[House Hearing, 117 Congress]
[From the U.S. Government Publishing Office]


                                    

                          [H.A.S.C. No. 117-7]

                        FINAL RECOMMENDATIONS OF

                    THE NATIONAL SECURITY COMMISSION

                       ON ARTIFICIAL INTELLIGENCE

                               __________

                             JOINT HEARING

                               BEFORE THE

                   SUBCOMMITTEE ON CYBER, INNOVATIVE 
                  TECHNOLOGIES, AND INFORMATION SYSTEMS

                                 OF THE

                      COMMITTEE ON ARMED SERVICES

                       MEETING JOINTLY WITH THE
                       
                   SUBCOMMITTEE ON NATIONAL SECURITY

                                 OF THE

                   COMMITTEE ON OVERSIGHT AND REFORM

                           [Serial No. 117-7]

                        HOUSE OF REPRESENTATIVES

                    ONE HUNDRED SEVENTEENTH CONGRESS

                             FIRST SESSION

                               __________

                              HEARING HELD

                             MARCH 12, 2021

                                     
[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]

                              __________

                    U.S. GOVERNMENT PUBLISHING OFFICE                    
44-410                   WASHINGTON : 2021                     
          
-----------------------------------------------------------------------------------                                     
  

                      COMMITTEE ON ARMED SERVICES
SUBCOMMITTEE ON CYBER, INNOVATIVE TECHNOLOGIES, AND INFORMATION SYSTEMS

               JAMES R. LANGEVIN, Rhode Island, Chairman

RICK LARSEN, Washington              ELISE M. STEFANIK, New York
SETH MOULTON, Massachusetts          MO BROOKS, Alabama
RO KHANNA, California                MIKE GALLAGHER, Wisconsin
WILLIAM R. KEATING, Massachusetts    MATT GAETZ, Florida
ANDY KIM, New Jersey                 MIKE JOHNSON, Louisiana
CHRISSY HOULAHAN, Pennsylvania,      STEPHANIE I. BICE, Oklahoma
    Vice Chair                       C. SCOTT FRANKLIN, Florida
JASON CROW, Colorado                 BLAKE D. MOORE, Utah
ELISSA SLOTKIN, Michigan             PAT FALLON, Texas
VERONICA ESCOBAR, Texas
JOSEPH D. MORELLE, New York

                Bess Dopkeen, Professional Staff Member
                Chris Vieson, Professional Staff Member
                         Caroline Kehrli, Clerk
                               
                               ------                                

                   COMMITTEE ON OVERSIGHT AND REFORM
                   SUBCOMMITTEE ON NATIONAL SECURITY

               STEPHEN F. LYNCH, Massachusetts, Chairman

PETER WELCH, Vermont                 GLENN GROTHMAN, Wisconsin, Ranking 
HENRY C. ``HANK'' JOHNSON, Jr.,          Member
    Georgia                          PAUL A. GOSAR, Arizona
MARK DeSAULNIER, California          VIRGINIA FOXX, North Carolina
KWEISI MFUME, Maryland               BOB GIBBS, Ohio
DEBBIE WASSERMAN SCHULTZ, Florida    CLAY HIGGINS, Louisiana
JACKIE SPEIER, California
                  Russ Anello, Majority Staff Director
                  Mark Marin, Minority Staff Director
                Dan Rebnord, Subcommittee Staff Director
                            
                            C O N T E N T S

                              ----------                              
                                                                   Page

              STATEMENTS PRESENTED BY MEMBERS OF CONGRESS

Grothman, Hon. Glenn, a Representative from Wisconsin, Ranking 
  Member, Subcommittee on National Security, Committee on 
  Oversight and Reform...........................................     7
Langevin, Hon. James R., a Representative from Rhode Island, 
  Chairman, Subcommitteeon Cyber, Innovative Technologies, and 
  Information Systems, Committee on Armed Services...............     1
Lynch, Hon. Stephen F., a Representative from Massachusetts, 
  Chairman, Subcommittee on National Security, Committee on 
  Oversight and Reform...........................................     5
Stefanik, Hon. Elise M., a Representative from New York, Ranking 
  Member, Subcommittee on Cyber, Innovative Technologies, and 
  Information Systems, Committee on Armed Services...............     4

                               WITNESSES

Clyburn, Hon. Mignon, Commissioner, National Security Commission 
  on Artificial Intelligence.....................................    13
Louie, Hon. Gilman, Commissioner, National Security Commission on 
  Artificial Intelligence........................................    14
Schmidt, Dr. Eric, Chairman, National Security Commission on 
  Artificial Intelligence........................................     8
Work, Hon. Robert, Vice Chairman, National Security Commission on 
  Artificial Intelligence........................................    10

                                APPENDIX

Prepared Statements:

    Langevin, Hon. James R.......................................    51
    Schmidt, Dr. Eric, joint with Secretary Work, Ms. Clyburn, 
      and Mr. Louie..............................................    54

Documents Submitted for the Record:

    [There were no Documents submitted.]

Witness Responses to Questions Asked During the Hearing:

    [There were no Questions submitted during the hearing.]

Questions Submitted by Members Post Hearing:

    Mrs. Bice....................................................    71
    Mr. Moulton..................................................    69


FINAL RECOMMENDATIONS OF THE NATIONAL SECURITY COMMISSION ON ARTIFICIAL 
                              INTELLIGENCE

                              ----------                              

        House of Representatives, Committee on Armed 
            Services, Subcommittee on Cyber, Innovative 
            Technologies, and Information Systems, Meeting 
            Jointly with the Committee on Oversight and 
            Reform, Subcommittee on National Security, 
            Washington, DC, Friday, March 12, 2021.

    The subcommittee met, pursuant to call, at 11:03 a.m., in 
room 2118, Rayburn House Office Building, Hon. James Langevin 
(chairman of the Subcommittee on Cyber, Innovative 
Technologies, and Information Systems) presiding.

 OPENING STATEMENT OF HON. JAMES R. LANGEVIN, A REPRESENTATIVE 
FROM RHODE ISLAND, CHAIRMAN, SUBCOMMITTEE ON CYBER, INNOVATIVE 
   TECHNOLOGIES, AND INFORMATION SYSTEMS, COMMITTEE ON ARMED 
                            SERVICES

    Mr. Langevin. Good morning, everyone. I call this joint 
subcommittee hearing together with the Subcommittee on Cyber, 
Innovation Technologies, and Information Systems, along with 
the Subcommittee on Oversight and Government Reform--National 
Security Subcommittee.
    Good morning, everyone. Before I begin my opening 
statement, I'm just going to give a brief technical readout for 
those members that are participating remotely. I think most 
members are participating remotely this morning.
    But, again, I want to welcome members who are joining us at 
today's joint hearing remotely. Members who are joining 
remotely must be visible on screen for the purposes of 
identifying--identity verification, establishing and 
maintaining a quorum, participating in the proceeding, and 
voting.
    Those members must continue to use the software platform's 
video function while in attendance, unless they are 
experiencing connectivity issues or other technical problems 
that render them unable to participate on camera.
    If a member experiences technical difficulties, they should 
contact the committee staff for assistance. Video of members' 
participation will be broadcast in the room and via the 
television internet feeds.
    Members participating remotely must be--must seek 
recognition verbally, and they are asked to mute their 
microphones when they are not speaking.
    Members who are participating remotely are reminded to keep 
software platform videos--video function on the entire time 
they attend the proceeding. Members may leave and rejoin the 
proceeding.
    If members depart for a short while for reasons other than 
joining a different proceeding, they should leave the video 
function on. If members will be absent for a significant period 
or depart for a--to join a different proceeding, they should 
exit the software platform entirely and then rejoin if they 
return.
    Members may use the software platform's chat feature to 
communicate with staff regarding technical or logistical 
support issues only.
    Finally, I've designated a committee staff member to, if 
necessary, mute unrecognized members' microphones to cancel any 
inadvertent background noise that may disrupt the proceeding.
    So with that technical message out of the way, I'm now 
going to proceed with my opening statement and then turn to 
Ranking Member Stefanik.
    Well, I want to welcome everyone to our joint hearing with 
the House Committee on Oversight and Reform Subcommittee on 
National Security, which will review the final recommendations 
of the National Security Commission on Artificial Intelligence.
    We welcome the subcommittee chairman, Stephen Lynch, my 
good friend from Massachusetts, and Ranking Member Glenn 
Grothman, and we are also pleased to host our House Armed 
Services Committee Chairman Adam Smith and Ranking Member Mike 
Rogers. Looks like we have a full house today. So I'm looking 
forward to this very meaningful and exciting hearing.
    So I'm pleased to welcome, of course, most especially, four 
commissioners for the National Security Commission on 
Artificial Intelligence, a commission that was created by the 
House Armed Services Committee in the National Defense 
Authorization Act for Fiscal Year 2019 to help us advance the 
development of artificial intelligence, machine learning, and 
associated technologies to prepare the defense enterprise for 
the national security challenges of the future.
    We asked this commission to produce a bipartisan whole-of-
government effort focused on solving national security issues. 
We appreciate the leadership and hard work of our witnesses and 
we owe each of them an immense debt of gratitude.
    Today, we welcome Dr. Eric Schmidt, chairman of the 
Commission; also, the Honorable Robert Work, vice chairman; the 
Honorable Mignon Clyburn, commissioner on the workforce and 
ethics lines of effort; and Dr. Gilman Louie, commissioner on 
the lines of effort focused on protecting and building on AI 
advantages, marshaling global cooperation, and threat analysis 
and response actions.
    While many of the commissioners hail from the tech sector, 
the world's understanding of artificial intelligence truly 
began in government defense labs, and specifically with 
investments by the Department of Defense Advanced Research 
Projects Agency, DARPA, and the Office of Naval Research.
    Now, decades later, we must redouble our focus on the power 
of defense science and technology research to propel us into 
the future. So I look forward to hearing the commissioners' 
recommendations on investments in basic and applied research 
and how to encourage faster adoption of innovative and cutting-
edge capabilities.
    The next-generation challenges are upon us, and in our last 
subcommittee hearing we talked about how the Department can 
transform innovation into reality, specifically by orienting 
ourselves, the software, and data capabilities that are often 
the beating heart of the platforms we require and that promise 
to dramatically improve decision-making and optimization 
processes.
    The battlespace of the future will be a complex web of 
software, networks, and data--and data integrated across 
domains and among our allies. Artificial intelligence and other 
next-generation innovations will be crucial in order to harness 
the power of data to give our men and women in uniform an edge 
in any future conflict.
    Our potential adversaries, of course, are already investing 
heavily in this future as well. So this Commission has 
undertaken the difficult task to articulate the potential of 
artificial intelligence and the risks and benefits that lie 
ahead.
    They have worked through these issues and identified 
recommendations related to research and software development, 
opportunities for international partnerships, safeguarding 
against our adversaries' advancements in this space, and 
cultivating a 21st century workforce.
    Above all, the Commission has crucial recommendations 
related to building and deploying AI [artificial intelligence] 
in an ethical manner that is respectful of human rights. 
Indeed, that last category is what sets our Nation apart.
    I commend the Defense Innovation Board, which was also 
chaired until last year by Dr. Schmidt, to helping the 
Department begin important discussions on ethics in AI.
    Last year, Ranking Member Stefanik and I, along with 
Chairman Smith and Ranking Member Thornberry, championed a 
package of provisions based on the Commission's first quarter 
recommendations that yielded 13 provisions in the National 
Defense Authorization Act for Fiscal Year 2021, with the 
majority deriving from the Commission's call to strengthen the 
AI workforce.
    That STEM [science, technology, engineering, and 
mathematics] talent is required today as much as ever to solve 
our most pressing national security challenges. Indeed, great 
power competition is also a race for talent. We must move past 
old models of training and learning, and establish a system to 
dynamically upskill our workforce as the technology evolves.
    Ranking Member Stefanik and I were pleased to invite 
Commission representatives for a review of the interim 
recommendations last fall and we look forward to hearing about 
your recently released final recommendations to Congress.
    Incredibly, there are over 100 in total and over 50 related 
to the purview of the Armed Services Committee.
    So we commend you for all your work that you put into this 
effort these past 2 years. We are grateful for that work and 
that due diligence, and we look forward to receiving your 
testimony today.
    So with that, I'll now turn to Ranking Member Stefanik for 
her remarks.
    [The prepared statement of Mr. Langevin can be found in the 
Appendix on page 51.]

