rights?
Please explain your view.
A.2. The strong enforcement of our trade and export control
laws is an imperative. This Administration has continued to
demonstrate vigilance in the application of these laws to a
high degree. For trade matters, any relaxing of the application
of such laws and regulations would be done where the United
States reaches agreements that bring greater benefits to our
national and economic security.
Q.3. In your written testimony, you observed that ``the key to
ensuring that [Commerce Department's Bureau of Industry and
Security] and other export control agencies are able to carry
out their missions and the new responsibilities under ECRA is
additional funding and resources. If we are serious about
addressing the current and future loss of U.S. emerging and
foundational technology, if we want to ensure that the United
States continues to be a global leader for innovation,
security, and freedom, it is critical that such funding and
resources is provided.'' What additional funding and resources
would you prescribe?
A.3. For BIS, funding could be provided for more end-use checks
to be conducted per annum based on past performance, and BIS
could use help on targeting of end-use checks in Hong Kong
through upfront research on no license required shipments prior
to post-shipment verification requests, enhanced and continued
intelligence sharing within BIS, and the utilization of
intelligence information to help identify appropriate end-use
checks, among other considerations.
Q.4. In your written testimony, you observed that ``it is
important that we have a system where R&D works here in the
United States, but also that key technology does not leave our
shores, especially where there is a national security/military
nexus.'' The discussion around export controls focuses
significantly on China and other external challenges, but I
want to further explore the domestic policies that we can pair
with our export control laws in order to drive innovation here
at home. Do you believe that significant increases in federally
funded basic and applied research could be complementary to our
efforts to address controls on emerging and foundational
technologies? Please explain your view.
A.4. On August 30, 2019, the Trump administration issued its
``Fiscal Year 2021 Administration Research and Development
Budget Priorities.'' In its memo, the Administration states
that ``While the private sector funds and performs the majority
of U.S. R&D, the Federal Government has an important role in
funding R&D in areas that industry does not have a strong
incentive to invest in and in areas of critical importance to
national and economic security.'' I fully agree with that.
Further, the Administration has prioritized Federal R&D funding
into ``Industries of the Future,'' such as artificial
intelligence and quantum information science. These industries
include emerging and foundational technology. Thus,
prioritizing Federal R&D funding toward these sectors will help
to build our capabilities and innovations in the sectors that
include emerging and foundational technology. Addressing
controls of these technologies is a separate funding need,
meant for enforcement of our export controls.
------
RESPONSES TO WRITTEN QUESTIONS OF
SENATOR CORTEZ MASTO FROM NOVA J. DALY
Q.1. You suggested in your testimony that the United States
could implement an ``Intellectual Property Entities List,''
similar to the United States Trade Representative's Notorious
Markets List.
Could you elaborate on that idea, and how such an entities
list would differ from the Notorious Markets List? What kind of
enforcement tools would you want to see created?
A.1. I am still working through the mechanics of such a regime
and would be happy to discuss the matter further with the
Senator and/or staff. The broad consideration is to establish
an interagency committee that would have the power to apply
remedies to repeated IP offender entities to include bans or
limitations on certain investments, procurements, U.S.
companies doing business with such entities, as well as
possible financial sanctions, etc.
Q.2. This spring, the Trump administration placed Zhongxing
Telecommunications Equipment Corporation (ZTE) and then Huawei
on the Commerce Department's ``Entity List'' for export
controls for involvement in activities ``determined to be
contrary to the national security or foreign policy interests
of the United States.'' In your testimony, you called these
actions ``necessary and long overdue.''
As you noted, Congress has also effectively prohibited the
Federal Government from purchasing equipment from Huawei and
ZTE. How best can Congress continue to support this national
security effort?
A.2. Congress has already done a good deal to address the
national security concerns arising from the entities mentioned
and the presence of their equipment in the U.S. market. As
noted, section 889 of the 2019 NDAA prohibits agencies from
procuring Huawei and ZTE equipment. The issue of State and
local procurements of Huawei and ZTE equipment remain a
concern. Congress can continue to support diligence in
oversight of the implementation of section 889. Congress can
further promote American economic and military competitiveness
while addressing this issue by assisting the Administration
with actions and funds that advance U.S. artificial
intelligence (AI) development and innovation. The President's
American AI Initiative could also use the backing and focus of
Congress. Continuing to support America's decades-long
leadership in AI research and development will increase
national security while growing innovative industries and
creating cutting-edge, transformative technologies.
Q.3. Do you believe that this action will be disruptive to
American manufacturers that supply components to these
companies, and if so, do you think there is any way we should
address the collateral economic impact?