  STATEMENT OF HON. ELISE STEFANIK, A REPRESENTATIVE FROM NEW 
    YORK, RANKING MEMBER, SUBCOMMITTEE ON CYBER, INNOVATIVE 
   TECHNOLOGIES, AND INFORMATION SYSTEMS, COMMITTEE ON ARMED 
                            SERVICES

    Ms. Stefanik. Thank you, Chairman Langevin.
    The National Security Commission on Artificial Intelligence 
is a critical step forward that I'm proud to have championed in 
the House with my colleagues on a bipartisan basis, and today's 
hearing is a culmination of years of hard work of our 
Commission
    Chairman Schmidt, Vice Chairman Work, Commissioner Clyburn, 
and Commissioner Louie, thank you for serving on the Commission 
and for testifying today. Your efforts will serve as a 
blueprint for how our country will respond to, develop, and 
lead the world in artificial intelligence capabilities.
    As we know, AI not only brings immense technological 
opportunities and innovation, but AI will also bring 
significant risks as our adversaries will deploy AI to 
challenge American interests and securities on our shores and 
abroad.
    And importantly, as you laid out in your final 
recommendations, AI will affect every facet of life going 
forward, from civil society to our economy and, of course, 
national security.
    For the Department of Defense [DOD] specifically, this 
final report is stark in its assessment. China will surpass the 
United States in AI leadership and win the innovation race if 
we fail to invest in emerging technologies and if we fail to 
take a whole-of-government approach to AI.
    The impact on our national security is profound and 
disturbing, and the report concluded that China will achieve 
superiority over the U.S. within the next decade if we don't 
solve our organizational and investment challenges by 2025, 
just 4 years from now.
    We face hard choices, given our limited resources to 
maintain that technological advantage over China. Future 
conflicts will take place on an AI battlefield and we must 
consider the future of systems that are not AI enabled.
    Simply put, DOD must be willing to take on risk and 
Congress should support those efforts. Further, the U.S. cannot 
win this competition if we don't have the right workforce. The 
Commission highlighted our talent deficit and concluded this 
problem is the greatest impediment to being AI-ready by 2025.
    Chairman Langevin and I are committed to solving this 
talent deficit, and last year we introduced legislation to 
retain technical talent here in the U.S. I also look forward to 
hearing more about the Digital Service Academy recommendation 
and other ways we can develop the necessary workforce within 
the DOD.
    Alternatively, our private sector is driving many of the 
advancements in AI and we should encourage increased 
collaboration between the Department and private sector 
partners.
    This subcommittee understands the issues many companies 
have interacting with DOD, primarily the onerous acquisition 
process. This report underscores the importance of reducing red 
tape so the Department doesn't hinder cooperation with the 
private sector.
    Again, I'm very proud of the work accomplished by this 
Commission and the work that we did to include many of the 
recommendations in last year's NDAA [National Defense 
Authorization Act].
    But more work must be done. Our warfighters must have most 
advanced technological capabilities to deter and defeat our 
adversaries in an AI environment.
    To improve the lethality capabilities of our forces, we 
must continue supporting the Joint Artificial Intelligence 
Center and enable the services and combatant commands to 
develop, tailor, and deploy AI systems to the battlespace.
    I look forward to the presentation today and the 
discussion, and I yield back.
    Mr. Langevin. Thank you, Ranking Member Stefanik.
    I now recognize Chairman Lynch for his remarks.

    STATEMENT OF HON. STEPHEN LYNCH, A REPRESENTATIVE FROM 
  MASSACHUSETTS, CHAIRMAN, SUBCOMMITTEE ON NATIONAL SECURITY, 
               COMMITTEE ON OVERSIGHT AND REFORM

    Mr. Lynch. Good morning, Mr. Chairman. Thank you very much.
    Before I begin, I do want to thank you, Mr. Chairman, for 
your continued leadership in the areas of cyberspace 
operations, artificial intelligence, and other developing 
technologies, all with critical implications for our national 
security.
    I'm pleased to join Chairman Smith, Ranking Member 
Stefanik, and Ranking Member Grothman as our subcommittee 
conducts today's important work to examine the final report 
released earlier this month by the National Security Commission 
on Artificial Intelligence.
    AI carries the remarkable potential to enhance and even 
transform our national security. We're already beginning to 
integrate AI algorithms, applications, and systems to 
facilitate intelligence collection and analysis, including to 
detect and prevent future terrorist attacks.
    We're also deploying AI to support battlefield medical 
evacuations, logistical missions, and military operations in 
Iraq, Syria, and Afghanistan, and other conflict zones.
    On the Financial Services Committee, where I serve as 
chairman of the Task Force on Financial Technology, we're 
seeing the use of AI and machine-learning technology to enhance 
international investigations to combat terrorism, financing, 
and money laundering.
    However, the evolution of artificial intelligence has also 
heightened the prospect that America's adversaries will win the 
race to develop and deploy AI, and to do so for malign 
purposes.
    To the great detriment of our national security, as 
reported by the National Security Commission on Artificial 
Intelligence, and this is a quote, ``AI is expanding the window 
of vulnerability that the United States has already entered. 
For the first time since World War II, America's technological 
predominance, the backbone of its economic and military power, 
is under threat,'' closed quote.
    Clearly, cybersecurity has become synonymous with national 
security, and our fundamental duty to protect our democracy 
requires that we become, quote, ``AI-ready,'' with resources, 
personnel, and strategies necessary to meet these urgent 
challenges.
    According to the Commission, however, we are a long way 
from that goal. Absent shifting trends, quote, ``China 
possesses the might, the talent, and ambition to surpass the 
United States as the world's leader in AI in the next decade.''
    In his 2020 book, ``The Kill Chain: Defending America in 
the Future of High-Tech Warfare,'' Christian Brose, who is a 
former staff director of the Senate Armed Services Committee 
under Chairman John McCain, articulates that, quote, ``A core 
pillar of the Chinese Communist Party's plan in harnessing 
emerging technology is to leapfrog the United States and become 
the world's preeminent power.''
    In fact, the 2017 development plan on artificial 
intelligence issued by China's State Council envisioned that 
China will lead the international AI sector as soon as 2030.
    It's also worth noting that in the race to develop and 
deploy AI, our adversaries, such as the Russian Federation and 
the People's Republic of China, do not struggle with the moral 
restrictions faced by democratic governments on the use of AI-
enabled autonomous weapons, nor are they hindered by moral 
considerations regarding the impact of AI on civil liberties.
    However, thanks to the expertise and the dedication of the 
NSCAI [National Security Commission on Artificial Intelligence] 
commissioners and their staff, this final report that they have 
released earlier this month sets forth a comprehensive 
blueprint to help the new administration and Congress allocate 
appropriate Federal resources toward the advancement and 
integration of AI technologies, technical infrastructure, and a 
digitally proficient workforce.
    We must also work together to ensure that these efforts 
maximize the opportunity for robust oversight, transparency and 
accountability that reflect our compelling national interest in 
safeguarding the civil liberties of all Americans.
    To that end, I'm proud to be an original co-sponsor of 
Representative Lori Trahan's upcoming legislation to establish 
a Digital Service Academy. The creation of a fully accredited 
university to train future public servants in artificial 
intelligence and other digital fields is a principal 
recommendation included in the Commission's final report.
    I want to thank you, Mr. Chairman. I look forward to 
today's hearing and discussing these issues with our 
distinguished panelists.
    And I yield back the balance of my time.
    Mr. Langevin. Thank you, Chairman Lynch. And again, I'm 
grateful that you and I could team up to bring our two 
subcommittees together on this very important topic, and I 
thank you for your leadership.
    With that, we'll now turn to Ranking Member Grothman for 
his remarks.

    STATEMENT OF HON. GLENN GROTHMAN, A REPRESENTATIVE FROM 
 WISCONSIN, RANKING MEMBER, SUBCOMMITTEE ON NATIONAL SECURITY, 
               COMMITTEE ON OVERSIGHT AND REFORM

    Mr. Grothman. All very pleasant today. Great to start off 
the day by hearing a nice Boston accent.
    I first want to thank Chairman Langevin and Ranking Member 
Stefanik for inviting us to join in this important meeting. I 
thank Chairman Lynch as well for having us.
    Advancing American technology safely and effectively should 
be a bipartisan policy priority. I also want to thank our 
witnesses here today, and I particularly thank you for not 
having us see you on a Zoom. To see you live and in person is a 
real treat for us Congressmen.
    You authored an impressive and thorough report on the 
future of artificial intelligence, or AI, and gave Congress and 
the executive a roadmap on how to proceed. Government use of 
artificial intelligence poses significant and potentially 
positive outcomes, but also significant challenges, 
particularly surrounding ethical use and data security.
    I think it's the duty of Congress to examine both the 
positives and negatives of AI prior to authorizing what is 
likely to be billions of dollars for decades.
    Your report highlights much of this but I want to focus on 
two main topics: improving the government and ensuring privacy.
    The purpose of civilian government use of AI should be to 
decrease the footprint and size while increasing the efficiency 
and effectiveness of the Federal Government. It would defeat 
the purpose of massive investment in an automation technology 
to simply expand the size and scope of agencies instead of 
streamlining the workforce.
    An analysis by Deloitte suggests that smart use of AI can 
save billions of man-hours and billions of dollars. The level 
of--this level of savings can only be experienced if the 
government makes cuts where AI allows us. We can see these 
benefits already taking place all over the government, like at 
the Social Security Administration and the Patent and Trademark 
Office.
    As the technology grows and advances, so must our 
workforce. The government growth--let me see here. The 
government must get better at recruiting and retaining top 
talent. To achieve the benefits of AI, we must be able to 
assure our fellow Americans that the data is safe and the 
technology is being used ethically.
    AI can be prone to false positives and negatives and 
overreliance on suspected patterns. It also relies on massive 
amounts of data in order to continue to learn and evolve. We 
must protect this data through data stewardship requirements, 
data transparency and disclosure rules, data governance rules, 
and data collection rules.
    These protections must be put in place. We can see the 
dangers of runaway AI use in China. Using AI to support 
genocide in Xinjiang and suppress democracy in Hong Kong 
provides insight into how our adversaries view and use this 
technology--I'll say rather than suppress democracy, I guess I 
should say suppress freedom--as a way to suppress dissent and 
to become a global and economic military power.
    It is vital that the U.S. counter these actions. I look 
forward to hearing from our witnesses today about how we can 
balance the new global arms race with government efficiency and 
privacy.
    Thank you. I yield back.
    Mr. Langevin. Thank you, Ranking Member Grothman, for your 
remarks. We will now hear from our witnesses, then move into 
the question and answer session.
    With that--and I just want to make sure that Chairman Smith 
or Ranking Member Rogers didn't have any opening comments.
    Mr. Smith. I'm good. Thanks, Jim.
    Mr. Langevin. Okay, very good. And I don't know if the 
ranking member is on, if he had any comments.
    Hearing none, okay. Then we will--we will now turn to our 
witnesses, as I said, and then move to the question and answer 
period.
    I'd like to now recognize Dr. Eric Schmidt, Chairman of the 
Commission. Dr. Schmidt is the co-founder of Schmidt Futures, 
was the technical adviser to the board of Alphabet, and before 
that, the CEO [chief executive officer] of Google. He has a 
distinguished record of contributions to the national security 
technology community, including chairing the Defense Innovation 
Board.
    Dr. Schmidt, as a commissioner on the Cyberspace Solarium 
Commission, I just want to say a big thank you for your 
commitment to ensuring that the two Commissions work together, 
and let me say how pleased I am to see you champion some of the 
Solarium's recommendations as well.
    So we are deeply in your debt. I'm grateful for your 
contributions in these areas, and I now recognize you to 
summarize your testimony for 5 minutes.

  STATEMENT OF DR. ERIC SCHMIDT, CHAIRMAN, NATIONAL SECURITY 
             COMMISSION ON ARTIFICIAL INTELLIGENCE

    Dr. Schmidt. Well, thank you, Chairman Langevin, Ranking 
Member Stefanik, Chairman Lynch, Ranking Member Grothman, and 
the members of all the committees.
    I am very pleased on behalf of all the commissioners to 
present 751 pages, which you asked us to produce more than 2 
years ago, and I could not be prouder of this report and its--
both the quality and the heft are worth noting.
    More than 2 years ago, this committee, the House Armed 
Services Committee, foresaw the huge impact of AI on our 
society and our national security, and also foresaw the future 
and potential, perhaps likely, threats from our opponents.
    And that process, along the way we worked with you in the 
NDAA a year ago to get some important changes in the 
legislation that will really help our Nation.
    Just overall, I cannot say enough about the way we've 
worked together, the support that you all have given us, and an 
opportunity that you've given us to serve the Nation.
    What I thought I'd do right now is just give a quick 
summary of where the report is, and my fellow commissioners can 
take you into some of the very interesting detail.
    We reached a number of overarching judgments. The first is 
that the government is not organized nor resourced to win the 
technology competition against a committed competitor, and it's 
not prepared to defend against AI-enabled threats and we 
strongly believe that our Nation needs to be AI-ready by 2025 
to defend and compete in the coming era of AI-accelerated 
competition and conflict.
    So we put the report into two parts. The first part is 
``Part I: Defending America in the AI Era,'' and it's 
fundamentally how the U.S. government can use AI technologies 
to protect the American people and our interests. It focuses on 
the implications of applications of AI for defense and 
security.
    The second part is ``Winning the Technology Competition.'' 
It's obvious, by the way, we should win that. Recommends 
government actions to promote AI innovation, promote national 
competitiveness, and protect critical U.S. advantages in the 
larger strategic competition with China.
    In the idea of simplifying what we need to get done, we 
came up with four priorities with a great many details, as 
you'll hear about.
    The first one is leadership. The government isn't quite 
ready for this fight. It's not organized in the right way. We 
need organizational structures that accelerate the government's 
integration of AI and the promotion of AI across the country.
    There needs to be something at the White House. We're 
proposing a Technology Competitiveness Council reporting in to 
the Vice President that would precisely monitor and drive this 
transformation that we need.
    And by the way, it's not just the government. It's also in 
private sector.
    Talent. As you've identified--a number of you have in your 
opening comments--there's a huge talent deficit in the 
government. We need to build new digital talent pipelines and 
expand existing programs.
    We need to cultivate AI talent nationwide and ensure the 
best technologists come to the U.S. and stay in the U.S. and 
don't go to our competitors. Seems sort of obvious but 
incredibly important to emphasize.
    In hardware, the AI systems are critically dependent upon 
powerful hardware and we, as a country, are too dependent on 
semiconductor manufacturing in East Asia and Taiwan in 
particular.
    Most cutting-edge plants are produced in a specific plant 
that's 110 miles from China. That's got to be an issue. We must 
revitalize U.S. cutting-edge semiconductor fabs and implement a 
national microelectronics strategy.
    We state very clearly in our report that the objectives is 
to stay two generations ahead of the Chinese effort. It could 
not be clearer, in our view.
    And the fourth, of course, is innovation. AI research is 
very expensive. We need the government to help set the 
conditions for broad-based innovation across the country.
    We need, for example, a national AI research infrastructure 
so more than the top five companies have the resources to 
innovate, and in particular, startups and universities need 
this facility.
    And we also need to add, we think over 5, 6, 7 years, up to 
$40 billion in annual funding in the next 5 years to cover AI 
R&D [research and development] for defense and nondefense 
purposes.
    And as you highlighted, Mr. Chairman, in your comments, 
there are other things that are crossing edge. The first is 
partnerships. We need to build coalitions with like-minded 
nations, the technology democracies--the techno democracies, in 
my own verbiage--to advance the development and use of AI in 
emerging technologies that support our values, which is 
critical. We spent a lot of time on our report talking about 
values.
    And the second one, consistent with the values, is 
responsible use. In the face of digital authoritarianism, we 
need, we, the U.S., need to present a democratic model of 
responsible use of AI for national security.
    You can imagine the opponents and how they might use or 
misuse these things. The trust of our Nation, the trust of our 
citizens, will hinge on justified assurance that the 
government's use of AI will respect privacy, civil liberties, 
and civil rights. We have a set of recommendations along those 
lines.
    I really thank you all for giving us this opportunity. It 
has been a true privilege for me to be part of this and to help 
lead it.
    Thank you very much.
    [The joint prepared statement of Dr. Schmidt, Secretary 
Work, Ms. Clyburn, and Mr. Louie can be found in the Appendix 
on page 54.]
    Mr. Langevin. Thank you, Chairman Schmidt.
    With that, next we will hear from the Honorable Robert 
Work, Vice Chair of the Commission. Secretary Work is a former 
Deputy Secretary of Defense and Under Secretary of the Navy.
    Secretary Work's commitment to innovative strategic 
thinking is well known, and I welcome him back before us today.
    Dr. Work, you are now recognized to summarize your 
testimony for 5 minutes.

STATEMENT OF HON. ROBERT WORK, VICE CHAIRMAN, NATIONAL SECURITY 
             COMMISSION ON ARTIFICIAL INTELLIGENCE

    Secretary Work. Thank you, Chairman Langevin, Ranking 
Member Stefanik, Chairman Lynch, Ranking Member Grothman, and 
members of the committee. Thank you for having us today. It's a 
great opportunity to testify before you.
    I'd like to follow Eric's broad overview with a focus on 
the parts of our report that deal with national defense. The 
Commission fears that our Armed Forces will lose their 
competitive military technical advantage within the next decade 
if they do not accelerate the adoption of AI across all their 
military missions.
    Now, the intelligence record, we think, is quite clear. I 
also would just like to note that we have a classified annex 
that I would commend to all of the members. It is a summation 
of the intelligence record of what we think or know what our 
competitors are doing with AI.
    We're not going to be able to defend against AI-enabled 
threats without ubiquitous AI capabilities of our own, and new 
warfighting concepts and paradigms. Without question, an AI-
enabled force is going to be more effective. AI-enabled systems 
can make targeting more discriminant and precise, thereby 
reducing civilian casualties and damage to civilian 
infrastructure and other protected entities.
    It will improve the tempo, speed, and scale of operations 
and it will enhance the way the battlefield can be monitored. 
It will help the way that commanders understand what is 
happening in the battlespace.
    It will also augment the abilities of service members 
including the way they perceive, understand, decide, adapt, and 
act in the course of all their missions.
    As Eric has said, if we are going to win this competition, 
we think we need to be what we call AI-ready by 2025, which by 
that we mean we will have the foundation in place for the 
widespread integration of AI across the force.
    There are four main ingredients to achieve this vision. 
First is top down leadership and strategic direction. On a 
transformation of the scale that we believe is necessary, you 
have to have strong top-down leadership.
    They set the priorities. They overcome the barriers to 
change. We think the JAIC's [Joint Artificial Intelligence 
Center's] new reporting structure established in the NDAA is a 
strong first step. Congress should also direct the Department, 
in our view, to form a steering committee on emerging 
technology that includes representation from the intelligence 
community, and this steering committee would drive action on AI 
and emerging technologies.
    Priorities should be implemented through a technology annex 
in the National Defense and Intelligence Strategies. The 
committee would align priorities, strategy, and resources 
across OSD [Office of the Secretary of Defense], the Joint 
Staff, and the intelligence community.
    Effective integration of the AI is absolutely going to 
require a close partnership between the technologists and the 
warfighters. To ensure technical expertise informs capability 
and requirements decisions at the highest level, we recommend 
that the U.S.--the Under Secretary of Defense for Research and 
Engineering be made the co-chair and the Chief Science Advisor 
for the Joint Requirements Oversight Council, or JROC.
    We also believe the Department should set specific AI 
readiness performance goals by the end of fiscal year 2021, and 
this will drive the outcomes of being AI-ready by 2025.
    Second, the Department, we think, must ensure it has in 
place the resources, processes, and organizations to enable AI 
innovation. The Department needs to establish a common digital 
ecosystem.
    It's called the Joint Common Foundation in the Department 
of Defense. That's good for a start. That's the technical 
foundation for all AI development fielding. So it's going to 
include access to a secure cloud, AI software, trained models, 
data, and algorithms, as well as high-performance computing 
power and a development environment that allows the entire AI 
stack to be put together in a way that is secure and will do 
what we expect it to do on the battlefield.
    We think JAIC should be designated as the Department's AI 
accelerant. The JAIC, in our view, should focus on 
applications, not on hard research. Essentially, the JAIC's 
role is to try to get as many applications into the field as 
possible, and they provide the resources through the digital 
ecosystem or the Joint Common Foundation.
    They also can provide subject matter expertise to support 
AI efforts across the Department without becoming a central 
clearinghouse. The Department has to expand the use of 
specialized acquisition pathways and contracting authorities to 
source and deliver the best AI.
    Ranking Member Stefanik mentioned this. Software and 
algorithms are just a different kettle of fish than ships, 
airplanes, missiles, et cetera, and we have to come up with 
ways that are specific to get those algorithms and models 
developed and into the hands of our warfighters.
    We have to come together also to reform the planning, 
programming, budgeting, and execution process. Congress has 
provided us--excuse me, provided the Department of Defense with 
an expanded toolkit of acquisition and contracting approaches, 
and the Department's effort to adopt AI will be impeded by 
processes that are unsuited to digital technologies and the 
pace of development of AI right now.
    The Department should also increase its overall S&T 
[science and technology] spending and increase AI R&D to $8 
billion annually by 2025.
    We think that's totally appropriate within the size of the 
DOD R&D budget, which is the largest in its history.
    Third, we think AI adoption has to be accelerated. We think 
one of the ways to get these algorithms and models across the 
valley of death and into the hands of the warfighter would be 
if Congress could create a dedicated AI fund specifically 
designed to speed operational prototyping and transition, 
overseen by the Under Secretary of Defense for R&E.
    We think the Department should prioritize adoption of 
commercial AI solutions, especially to its core business 
processes and administrative processes, as well as logistics 
and sustainment systems.
    Technologists should be integrated at every level in the 
Department, in the administrative side as well as the 
operational side. This would mean, for example, standing up AI 
development teams at the COCOMs [combatant commands].
    And fourth and finally, as Eric said, we need the adoption 
of these technologies among our allies and partners, and 
promote AI interoperability.
    Thank you again for the opportunity to testify here today 
and I look forward to your questions.
    Mr. Langevin. Thank you, Secretary Work. Greatly appreciate 
your contributions and efforts on this extraordinary report and 
effort. Thank you.
    We'll now receive testimony from the Honorable Mignon 
Clyburn. Commissioner Clyburn spent 9 years on the Federal 
Communications Commission where she worked to close the digital 
divide.
    She has a distinguished career fighting for diversity in 
the communications sector, and I welcome her back to share more 
of her thoughts on workforce and ethics from the Commission.
    Commissioner Clyburn, you are now recognized to summarize 
your testimony for 5 minutes.