A.3. There are both positive and negative impacts on U.S.
manufacturing and the supply of components as a result of the
implementation of section 889 of the NDAA. For this reason, it
is important to also ensure that the application of export
controls continues to be targeted and surgical. Many State and
local providers who have equipment from these entities and who
contract with Federal agencies will need assistance. Congress
must be ready, where appropriate, with funding to ensure that
these State and local entities can address the economic
consequences of the law and avail themselves and their systems
with other technologies from trusted parties.
Q.4. You said in your testimony you believe the only area for
effective export control is computing hardware, which will
require multilateral collaboration with countries that have a
large amount of hardware engineering expertise. Which countries
do you believe are most pivotal for the United States to work
with in this regard?
How should Congress and U.S. Departments and agencies
decide which computing technologies should be subject to export
controls, and which should be areas where free and open
exchange of technology could contribute to the greater good?
A.4. Thank you for this question. I believe, however, that Dr.
Ben Buchanan in his opening statement testified on this matter.
As such, this question would be better answered by him or
someone with similar expertise.
Q.5. The rapid development of artificial intelligence brings
exciting possibilities. While it is important to safeguard our
technology, collaboration with global partners could help bring
mutually advantageous developments in the field.
Do you believe there is space to collaborate with China on
AI?
A.5. This is an important question with difficult answers.
Currently, U.S. companies do collaborate with Chinese entities
on AI technologies due to global manufacturing and supply
chains. Many U.S. companies manufacture in China. However,
given China's theft of intellectual property and other actions,
it is imperative that we deeply assess where and the degree to
which we have and continue to have such collaboration. We must
ensure that the United States and U.S. companies gain long-term
benefit from any such collaboration.
------
RESPONSE TO WRITTEN QUESTION OF SENATOR SINEMA
FROM NOVA J. DALY
Q.1. There appears to be consensus that a multilateral approach
to export controls is most effective in mitigating technologies
that threaten U.S. industry and national security. It also
appears there is a consensus that multilateral efforts will
work best in restricting divisive Chinese technology and
infrastructure. Given the importance of a multilateral approach
and the serious national security threats China poses, are you
at all concerned that the Administration's policies and
rhetoric on trade could undermine the necessary goodwill to
work collaboratively with our trading partners to hold China
accountable?
A.1. Collaboration with our international allies is always the
best response to address bad actors where it concerns the
protection of intellectual property. However, trade tensions
between the United States and its allies have arisen under
nearly all Administrations. One need only look at the yearly
``National Trade Estimate Reports on Foreign Trade Barriers''
issued by the United States Trade Representative to see that we
have had market barrier issues arising from allied countries
going back many years. While this Administration has taken a
stronger stance on addressing these issues, the United States
and its allies have historically found ways to work to address
trade tensions. So long as we hold in common the principles of
freedom, democracy, and the rule of law, the United States and
its allies will continue to find common ground on matters of
free and fair trade and also address together the negative
aspects of China's lack of IP enforcement and technology theft.
------
RESPONSES TO WRITTEN QUESTIONS OF CHAIRMAN CRAPO
FROM BEN BUCHANAN
Q.1. Big Data and AI--I know this is a little off topic, but
the Banking Committee held three hearings on privacy in the
``big data'' era, including how data is used to segment, score
or otherwise make predictions about an individual's
creditworthiness, employability, or general reputation. AI is
at the center of this discussion and I am concerned with the
extent to which individuals' data is collected and processed
without their knowledge, consent, or any real understanding of
use or scope. I believe individuals should have rights over
their data, including to access, control, correct and delete
it.
How do AI systems complicate or challenge the ability of
individuals to exercise data rights?
A.1. AI systems enable much better analysis of data. In this
sense, they increase the incentive for corporations to collect,
store, and examine data on wide swaths of Americans. Simply
put, deeper analysis is possible now than ever before because
of AI, much of it outside of the view of Americans.
Q.2. What risks are associated with AI in this context and how
can they be mitigated in any future legislative effort?
A.2. One substantial risk is that consumers do not understand
the way in which their data is being used to draw inferences,
via machine learning technology, about them. These inferences,
such as their buying preferences, can then be used to drive
advertising campaigns. While this risk has long existed,
machine learning technology and the associated rise of data
analytic tools amplifies it tremendously. While I do not have
specific legislation to propose, it seems to me that Congress
might investigate whether consumers are meaningfully consenting
to the way in which their data is being used.
Q.3. Google Data Privacy--Your testimony describes export
controls as relatively ineffective in stopping the export of
algorithms given the rate of innovation and the fact that AI is
a fairly open resource. You also identify the mass of personal
and behavioral data as the competitive advantage for large
technology companies, as opposed to their AI systems. It would
seem to me then that that data could also be the real
vulnerability, if for instance, a foreign adversary were to
obtain all of Google's consumer data.