   STATEMENT OF HON. MIGNON CLYBURN, COMMISSIONER, NATIONAL 
         SECURITY COMMISSION ON ARTIFICIAL INTELLIGENCE

    Ms. Clyburn. Thank you very much.
    Chairman Langevin, Ranking Member Stefanik--sorry, 
Stefanik. Stefanik. We're friends.
    [Laughter.]
    Ms. Clyburn. Chairman Lynch, Ranking Member Grothman--I'm 
having a tongue-tied morning--and members of the committee, 
thank you for the opportunity to appear before you to highlight 
the Commission's workforce recommendations.
    But first, I would like to thank the members of the 
committee, particularly Chairman Langevin and Ranking Member 
Stefanik, for your leadership and for advancing many of the 
Commission's recommendations in this area.
    Each time my fellow commissioners discussed the workforce, 
we arrived at the same conclusions. The military needs to have 
expertise both in and out of uniform or it will be unable to 
build the systems to perform the task described in our report, 
and the DOD is unlikely to develop expertise quickly enough on 
its own.
    As a result, if the Department of Defense is going to 
become AI-ready, especially by 2025, as we have recommended, 
congressional action will be needed.
    Allow me to briefly describe four high priority 
recommendations in the report.
    First, and most critical for the AI workforce, is the need 
for military and civilian career fields in software 
development, data science, and artificial intelligence. The 
inability of military digital subject matter experts to spend 
their careers working in digital fields is, arguably, the 
single most important issue impeding modernization.
    Without this career path, DOD will continue to struggle to 
recruit new talent, identify talent, and retain the talent it 
already has. I should note that many of the military and 
civilian experts we spoke with when the Commission started have 
since left government service because they were unable to 
continue working on AI.
    We must stop bleeding talent. This is a well-known problem 
with a relatively straightforward solution. Unfortunately, we 
have not seen enough progress and it is time for us to take 
concrete steps to address the hemorrhaging when it comes to 
talent.
    Our second priority is training junior leaders. We must 
fundamentally change how junior leaders use and interact with 
AI and other information-processing agents. Junior leaders must 
not only understand how to team with machines, but learn when 
to trust machine outputs.
    We recommend the military services integrate AI topics into 
pre-commissioning and entry level training for junior officers 
and training for both junior and senior noncommissioned 
officers.
    Our third priority is to incentivize emerging technology 
literacy among senior officers. We often speak of the need for 
a cultural change in DOD. But the most effective way to change 
culture is to change incentives.
    Using the Goldwater-Nichols Act's incentivization of joint 
competency as a model, Congress should require DOD to create an 
emerging technology certification process in critical billets.
    Service members should earn their certification by serving 
in noncritical emerging technology billets, fellowships with 
industry and academia, graduating certified courses, and 
earning commercial certifications.
    Finally, we recommend a United States Digital Service 
Academy, an accredited degree-granting university. The academy 
would help meet the government's needs for expertise in 
artificial intelligence, software engineering, electrical 
engineering, computational biology, and several other areas.
    Students would attend the school tuition free and receive a 
highly technical education. Graduates would then enter the 
government as civil servants with a 5-year service obligation.
    Our staff will be available to work with you on further 
details of each recommendation. But for your convenience and 
consideration, we have produced a draft--we have produced draft 
legislative language for your review.
    Thank you again for the opportunity to appear before you 
and I look forward to any questions you may have.
    Mr. Langevin. Thank you very much, Commission Clyburn. We 
appreciate your testimony today and what you had to say.
    And, lastly, we'll receive testimony from Mr. Gilman Louie. 
Mr. Louie is a co-founder and partner of Alsop Louie Partners, 
an early-stage technology venture capital firm, and he was the 
very first CEO of In-Q-Tel, the venture capital firm 
established with the backing of the Central Intelligence 
Agency.
    I welcome him now to share his insights from the Commission 
on technological advantages, global cooperation, and threat 
analysis.
    Commissioner Louie, you are now recognized to summarize 
your testimony for 5 minutes.

STATEMENT OF HON. GILMAN LOUIE, COMMISSIONER, NATIONAL SECURITY 
             COMMISSION ON ARTIFICIAL INTELLIGENCE

    Mr. Louie. Thank you very much.
    Chairman Langevin, Ranking Member Stefanik, Chairman Lynch, 
Ranking Member Grothman, and members of the committee, thank 
you for the opportunity to testify today.
    Our report took a broad view of national security to 
compass economic competitiveness, as Eric has noted, as well as 
defense, as Bob has discussed, and wanted to focus on a series 
of cross-cutting national security problems related to AI that 
needs urgent attention.
    What these have in common is that our adversaries are 
aiming to take advantage of the free and open nature of our 
society.
    First, our society's digital dependency leaves us 
vulnerable to emerging AI-enabled threats. For example, 
adversaries are using AI to enhance disinformation campaigns 
and cyber attacks.
    They're also harvesting data on Americans to build profiles 
of their beliefs, behavior, and biological makeup to be used 
for tailored attempts to manipulate or coerce individuals.
    This is a gathering storm of foreign influence and 
interference, and requires organizational and policy reforms to 
bolster our resilience.
    You should stand up a task force and 24/7 operation center 
to confront digital disinformation. The government needs to 
better secure its own databases and prioritize data security 
and foreign investment screening, supply chain risk management, 
and national data protection legislation.
    We need AI-enabled cyber defenses to protect against AI-
enabled cyber attacks. And as the pandemic has made clear, 
biosecurity must become a top tier priority in national 
security policy.
    Second, competitors are making every effort to steal our 
technology, research, and intellectual property. As the margin 
of U.S. technological advantages narrows and foreign efforts to 
acquire American know-how increases, we need to examine how to 
best protect our ideas, universities, labs, and companies 
without unduly hindering innovation.
    We need to modernize export controls and foreign 
investments screening to better protect dual-use technologies 
like AI. We need to protect U.S. research institutions as 
national assets. They need tools and resources to assess risks 
and share information as well as cybersecurity support, and we 
need to elevate intellectual property [IP] policy reforms as a 
national security priority in light of China's effort to 
leverage and exploit IP policies to its own advantage.
    Finally, to protect our country in all these areas, we need 
to better--have better intelligence. The report makes 
significant judgments that intelligence will benefit from AI 
more than any other national security mission. The intelligence 
community should integrate AI across all aspects of its work, 
from collection to analysis.
    We need to empower science and technology leaders in the IC 
[intelligence community]. We need to leverage open source 
information, and we need new approaches to intelligence fusion 
and human-machine teaming to develop better insights and 
augment human judgment.
    Let me close by saying that just as AI is posed [to] impact 
all sectors of society, it also poses to impact all dimensions 
of national security.
    I urge Congress to review the full range of national 
security problems addressed in this report and adopt our 
recommendations to address them.
    Thank you, and I look forward to your questions.
    Mr. Langevin. Thank you, Commissioner Louie. I deeply 
appreciate you lending your expertise and efforts to this 
Commission, and thank you for your testimony today.
    We'll now turn to member questions, and we'll recognize 
members for 5 minutes. I'll begin by recognizing myself for 5 
minutes.
    Dr. Schmidt and Mr. Louie, are there better ways that DOD 
could leverage industry and academia to field the AI systems 
more quickly than going through the normal acquisition 
pipeline?
    Will the software acquisition pathway, by way of example, 
provided in the FY [fiscal year] 2020 NDAA and the Budget 
Activity 8 software budgeting pilots help DOD's efforts?
    Dr. Schmidt. Thank you, Mr. Chairman.
    The pathway that you all provided is helpful but is not 
sufficient, and the cultural aspect of training people to treat 
software differently has taken--has been harder and taken 
longer than I thought.
    I do know many, many software companies who want to work 
with the government and, in particular, the DOD, and they 
cannot find a corresponding customer or user or buyer or 
someone who can work with them.
    My own view is that the DOD should set up some kind of 
technology insertion program where they, literally, go and try 
to get this stuff in because it's so strategic and so important 
to the mission of the DOD.
    Gilman.
    Mr. Louie. Thank you, Dr. Schmidt.
    Software and AI are joined at the hip, and until the 
Department is able to acquire software as software, not as 
hardware, not in the form of block upgrades, but as consumed as 
a field that fuels our system.
    We have a saying that we did back who worked--many of us 
who worked on SWAP [Software Acquisition and Practices] report. 
We said software is something that never ends. It's a continual 
process. But all of our acquisitions are designed for building 
big systems in these kind of monolithic upgrades.
    Our adversaries are not doing that. For us to be 
competitive and for us to have the best software as it's 
happening, we need to reform on how we do it.
    A single color of money is a start, but it's only a start. 
The culture needs to change. We need professionals who know how 
to acquire software, and understands the basic underpinnings of 
AI.
    Mr. Langevin. Very good. Thank you, Dr. Louie.
    If I could continue with you, Dr. Louie--Commissioner 
Louie. We're all aware that China views talent as central to 
its technological advancement. The Commission addressed many 
recommendations towards the U.S.'s need to attract and retain 
the best foreign talent to study, live, and work in the United 
States. Can you speak to why this is so important to our 
national security and how others will capitalize on our 
policies if we do not find a way to keep the best talent here?
    Mr. Louie. Sure.
    First of all, China realizes it has a major disadvantage 
when it comes to attracting talent. Talent comes to democracies 
where good ideas flourish and individuals can chase their own 
pursuits in order to improve research and development.
    In doing so, we must not give away that advantage. We 
attract the best and the brightest from all over the world for 
the past 50 years. It's a fundamental U.S. advantage. The 
Chinese realize that they cannot compete with our top 1 percent 
when it comes to AI.
    Their strategy is to use their talents to apply what we 
discover. That discovery capability lies in our ability to work 
not only in our universities and research labs from individuals 
both from here domestically and from all over the world, but to 
share those ideas in the open and shared platform called Open 
Science.
    China is not an open society. It does not believe in open 
science. We do. Please don't give up that advantage.
    Mr. Langevin. Yeah. Well said, Commissioner Louie.
    Finally, Dr. Schmidt, of all your recommendations that 
focus on the Department of Defense and its adoption of AI, what 
do you believe are the Commission's most consequential 
recommendations that Congress has not yet acted upon?
    Dr. Schmidt. There are many--and Commissioner Work can 
probably add here--there are many aspects of the 
recommendations around talent that have not been adopted, in 
particular, retaining specialized talent.
    And the other issue has to do with the regulatory 
structure. As you highlighted in your earlier comment, the 
regulations are, essentially, antithetical to prioritizing AI.
    They're built around large weapon systems of a hardware 
kind, and the real strength of our Nation will come from the 
strength of our software and AI activities.
    Mr. Langevin. Commissioner Work, did you have anything to 
add, Mr. Secretary?
    Secretary Work. Sir, I'll just say that there's three 
general areas. I don't have a single answer. So with your 
forbearance, I'll give you three.
    The first is under leadership and strategy. We think that 
tri-chair steering committee is absolutely central to provide 
the top-down leadership and push for the integration of AI 
throughout the force. And having the JROC as the co-chair--
excuse me, having the Under Secretary of Defense for R&E as the 
co-chair for the JROC will make sure the technology and 
capabilities and requirements are absolutely synced up.
    Under enabling the resources, processing, and 
organizational constructs, getting that common digital 
infrastructure, which means secure cloud, we hope we can get to 
a secure cloud for the Department.
    But having the algorithmic libraries, et cetera, and having 
JAIC established as the AI accelerator is probably the most 
important thing at the applications level. And then under 
accelerated adoption, it's trying to get these teams out to the 
field, out to the COCOMs.
    Dr. Schmidt and I visited SOCOM [U.S. Special Operations 
Command], for example, and what they're doing down there, where 
the combatant commander himself has taken this as a personal 
mission and established the talent and team to push it, has 
really seen some remarkable advances in special operations 
capabilities.
    We need to have those type--same type of teams and 
approaches in the Indo-Pacific Command and European Command. So 
thinking of it in terms of leadership, enabling resources and 
processes, and accelerating adoption would be the three things 
that I would say are the most consequential.
    Mr. Langevin. Thank you, Secretary Work, and thank you all 
for your answers.
    With that, Ranking Member Stefanik is now recognized for 5 
minutes.
    Ms. Stefanik. Thank you, Chairman Langevin.
    An idea that each witness and each commissioner has 
highlighted today, as well as the chair of the Oversight and 
Government Reform Committee, is this bipartisan proposal for a 
Digital Service Academy and how the issue of workforce and 
talent is going to make or break whether the United States is 
able to lead in AI.
    So my question is for Chairman Schmidt. Can you expand upon 
this recommendation to include how such an academy should be 
established? And then I also want to hear what have been the 
impediments to this recommendation.
    Dr. Schmidt. If it's okay, I prefer that Commissioner 
Clyburn answer.
    Ms. Stefanik. Great.
    Ms. Clyburn. Thank you, again, very much.
    The need is very clear. We have a clogged pipeline, and so 
the Commission was very bold in its recommendation. It 
recognizes that there are a number of impediments.
    Many of them are economic, and so what we attempted to do 
here was to identify what could be a platform that would be 
targeted when it comes to STEM education, be unapologetic about 
educating to meet the needs that we have in government from a 
civilian point of view, and this academy is a part of the 
solution.
    Some of the things that I have heard that have been maybe a 
little less embracing is what does it mean in terms of the 
other service academies. You know, do we need another 
institution? How expensive will it be? How long will it be to 
onboard?
    It is not going to be inexpensive. It will take up to 7 
years to graduate the first class. But what I will say is we--
that pipeline problem that I recommend--that I mentioned, 
everyone knows that it's there.
    We need to be big, bold, and targeted and intentional, and 
this is one way we thought it was unwaveringly, you know, 
obvious that we were serious about addressing the pipeline 
deficit.
    Ms. Stefanik. Thank you, Ms.----
    Dr. Schmidt. May I--may I----
    Ms. Stefanik. Oh, sure. Go ahead, Eric.
    Dr. Schmidt. Yeah, let me just--let me add that a number of 
universities have offered to help us get this set up if the 
legislation proceeds as we've described, and it should 
ultimately be very good economically because of the payback 
requirement. If you graduate, you have to work for the 
government for 5 years.
    We strongly endorse this idea.
    Ms. Stefanik. Well, I appreciate that. I think this is 
going to be one of the most important ways and important 
strategies that we need to work on a bipartisan basis to 
embrace in order to make sure that we maintain a competitive 
edge.
    And my other question is on the timeline issue related to 
workforce and talent.
    Commissioner Clyburn, you talked about the 7-year time 
periods to even graduate students from a Digital Service 
Academy. But yet the report talks about this 2025 timeline in 
order for the--for the United States to maintain that edge 
against China.
    Are there any steps we can take now to address this talent 
deficit before 2025 or before that 7-year period with the 
Digital Service Academy, and if yes, what are those steps?
    Ms. Clyburn. Yes, ma'am. Thank you for that, and forgive me 
for struggling with your name. We say a lot of long A's and 
long E's so forgive me. I'm from South Carolina.
    Ms. Stefanik. It's okay.
    Ms. Clyburn. So one of the things that we recommend is an 
agency-specific digital corps, and this would be modeled after 
the Army's Medical Corps. And this would allow for specialized 
personnel and policies and guidelines for promotion.
    One of the issues, when I mentioned those who have left 
government, is there was no way to stay on an AI track and to 
get promoted and for us to benefit from that. The National 
Reserve Digital Corps, again, another civilian track, you know, 
modeled after the other, you know, reservist opportunities.
    We can quickly get those committed and on board to offer 
38--at least 38 days a year for service. They could triage. 
They could help. They could assist and they could augment.
    The scholarship--we have, you know, scholarship service, 
you know, programs that we need to expand and we need to expand 
the cyber--well, again, the cyber corps for civilian--excuse 
me, scholarship service.
    That would be NSF [National Science Foundation] managed. 
That would allow those who qualified to work in an agency. 
There are some things that we can do right now that will ensure 
a pipeline while we wait for that first graduating class.
    Ms. Stefanik. Great. Thank you so much, Commissioner 
Clyburn. Thank you, Chairman Schmidt, and all the commissioners 
for your emphasis on the workforce challenge.
    Yield back.
    Mr. Langevin. Thank you, Ranking Member Stefanik.
    The chair now recognizes Chairman Lynch.
    Dr. Schmidt. Chairman Work wanted to make a comment.
    Mr. Langevin. Oh.
    Mr. Lynch. Thank you, Mr. Chairman.
    First of all, I want to thank the panel for all of your 
work. First of all, coming here today in person, I think, 
speaks to the urgency of the issue, and we appreciate that.
    And also, we're thankful for this report. You know, 
normally in Congress when we see a 750-page document, we assume 
that it was compiled for the purpose of defending itself 
against the risk of being read.
    But I'm happy to say that your document is an exception. 
And I want to just point out to my colleagues and staff that in 
Appendix D, so not only did it make recommendations about what 
might be done to address the problem, but it also includes 
draft legislation in rudimentary form to really--so it's all 
worked out, pretty much.
    It can be refined a bit. But it's a head start on a lot of 
the work that we are engaged in here at the Capitol, and I 
appreciate your work.
    I do want to follow up on Ms. Stefanik's inquiry regarding 
several of you have talked about the dearth of talent, and the 
fact that we have got to get more young people into the 
pipeline.
    I am a co-sponsor of Ms. Trahan's bill for the academy. 
But, you know, I've had a similar problem in my own district. I 
represent a big part of Boston. A lot of the jobs in my area 
that are being created were really heavily reliant on a math 
and science background, and in the traditional public schools 
we weren't getting that.
    So I actually founded a charter school that triples the 
amount of math and science that the kids would have gotten had 
they gone to the traditional public schools.
    So--and there are a lot of schools around the country that 
are doing this, both traditional public schools and charter 
schools.
    Is there a way that we might be able to incentivize that 
type of activity? Because we have got to not just think about 
people who might serve in government tomorrow, literally, but 
also increasing that or animating that pursuit around issues 
like cyber, artificial intelligence, you know, so many other 
areas that are coming at us at a pace that is unprecedented.