These companies are incentivized to secure their systems,
but that may not be enough. My question then is what
comprehensive privacy controls or practices could help mitigate
the risk of big data being used in this way?
A.3. It is important to differentiate between privacy and
security. In general, I think top-tier tech companies like
Google have adequate incentive to secure their systems; other
companies do not take security nearly seriously enough, as many
years of breaches have obviously shown. When it comes to
privacy, the risk is not that a foreign hacker will access the
data--a security concern--but that the company itself will
misuse the data in a way that the consumer does not understand
or permit. As I indicated in my answer above, I think that is a
very serious risk, and the capability of machine learning
systems for ever-deeper analysis amplifies it further. It is
vital that American consumers understand what is happening and
consent to the terms when they interact with modern technology
companies.
Q.4. The United States has a special treatment arrangement with
Hong Kong with regards to export controls. While it is in the
United States interests to have a strong economic relationship
with Hong Kong, there is a lot of concern about growing Chinese
encroachment on Hong Kong's autonomy and the potential
implications for safeguarding technology.
Is our current export control policy equipped to deal with
risk of diversion from Hong Kong to China?
What are some ways in which China is using or could use
Hong Kong as a vector for acquisition of technology that we do
not export to the Mainland?
What are your specific recommendations for strengthening
our export control regime in relation to these challenges?
A.4. Unfortunately, I do not claim any regional expertise on
Hong Kong or its relations with China. I have never studied
these subjects in any kind of depth, nor am I familiar with how
export controls apply to Hong Kong. I must defer to other
experts on these three questions as a result.
------
RESPONSES TO WRITTEN QUESTIONS OF SENATOR WARREN
FROM BEN BUCHANAN
Q.1. At least one U.S. company has been found to have provided
the Chinese government with a tool enabling it to monitor
Uyghur and Central Asian minorities, as part of what one Uyghur
activist described in April 9, 2019, testimony to the Senate
Foreign Relations Subcommittee on East Asia, the Pacific, and
International Cybersecurity Policy as ``an Orwellian mass
surveillance State'' where ``more than one million Uyghurs are
arbitrarily detained outside the legal system in concentration
camps.'' A bipartisan group of Senators introduced the Uyghur
Human Rights Policy Act, of which I am a cosponsor, which
states in part, that:
the Secretary of Commerce should review and consider
prohibiting the sale or provision of any United States-
made goods or services to any state agent in Xinjiang,
and adding the Xinjiang branch of the Chinese Communist
Party, the Xinjiang Public Security Bureau, and the
Xinjiang Office of the United Front Work Department, or
any entity acting on their behalf to facilitate the
mass internment or forced labor of Turkic Muslims, to
the ``Entity List'' administered by the Department of
Commerce.
Do you agree? Please explain your view.
A.1. While it is difficult to know with certainty what is
happening in Xinjiang, and while I do not claim particular
expertise on the subject, I have certainly read a number of
news reports that are both credible and alarming. Like you, I
am very worried about the role of technology in aiding
repression around the world. I do not think it is appropriate,
nor should it be legal, for American companies to aid
authoritarian regimes in any effort to crack down on dissent,
prosecute religious or ethnic minorities, or otherwise repress
their populations.
Q.2. Can you conceive of any circumstances under which it would
be appropriate for the United States to weaken our export
control laws and regulations, or the enforcement of those laws
and regulations, vis-a-vis China or any other foreign
competitor in order to extract concessions or other commitments
from that foreign competitor on matters related to trade or
human rights? Please explain your view.
A.2. As I have indicated in other answers, my view is that
export controls put in place for national security concerns
should not be negotiated away for trade concessions. Doing so
undermines the credibility of American export controls.
------
RESPONSE TO WRITTEN QUESTION OF SENATOR SINEMA
FROM BEN BUCHANAN
Q.1. There appears to be consensus that a multilateral approach
to export controls is most effective in mitigating technologies
that threaten U.S. industry and national security. It also
appears there is consensus that multilateral efforts will work
best in restricting divisive Chinese technology and
infrastructure. Given the importance of a multilateral approach
and the serious national security threats China poses, are you
at all concerned that the Administration's policies and
rhetoric on trade could undermine the necessary goodwill to
work collaboratively with our trading partners to hold China
accountable?
A.1. Yes, as I indicated in the hearing, I am concerned that
the rhetoric and policies of the trade negotiations can, for a
variety of reasons, undermine the real and perceived importance
of national security concerns. In my view, export controls put
in place for national security reasons are not something to be
negotiated away, since doing so undermines their credibility of
the stated national security concerns. Further, I believe
export controls are most effective when done in a multilateral
fashion, and any effort to weaken American alliances undermines
the potential for strong export controls.
Additional Material Supplied for the Record
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]