    The velocity of change is breathtaking in terms of what 
we're grasping, you know, both in Congress, but also in our 
society. Is there a way that we might incentivize that learning 
at a much lower level than we're talking about for this 
academy?
    Ms. Clyburn. Yes, sir, I believe we can. I think we need to 
just demystify what STEM is or what STEAM [science, technology, 
engineering, the arts, and mathematics] is. When people hear 
about that they think it's for a certain segment of the 
community. They don't recognize that if a young person creates 
any type of product, a designer that creates lipstick, that's 
science.
    And so what we need to do, I think, a better job of 
messaging is saying, this is what it is. It is a part of your 
everyday--you know, your everyday culture. This is science. 
This is engineering.
    And we need to help our teachers become better facilitators 
and supportive, and we really need to recognize some of the 
existing cultural barriers when it comes to especially women 
and underrepresented groups.
    So afterschool programs, that's important. NSF has, you 
know, programs. We need an all-of-the-above approach in order 
to erase some of the challenges that you see, and kudos to you 
for starting the charter school with that focus.
    Mr. Lynch. Mr. Chairman, we're about to consider several 
immigration bills in Congress. We have already begun debate 
about that. They're being put together.
    Is there an opportunity for us to use the visa program and 
that immigration bill as a way of getting talent? I see that 
Canada does a very good job at this. There's actually a little 
tension between the Chinese government and the Canadian 
government because Canada has been so successful in recruiting 
some really top-line talent from China and Asia.
    You know, I tend to think we should be doing the same 
thing, using the promise of America to attract, you know, some 
of the--some of the talent that would like to get to work on 
some of the issues that we have in the country.
    Is there a way to do that through our upcoming immigration 
debate?
    Dr. Schmidt. There's no question that the United States 
will be stronger if we encourage high-skills immigration. In 
the last administration, mostly what would happen is that the 
visas would be very difficult to get and that companies would 
park employees in places like Canada, Vancouver, and so forth, 
waiting for the H-1B lottery.
    This is not in America's interest. So everything that we 
can do to get high-skills immigration into the right places is 
welcome.
    The argument is relatively simple. If we don't welcome 
them, they will create companies and efforts in countries that 
may ultimately not consistent with our best interest.
    The other obvious point is once we let them in the country 
or once we educate them in the country, we need to give them 
some way of staying in the country, consistent with the law and 
their good behavior.
    It's stupid, frankly, if I may say that, to fully educate a 
brilliant quantum physicist and then send him to China where he 
creates a quantum physics program that competes with our 
military activities, which, indeed, is what we did.
    Gilman, you may have something to add here.
    Mr. Louie. Thank you, Dr. Schmidt.
    We had some very specific recommendations in our report.
    First, we should expand our O/M/J visa programs. We have 
many talented individuals who want to stay in the United 
States, but they don't do it through the classical means of 
having published reports and things of that nature. For that, 
that's terrific.
    But AI is moving so rapidly, we need to open our aperture 
on who should be eligible for that class of visa.
    Second, we want entrepreneurs, entrepreneurs who come to 
the United States, create jobs for Americans, and we should 
develop and expand that lane for individuals, particularly in 
these areas of high competitions in science and technology. We 
want those entrepreneurs. We want them to create jobs here and 
build great American businesses.
    And finally, we made a recommendation. With the appropriate 
screening in place, we believe that anybody receiving a 
doctorate degree in the science and technology area or areas of 
critical requirements should be granted a visa.
    We want these individuals to stay, not go back and develop 
competing capabilities elsewhere.
    Mr. Lynch. Mr. Chairman, thank you.
    Mr. Langevin. Very good. Thank you, Chairman Lynch.
    Yes?
    Secretary Work. I was just going to say, and this goes back 
to a Ranking Member Stefanik's question, you know, we looked at 
this as a--we wanted to exploit our homegrown talent. We wanted 
to attract worldwide talent. And we wanted to ID [identify] and 
use the talent we have on hand.
    So exploiting homegrown talent, Commissioner Clyburn 
continually just said, look, we have got a lot of people who 
want to get into STEM. They don't have the opportunity to do it 
because they can't afford to go to college.
    So the national research--the National Digital Corps--
National Reserve Digital Corps was designed to do just that. 
It's like the NROTC [Naval Reserve Officers' Training Corps] 
program or the ROTC [Reserve Officers' Training Corps] program.
    Anybody can apply. They'd get a full ride, and when they 
graduate, they would owe--they would--38 days a year like a 
National Guardsman or someone in the National Guard where 2 
days out of every month, they would come into a unit and say, 
how can AI or these advanced technologies help you accomplish 
your mission better. I'm here to help you do that. And then 2 
weeks out of the year, they'd go to a military exercise and 
say, geez, if you just implied machine learning in this 
particular application, you're going to be 15 times more 
effective.
    And so that was really trying to attract homegrown talent. 
The National Digital Academy is designed to get people into the 
government for a long period of time so we can exploit them.
    Attracting the worldwide talent, Commissioner Louie and 
Commissioner Clyburn have already talked about this. But the 
third thing I wanted to say, which we haven't talked, is, 
there's a lot of talent in the Department of Defense right now. 
These young men and women, many of them are great coders. All 
they want to do is have an opportunity to get on a software 
development team and they will rock the world.
    So many of the suggestions we have is how do you identify 
those people? Give them a classifier that we can follow and 
assign, and that's how we're confident we could get to 2025 
before we're getting 700 graduates from the Digital Academy. So 
we took it from a holistic view, Mr. Chairman.
    Mr. Langevin. Thank you for that.
    Mr. Lynch. Mr. Chairman, I yield back.
    Mr. Langevin. Thank you, Chairman Lynch.
    Mr. Grothman--Ranking Member Grothman is now recognized 5 
minutes.
    Mr. Grothman. Sure, a few questions.
    There's nothing--I didn't anticipate the debate here would 
wind up over immigration. But it's an interesting topic. Do you 
think we should have some sort of math and science tests as we 
decide who's going to be able to immigrate in this country?
    Mr. Langevin. I don't think your microphone is on.
    Mr. Louie. I think math and science is critical. I think, 
particularly in our recommendations for the doctorate lane, 
clearly, S&T should be one of the testing variables that we 
should be looking at.
    That's the talent that we need in this country and we 
should be very explicit to the rest of the world. If you want 
to come to the United States and use your talents and use it 
well here, we welcome you.
    Mr. Grothman. Absolutely. When I talk to immigrants in my 
area, they sometimes complain that the smartest people are 
going to Australia, New Zealand, Canada, and we're getting the 
ones who aren't. So if the chairman wants to put a math and 
science requirement in immigration, it's something to look at.
    Now, we talk about ways and the government getting more 
involved in AI. Are there areas of the government that you 
think will be able to be cut as we improve in artificial 
intelligence, since presumably, it will, you know, improve a 
variety of things? I feel sorry; all of a sudden now Chairman 
Schmidt had to put a mask on. It was great seeing you without 
the mask.
    Dr. Schmidt. I'm trying to follow the rules.
    Bob, do you want to answer the government question?
    Secretary Work. Well, I--you've sent shivers down my spine 
because now I'm going back to my time as the Deputy Secretary, 
and every time a question like this came up, I would always 
hate to answer it.
    But, generally, we know that AI is going to have a 
tremendous impact on all the back office processes in the 
Department of Defense and the Federal Government, and it will 
become more efficient and you will require fewer people to do 
the work.
    So I would expect this to have some reduction in the 
overall Federal workforce as AI is completely implemented 
across the government and in the Department of Defense.
    Mr. Grothman. Do you think it would be a good idea to have 
targets right now? Because you know what happens. Nobody ever 
wants to cut the government.
    And I agree with you. Hypothetically, one of the benefits 
of artificial intelligence is that it should make things more 
streamlined and reduce the number of positions.
    But people around here don't like to do that and that's 
what I'm trying to ask you right now so we know in advance when 
we're putting together the budget in 2028 where we can expect 
to have reductions in personnel.
    Dr. Schmidt. I think it's reasonable to expect that some of 
the overhead functions will be smaller and that the specialized 
functions will be larger, and the easiest way to achieve that 
would be not to do what you're describing, but instead, put 
some guidance on the kind of people that are being hired now 
because the hiring pipeline takes a while.
    So the government loses, you know, 10 percent of the 
employees a year and it's hiring 10 percent new, some number 
like that. You all would know the exact number. You could 
establish a threshold that among the new hires that occur every 
year, a certain percentage of them have to meet a science and 
technology type threshold, a specialized threshold, and that 
would achieve your objective without having to go through the 
fighting over budgets argument.
    We're really focused on the people, not the money, because 
the people will drive everything else.
    Mr. Louie. I also think that the other part of the equation 
here is increasing the productivity of our people. Look, our 
adversaries are not holding back. You know, China has already 
set out its goals: to compete with the U.S. in AI by 2025--
underneath ``Made in China 2025''--to surpass the U.S. 
capabilities by 2030, and to win in any domain anywhere in the 
world in any kind of a hostile action by 2049.
    The issue is not so much the savings but where do we deploy 
our talent? Where do we put them in places where they are going 
to be the most productive and how are we going to compete with 
an adversary who's committed its entire nation to win in the 
fourth industrial revolution?
    We need to meet that challenge head on. We need to deploy 
our personnel, educate them, and skill them to be able to fight 
in this next arms race.
    Mr. Grothman. What majors right now in American colleges, 
to you if you're going to major in that, signals that you would 
be good in the AI field?
    Mr. Louie. Mathematics, clearly. Statistics, for example, 
is something that we should be teaching in our high schools 
because it's critical to the way we're thinking.
    Second, even if you're in non-S&T areas, understanding how 
AI--what you can trust it for, what you need to augment it 
with, what you need to question about it, is going to be 
critically important as well.
    Ms. Clyburn. And, sir, I would follow my niece in 
biomedical engineering.
    Dr. Schmidt. And let me add that there is good news here. 
Universities of the United States are generally seeing a huge 
supply of computer science graduates and majors, and in 
virtually every university I've studied, computer science is 
now the number one major ahead of, for example, economics, 
which is sort of a big surprise.
    That workforce is coming into the private sector. It's not 
coming into the government sector the way it should be. Because 
in the private sector, they get to work on this stuff. For the 
government, they have to do this billet and that billet and so 
forth.
    When I was running the DIB [Defense Innovation Board], we 
met a brilliant--we were doing a Russia review in the National 
Security Council--this brilliant young man who was busy doing 
cyber attack analysis, and I asked what was his career path. He 
said, next week I'm being transferred to a nontechnical 
position in another part of my rotation. That's insane, and 
that's how the government works. We need to address those 
issues, and those are covered in our report.
    Mr. Grothman. Thank you for giving me additional time.
    Mr. Langevin. Thank you, Ranking Member Grothman.
    Mr. Larsen is now recognized for 5 minutes.
    Mr. Larsen. Over here. Thank you, Mr. Chairman, for 
recognizing me.
    Twenty years ago when I came to Congress, I don't think I 
would ever predicted 20 years later I would have said, I'm 
really excited to read a report on AI. But I am, and two 
particular sections or pages, I'm real excited about pages 77 
and 78 and about 297 through about 300. Those are the education 
bits, and I'll get the--Mr. Work, I'll let you know. You don't 
have to read it.
    But on that, a few years back we included in the NDAA 
direction to the JAIC to develop an AI education strategy, 
which they are beginning to roll out. My vision of it is to go 
as deep as possible with as many folks--if I could just ask the 
guys on the clock, they didn't reset the time.
    Sorry about that. Yeah.
    I think every woman and man in uniform and civilian in the 
DOD should have some basic AI education. They don't have to be 
coders. They don't have to be the ones leading the project, but 
have some basic education, which is why page 77 is interesting. 
I'm excited about having an AI-ready DOD by 2025 or warfighters 
are enabled with baseline digital literacy and so on.
    And then getting to pages 297 to 300, implementing that 
even more. So I want to start with you, Commissioner Clyburn. 
What exactly, you know, is your vision within the DOD on AI 
education? Again, not everyone needs to be a coder. Not 
everyone needs to be writing the software. But my idea is that 
they should at least understand the weapon they're using, 
whether it's an electronic or a software weapon, a hardware 
weapon, the weapon someone else is using against us.
    What's the--what was the vision of the Commission when you 
came up with these recommendations?
    Microphone.
    Ms. Clyburn. It was a recognition that the next major 
conflict will likely not be on the ground. But it will be AI-
inspired. And so you're right, from headquarters to the 
tactical edge is how we phrase it and I really embrace that.
    We have to be AI-ready. We have to be AI-trained. Those who 
protect us should have that digital foundation. Those who excel 
in computational thinking within the ranks, they should be 
identified and supported.
    It's a shame that we're losing all that talent, and they 
should be able to advance. You know, those who are employ--we 
need upskilling within the ranks. Junior officers from the 
beginning need to be trained in AI concepts and they need to 
continually be educated and certified.
    So we can't stand still. We cannot ignore the fact that, 
again, we are in an AI--we might not recognize it, but we are 
in a AI revolution and we need to be prepared.
    Mr. Larsen. Did you look at the JAIC's current work done on 
this education strategy with these recommendations?
    Ms. Clyburn. Yes, we did, and I will yield the rest of my 
time to Mr. Work.
    Mr. Larsen. Yeah, I was going to ask him. Good.
    Secretary Work. Well, I'd just like to start, sir. Andrew 
Moore--Dr. Andrew Moore, formerly of Carnegie Mellon, now at 
Google, is one of our commissioners who leads line of effort 
one, which is the research and development line of effort.
    And I remember quite clearly the first time he was talking 
to us and he said, look, you can get a lot accomplished with 
young men and women who aren't a computer science graduate.
    What they have to have is understanding of is computational 
thinking, and that had a big impact on the way we were thinking 
of this because he made the case that there's a lot of innate 
talent in the force, we just have to identify those people.
    He recommended, for example, if we gave just one class, 
like, at the 7th grade, and another class in the 11th grade on 
computational thinking, that you would have people who graduate 
who would be immediately able to step onto a COCOM development 
team.
    So this led us to say we ought to add a section on 
computational thinking in the Armed Forces--the ASVAB, the 
Armed Services Vocational Aptitude Battery--and identify those 
folks so that we could do it.
    So computational thinking is as important for the whole 
force as the hardcore Ph.D.s who would go into dedicated R&D 
billets. So that would be one thing that I would say we all 
need to think about how we would do that, and we recommend a 
new National Defense Education Act II, and I would recommend 
that Congress think about, you know, should there be 
computational thinking type requirements that we establish to 
allow the entire workforce that's going into the government and 
into commercial sector, academia, et cetera, to expand these 
things.
    I'd also like to go back to one of the things on----
    Mr. Larsen. If I--if I could, I wanted to go to Mr. 
Schmidt.
    Secretary Work. I'm sorry, sir.
    Mr. Larsen. The--I'll be really quick with a question. We 
talked in the past about China's declarative policy to want to 
lead by 2030 in AI, and so why don't we just have a declarative 
policy we're going to lead by 2029 and just go do it.
    We have no idea how China is going to do it. They're just 
going to say they're going to do it. And, frankly, whether or 
not they're successful they would probably say they are 
leading, in my view.
    Why don't we just have a declarative policy we're going to 
lead in AI? Does being AI-ready by 2025 in the DOD, as you 
state in your report, does that equate with the general U.S. 
leadership if we do the things in this report?
    Mr. Langevin. If you can answer briefly.
    Dr. Schmidt. It's a component but not sufficient. We 
actually call for a technology competitive council to cover not 
just AI but some of the other key areas. It's crucial that the 
United States have a national plan for competitiveness globally 
that addresses AI, semiconductors, synthetic bio, and a few 
others, energy, et cetera.
    Mr. Larsen. Yeah. Thank you.
    Thank you, Mr. Chairman, for your indulgence. Thank you.
    Mr. Langevin. Thank you, Mr. Larsen.
    The ranking member of the full committee, Mr. Rogers, is 
now recognized.
    Mr. Rogers. Thank you, Mr. Chairman. Can you hear me?
    Mr. Langevin. Yes, I can.
    Mr. Rogers. Great.
    Dr. Schmidt, just to follow up on Mr. Larsen's question, is 
China leading in AI?
    Dr. Schmidt. We spent a lot of time looking at this 
question, and, in general, we are a little bit ahead but not 
very much. They have more people coming. We're doing better in 
algorithms, but they're coming.
    They are doing better in some industries, in particular, in 
financial transactions, electronic commerce, and surveillance. 
We're doing better in some other areas of research. It's a 
close race.
    At the moment we're ahead. Our report specifically says 
that we can lose our lead within a few years.
    Mr. Rogers. Well, I share Mr. Larsen and Mr. Lynch's 
observation. This Commission did some great work. This report 
is not just going to sit on a shelf. We're going to do some 
stuff here that's really meaningful. So I want to commend all 
of you and thank you for your time in putting this together 
because it's going to make a difference.
    Commissioner Clyburn, I wanted to follow up on Ranking 
Members Stefanik's questions about how we can get folks into 
these jobs earlier than the 7-year timeline it would take to 
graduate our first class.
    Are there some scholarship opportunities that we can 
provide people now to go to an existing university? We have got 
some fine existing universities in this country right now that 
can get some people trained earlier while we're standing up the 
Digital Service Academy and getting the first few people 
through there.
    Ms. Clyburn. So we talk about the Reserve Officer Training 
Corps. We talk about expanding that CyberCorps Scholarship for 
Service program that is--that it is in place.
    So, yes, sir, and there are some things that--outside of 
that that I think are possible. So being targeted and sending 
the signal. If you were to highlight just the two or three 
things that I mentioned, I think that will encourage others to 
follow.
    So it's that down payment that we should make with the 
existing programs and others modeled after successful programs 
that I think will make a difference and enable us to move in 
the right direction.
    Mr. Louie. Let me also add that not every field requires 
the degree.
    Mr. Rogers. Exactly.
    Mr. Louie. We have a great--we have a great enlisted force. 
I mean, amazing. These kids know all about AI. How do I know 
that? They're playing it on their video games.
    Mr. Rogers. Yes.
    Mr. Louie. What they're missing is the linkage between what 
they're seeing in those games and what they're seeing in their 
Department of Defense.
    Mr. Rogers. Yes.
    Mr. Louie. So with the appropriate certification and 
badging program, we could also get our enlisted teams up and 
ready and AI-certified and ready.
    Dr. Schmidt. And imagine----
    Mr. Rogers. Yeah, and that's what I was getting at, too.
    Dr. Schmidt. Sir, imagine if you had a call for talent 
where there was some kind of a test and a competition. I think 
all of the commissioners would say to you you're going to be 
really pleasantly surprised by the existing talent in the 
government that's not correctly being used.
    Mr. Rogers. Yeah, and I agree, and that's one of the things 
I was hoping is that, you know, when we have this structure, 
set up this Digital Service Academy, that it's not just focused 
on getting people a BA [bachelor of arts] or an MS [master of 
science] or Ph.D., that there are--other folks can get 
certified that can go ahead and get in the workforce right 
away.
    And I'm also hopeful--you know, I first talked about this 
concept with Dr. Schmidt about a year ago and very excited 
about it. I'm curious to know, do you envision this being a 
purely public sector entity that we stand up or is it going to 
be a public-private partnership? How do you think it would be 
structured? And also, I would ask this. You know, obviously, 
this committee is concerned with the Defense Department making 
sure we can defend ourselves with these technologies and this 
skilled workforce.
    But as you know, the Department of Homeland Security, the 
Department of Treasury--there are other government agencies 
that have the same shortfall in cyber and digital employees. 
Can you see this [inaudible] as one day being opened up to 
other agencies within the Federal Government?
    Dr. Schmidt. We foresee it immediately being available for 
all of the Federal activities, and my--I always prefer public 
partner or public partner--public-private partnership, excuse 
me--PPP--for the reasons that I think that's when America is 
strongest.
    Mr. Rogers. Great. Commissioner----
    Ms. Clyburn. Sir, if you'll allow me a quick----
    Mr. Rogers. Yes, go ahead.
    Ms. Clyburn. I'm sorry. If you'll allow me a quick bite at 
the apple. I could not go back home without affirming that the 
2-year colleges offer substantial support and opportunities, 
and they don't just offer 2-year degrees. They offer other 
certifications that will enable this, too. So I couldn't go 
back home without mentioning that.
    Mr. Rogers. Well, I'm glad you did because that's exactly 
what I was thinking about. It's a resource that we definitely 
need to be tapping into because it would help fill a need that 
we have got.
    Commissioner Louie, do you believe the private sector even 
wants to work with the Department of Defense or the government 
on these kind of national security issues involving AI and 
cyber?
    Mr. Louie. I had the fortunate opportunity of standing up 
In-Q-Tel, and there were a lot of people betting against it. 
They asked the question, why would a young entrepreneur walk 
across Sand Hill Road and ring the doorbell for In-Q-Tel or a 
Federal intelligence agency?
    Turns out that there are plenty of Americans who want to 
serve and put their technologies in a way that protects this 
country. There are some, of course, who will opt out, and 
that's the American way.
    We're not China, right. We can't--we do not compel 
companies and individuals to work for the government or give up 
their information. That sets us different--in a different and 
better place than, I think, our competitors. So----
    Mr. Rogers. Great. Thank you, Mr. Chair.
    Mr. Langevin. Very good. Thank you, Ranking Member, for 
your questions.
    And chair now recognizes Mr. Welch for 5 minutes.
    Mr. Welch. Thank you very much. We really appreciate this 
hearing and the work of the Commission.
    One of the frustrations that I'm experiencing in listening 
to this is that there appears to be a consensus--a bipartisan 
consensus on the absolute urgency of following through on the 
recommendations that you make.
    Yet, on a practical level, there's impediments from us 
doing it. Some are bureaucratic and, to some extent, 
potentially congressional.
    And I'll start with you, Dr. Schmidt. You do mention it's 
the people who are the most important. But if they're in a 
structure where the mission that they're doing is about 
overcoming obstacles that hasn't integrated that mission into 
its stated policy, how do we overcome that?
    Dr. Schmidt. Part of our recommendations, as Chairman Work 
recommended earlier, is to take the AI efforts and cause them 
to report higher inside the military, and in the military, 
because it's very hierarchical, having AI be a major component 
of strategy at every level is one of the ways that the top-down 
can work.
    I always favor a top-down and bottoms-up approach. So I'd 
like to see a DOD statement around AI that's much stronger than 
we currently have with the JAIC. I'd also like to have a higher 
reporting level, more resources, and so forth. But I'd also 
like to have individual control at the commander level and the 
COCOM level where they have flexible teams which can be used to 
solve important national security problems, which would include 
AI expertise.
    You want to get both flexibility for the commanders as well 
as priority at the highest level.
    Mr. Welch. Mr. Work, let me go to you. Thank you for that.
    What are the implications of a decision that AI is now core 
and existential to our defense versus supplemental and 
discretionary with respect to how it would affect other weapons 
systems?
    We have--as one of the witnesses said, we have these huge 
platforms that have been components, major components, of our 
defense strategy and it appears that AI cyberwarfare is really 
the biggest threat.
    So can you just elaborate on what would be involved in a 
Pentagon shift in thinking and, frankly, a congressional shift 
in thinking where implications would affect jobs in many 
members' districts?
    Secretary Work. We do not talk about this specifically in 
the report. But very broadly, Mr. Welch, we are shifting into 
an era of systems warfare. Both our adversaries--I mean, our 
adversaries explicitly say this, and say the way we will defeat 
the U.S. military is to have better operational systems, and 
the better to way--and the way to get there is to inject AI 
applications and autonomy into the system so it operates at a 
faster speed and can operate more effectively.
    So this is a big shift in thinking, going from platform-
like thinking to systems thinking, and trying to figure out how 
these applications improve.
    And in my view, all you need to have is, you know, cross-
functional teams that look at, say, our sensor group, and the 
cross-functional team says the biggest return on investment is 
to do machine learning on the sensor so that it can go through 
the information and just pass on the data that is required, 
which would make everything go faster, wouldn't clog the pipes, 
et cetera.
    And you would have someone do the same thing for our 
command, control, and communications, and intelligence grids. 
How would we have an AI-enabled application that would help 
decision making?
    So this is--this will literally affect every operation 
mission that we do, and it's going to require a different way 
of training our commanders and our people, a different way of 
educating them and a different way of training them.
    Mr. Welch. But it also requires the confidence of our 
defense leaders that a change in direction is not only 
desirable but, really, essential, that has implications on the 
way we're doing business now--I mean, the higher education 
process and political process. Care to comment in my remaining 
12 seconds?
    Secretary Work. Amen, sir. It all starts with trust. As 
soon as we have demonstrated applications enough to where 
commanders trust them, then you will see an accelerated 
adoption. But trust is absolutely key.
    Ms. Clyburn. And education starts at the top and continues 
through the ranks.
    Mr. Welch. Thank you. Thank you, Commissioner. Thank you 
all.
    Mr. Langevin. Very good. Thank you, Mr. Welch.
    Ms. Foxx is now recognized for 5 minutes.
    Ms. Foxx. Thank you, Mr. Chairman. For some reason, when I 
turn on my mic, my video goes off. But there goes the video. So 
I hope you can hear me.
    Thank you for having this hearing, and it's been very, very 
informative and very enlightening. And it's nice to see such 
bipartisan agreement.
    Mr. Schmidt, I know you're most comfortable with monopolies 
from your days at Google, but why should we create a monopoly 
in 5G?
    Why can one company in concert with the Department of 
Defense have a monopoly on 5G spectrum when having multiple 
facility-based competition is what made the U.S. 4G market so 
successful and why the app economy originated here?
    Dr. Schmidt. Our report does not specifically suggest what 
you just asked me about. But let me comment in general, that we 
have taken a look at the 5G situation and China is perhaps 10 
times ahead of us both in terms of speed as well as the number 
of towers.
    They have planned 1.3 million towers, for example, this 
year in total, and it looks like the leadership that China is 
going to have in 5G will become a significant national security 
threat to the United States, partly because 5G is used in 
autonomy, which we discuss in our report, and also because it 
will create an ecosystem of applications.
    Imagine if the key applications that are used 5 years from 
now are based on the Chinese 5G network and not on American 5G 
leadership. Over the last decade, we have, essentially, given 
up leadership in that area, having--after having done a 
fantastic job in 4G.
    My view is that it is a national security problem that we 
are not leading in 5G. I'll let you all debate the specific 
solutions to that. I would encourage you to demand a strategy 
from the various players that gets us at least equal to China 
in terms of performance and coverage.
    Ms. Foxx. Well, thank you for that. You know, it's been 
reported that the spectrum-sharing arrangement would constitute 
a boom for Alphabet and other big tech firms. This sharing 
scheme seems to be closer to the nationalization of our 
spectrum markets than anything else. To be blunt, I think all 
our constituents, but particularly mine and the ones I hear 
from, are tired of the stranglehold big tech firms have over 
the markets.
    The unchecked power of big tech's ability to corner 
markets, marginalize competitors, and silence my fellow 
Americans cannot be tolerated any further.
    With that, Congress shouldn't be empowering market 
mobilization--monopolization in this critical space. Our 5G 
infrastructure creates digital highways that artificial 
intelligence can drive on, and this Congress should not be 
risking our competitive advantage over China by investing in 
this untested and unproven nationalization framework.
    Doing so would put broadband development--rural broadband 
projects, and even national security at risk. As I said, I 
believe that there's bipartisan concern about big tech's 
antitrust problems.
    I want to----
    Dr. Schmidt. May I respond? May I respond, madam?
    Ms. Foxx. Okay.
    Dr. Schmidt. So in the first place, I no longer work at 
Alphabet, but I have a large stock position in Alphabet. 
Alphabet is not one of the beneficiaries of any of these 5G 
activities at the moment, and the argument that you're making 
about sharing is not technically correct. Your argument is, 
fundamentally, that it would be better to have a highway 
occupied by one car rather than having a highway have lots of 
different cars on it.
    The CBRS [Citizens Broadband Radio Service] option, which 
is in the 3.5 gigahertz phase, proves that sharing works. We 
need to solve our bandwidth problem. Personally, I have my own 
technical views of how to solve this. But I do not want to 
allow your statement to go unchallenged with respect to the 
statement that we're okay.
    We're not okay and we need a solution in this space that is 
competitive with China.
    Mr. Louie. Let me add, since I was one of the co-authors on 
the 5G report for the Defense Innovation Board, we are not in 
favor of monopolies. We do not believe that one company should 
own it all. I think that's, fundamentally, anti-American.
    On the other side of that equation, for us to be 
competitive in 5G we need to have large continuous blocks of 
spectrum available and not little small segments. Think of it 
like this.
    Think of going down the freeway in which you sell each lane 
to a particular company, and you cannot change your lane once 
you get on that road. Sharing, particularly in the DOD 
spectrum, is the only way to allow broader use of our spectrum 
and protect our military systems.
    Our radars, our air-to-air, surface-to-air, our satellite 
comms [communications] require sharing. Without that sharing, 
we will not be competitive. The U.S. 5G is one of the slowest 
in the world where our average throughputs are less than 50 
megabits, compared to China, which is going at 300 megabits, 
going to 1 gigabit.
    If we want to be competitive and we want AI to drive on the 
information highway, we have got to free up the lane. One car a 
lane is not a solution.
    Mr. Langevin. Very good.
    Ms. Foxx. Thank you both. I want to build on something also 
that Representative Rogers mentioned and----
    Mr. Langevin. I believe the gentlelady's time has expired, 
but briefly.
    Ms. Foxx. Oh. Well, I wanted to push certification, Mr. 
Chairman. I think we need to get alternatives to baccalaureate 
degrees and push certification where we can.
    Thank you for indulging me.
    Mr. Langevin. Certainly. I thank the gentlelady for her 
line of questions.
    Mr. Khanna is now recognized for 5 minutes.
    Mr. Khanna. Thank you, Chairman Langevin, and thank you to 
Dr. Schmidt and particularly Commissioner Clyburn, and all the 
commissioners for your excellent work.
    Before I get to my question, I do have to address what 
probably should be the headline of this hearing, Representative 
Grothman's comments that somehow that the smartest immigrants 
are going to Australia and New Zealand and we're not getting 
the smartest immigrants.
    I have a number of comments, but I guess I just for 
educational purposes, maybe Dr. Schmidt, could you comment on 
the intelligence of Sergey Brin and Satya Nadella and Eric 
Yuan? I don't see Australia and New Zealand having produced all 
these tech companies. And what--how would you characterize the, 
quote/unquote, ``intelligence'' of immigrants coming to the 
United States?
    Dr. Schmidt. America has benefitted enormously from high-
skills immigration. I have personally been the beneficiary as 
the companies that I have been working for were founded by such 
people, and they're incredibly brilliant.
    We need them to drive our tech sector. We need to create 
wealth in our stock market. We need them to pay--help pay the 
taxes for our government. I can go on and on about the quality 
of immigration.
    At one point, I was sitting at Google and I realized half 
people in the senior executives were immigrants, many them from 
South, Southern Asia.
    Mr. Khanna. You know, the other thing that I think I 
understood Congressman Grothman to be saying is that we should 
have some kind of a math and science requirement for 
immigrants.
    I think the American public would love to have a math and 
science requirement for Members of Congress. I think that would 
probably be a better start. I would challenge Congressman 
Grothman.
    Maybe you and I can take a math and science test and ask 
all of our immigrants also take a hypothetical test, and my 
guess is I wouldn't be surprised what the results are. I hope 
you'll take me up on that.
    But what do the commissioners think of the idea of having a 
math or science test for immigrants? I don't understand.
    Did Madison or Jefferson put that in the Constitution that 
we should have math scores for who should let in--we should let 
into this country? Would any of the commissioners please 
comment on that statement?
    Mr. Louie. I'm happy to answer at least part of it. The 
particular line of questioning around having science and 
technology skills for our recommended visas for anybody with a 
degree in S&T is pretty straightforward.
    I don't think any of us are advocating that we should have 
a generic test like the way you have to pass the other 
immigration examination on American citizenship or any other 
kinds of activities.
    But we need science and technology skills and that is 
clear. We are not going to be competitive without that talent. 
We need to grow them internally, we need to get them from 
outside, and we need to encourage people who are educated in 
this country of our very best universities, both public and 
private, to allow them to practice their knowledge in the 
United States and not abroad.
    Mr. Khanna. Personally, I agree with that. But just to be 
clear, you, clearly, disagree with Representative Grothman's 
suggestion that somehow we should have a math or science 
requirement for our immigrants.
    Mr. Louie. I don't believe that was the question I was 
answering.
    Mr. Khanna. I'm saying you would disagree with his 
statement on that, correct, and the Commission would? I don't 
think anyone in America, a reasonable person, thinks we should 
have requirements for immigrants to take math and science 
tests.
    Dr. Schmidt. In the excess of clarity, the report does not 
make a claim in this area, and so speaking for the Commission, 
the Commission does not take a position on this. We have said 
repeatedly high-skills immigration in our Nation is very 
important.
    Mr. Khanna. Let me just change directions, because I was 
just so struck by those comments of Representative Grothman I 
don't--I thought they had to be addressed.
    But let me ask the final question, which is on a AI, and a 
two-part question. One, it seems to me that it's in our 
country's advantage to move towards forms of AI that aren't 
relying simply on data, and when you look at some of the work 
Jeff Hawkins has done on how the human mind works in terms of 
maps of reference, and you look at Tenenbaum at MIT 
[Massachusetts Institute of Technology] and what he's being 
able to do, saying, you know, a child doesn't need thousands of 
pictures to understand what a cat is but understanding how 
categories of human perception work, and that it would be 
valuable for the United States to invest in that kind of a 
general learning AI so that we're not dependent on data because 
China will have a huge advantage over data.
    If you could address that, and then second, address what it 
means for us to have human judgment still and human control so 
that the defense decisions are not being just made by 
algorithmic AI but by people, really, still having human 
judgment over that.
    Dr. Schmidt. Quick answer to the first part. Bob, you'll 
take the second part.
    You're exactly right that, right now, these algorithms need 
an enormous amount of data. There's very promising research 
about much more limited data training models, and I think 
eventually this issue around data will become less important 
and the rise of this next generation of algorithms that you 
have suggested will be the story.
    It's crucially important that this next generation of 
algorithms get invented in the United States.
    Bob.
    Secretary Work. DARPA describes AI coming in three waves.
    The first wave was what it refers to as expert systems. 
These are physics-based models that are quite capable. But they 
have now been supplanted by second-wave systems, which are 
statistical machine learning.
    There are a lot of applications still for first-wave 
systems. So I would agree with you that there will be 
applications that just are physics-based models or rules-based 
models, if then.
    They're very, very good to explain. Like, if you had a 
safety accident, you can go back through the coding and say 
this is exactly what caused the accident, and in machine 
learning sometimes we won't know exactly why the algorithm 
chose the action that it did.
    As far as defense decisions, the clearest expression that I 
can offer you, sir, is DODD, Department of Defense Directive, 
3000.09, for example, that says, ``For weapons with autonomous 
functionalities, they will be designed and operated to maintain 
appropriate human judgment over the use of force,'' period, end 
of story. The DOD Law of War manual says you cannot transfer 
responsibility to a machine under any circumstances.
    So in my view, the Department of Defense has been very 
clear that when it comes to decisions over human life that will 
always be a human making those decisions.
    We also make a clear recommendation that we should declare 
that machines will never ever be given the authority to order a 
preemptive nuclear strike. The use of nuclear weapons should be 
off the table and we should enter in discussions with all of 
our rivals to see if everyone would agree with that.
    Mr. Louie. Let me just add----
    Mr. Khanna. Thank you. Yeah, please.
    Mr. Langevin. Briefly. The gentleman's time has expired.
    Mr. Louie. Just quickly. Look, we were very clear not only 
in 3000.09 that we must continue to comply with international 
humanitarian laws. That is what separates the U.S. from some of 
our competitors, and that humans, commanders, to be held 
accountable for the deployment of any such weapons onto the 
battlefield.
    Mr. Langevin. Yeah. And it's why--I'll just comment, that's 
why it's so important that we have international engagement and 
leadership on AI. Although we may view things that way, other 
countries that are not as friendly to us may view things and do 
things very differently. We can't let that happen. So 
international engagement is going to be critical on that topic, 
going forward.
    Mr. Khanna, thank you for your questions.
    Mr. Khanna. Thank you.
    Mr. Langevin. Mr. Moore is now recognized for 5 minutes.
    Mr. Moore. Thank you, Chairman. I am encouraged by 
Commissioner Clyburn's focus on retention. It reflects many of 
my concerns about a greater retention issue that we have across 
sort of all branches of military among our Active Duty labor 
force.
    And, admittedly, while I'm new to the committee and new to 
my time in Congress--so I'm early in my tenure--some of that is 
anecdotal, as I speak with--as I speak with fighter pilots in 
the OPSTEMPO [operations tempo] that the incentives to go to 
the private sector are great and there's--it's growing in that 
regard. And I have--I have big concerns there and I hope to use 
my time in this committee to address that.
    And so with that same thought process, I want to make sure 
that we're not going down the same path, that maybe as we talk 
about the potential of an academy, which, with bipartisan 
support, is something that I could very well be excited about, 
you know, is 5 years enough, right.
    This concept of, you know, just putting a time limit on 
what you're expected to do, is that enough, or do we--are we 
going to see some of the same issues as soon as that 5-year 
timeline kind of hits there's--because there's so much need in 
the private sector right now with AI and cybersecurity. Are we 
going to see that same retention or attrition issues there?
    So I'll direct my question to Commissioner Clyburn. 
Anybody, any of you are welcome to comment. Thank you.
    Dr. Schmidt. Can I add something ahead of Commissioner 
Clyburn?
    Mr. Moore. Please.
    Dr. Schmidt. I was really struck in my work with the 
Defense Department of how many people work there for low pay 
and in difficult conditions because they were patriotic. And 
the ones that I spoke with did not, fundamentally, leave for 
money.
    They left because the opportunity in their career was more 
interesting in the private sector, that the work that they 
wanted to do, they could not do well as Federal or military 
employees.
    That's got to get fixed. To provide leadership at the 
national level, we're going to have to have places for these 
people to serve while they're in the government, and this is 
true not just for the DOD but any aspects of the Federal 
Government.
    We stated earlier we believe these people to exist, we 
believe they're already in your employ, and we believe they're 
underutilized.
    To me, that's a big priority for this committee to think 
about. How do we create it so we keep these people rather than 
allow them to become disaffected and then leave for higher 
paying jobs.
    Commissioner Clyburn.
    Ms. Clyburn. So I will do something that's unusual and be 
brief. I believe what we can do and what we should do is ensure 
that the tools and the infrastructure inside of government for 
those who want to stay, who are willing to stay, as the 
chairman mentioned, that they have the tools needed in order to 
be productive, in order to be challenged, in order for us to 
meet our national security objectives.
    That's the biggest issue. It's the frustration inside of 
the infrastructure that needs to be fixed. So if we fix that, I 
don't think we will have as much of a retention problem. Those 
who want to run after money, they will do so.
    But there are more people who want to serve, but they are 
not going to go to work every day and get frustrated about 
advancing, being onboarded, and a lot of the other issues that 
we enumerate in this 450-plus pages of light reading.
    Mr. Louie. Let me just add 10 seconds to this level.
    In our discussions both in the AI Commission and my work 
with the DIB, it was pretty clear we had a large number of 
junior officers departing and the reason why they were 
departing was because of the lack of understanding by their 
senior officers of what these technologies can do. That 
frustration forced them to go choose employment elsewhere.
    They are committed to this country. They are committed to 
the services. But we need to educate our seniors, not just our 
junior officers.
    Mr. Moore. There's an incredible amount of bureaucracy that 
does exist in these situations. I think that actually gets to a 
lot of your points. I support that and hope to be a part of 
finding ways for that labor force to be engaged and committed 
to continuing to move forward.
    Because while it's a great thing for private economies, to 
have such good training at the government level, we do have to 
address this. So I appreciate those comments and hope to work 
with you on all of that.
    Thank you. I yield back.
    Mr. Langevin. Thank you, Mr. Moore.
    Mr. DeSaulnier is recognized for 5 minutes.
    [No response.]
    Mr. Langevin. Is Mr. DeSaulnier there? Oh, you're on mute.
    Mr. DeSaulnier. I just have to have the mental adeptness to 
remember to turn my mute off, Mr. Chairman. Thank you.
    I just--I was saying that I want to concur with my 
colleague from the Bay Area, Mr. Khanna's comments about 
immigration, the best and the brightest and testing.
    Clearly, we have seen it here in my district and I know he 
has there in terms of the importance of what Dr. Schmidt said 
of attracting and having attracted so many good people to this 
field and others.
    I want to switch a little bit to--or a lot--to bots. I have 
had a fair amount of success around this issue in legislation, 
and some of the comments in your report about AI and botnets.
    There's a quote in the report from the Justice Department: 
``Control of hundreds of thousands or even millions of 
computers to advance their schemes using AI and botnets.''
    We know from our investigation of the last elections and 
the Mueller report how these were used. We know how they're 
used domestically and for foreign.
    So the Commission, again, says in its report that this may 
become more powerful with advanced AI, not just directly 
spreading malware but harvesting both computerization power and 
data to put forward further offensive training in ways that 
were not previously possible.
    Now, Mr. Louie, can you talk about this threat to our 
infrastructure--transportation, utilities, health, and 
financial--please?
    Mr. Louie. This is a real threat to our critical 
infrastructure flow. As problematic as botnets are, AI-driven 
botnets are operating at machine speeds against our defenses, 
which have, still, people in the loop rather than on the loop. 
It's a losing strategy and it's clear to us that our 
adversaries are using not only advanced technologies like 
botnets, they're going after supply chain as with the 
SolarWinds attacks and they're also using it in disinformation 
campaigns.
    And the next step is to use disinformation campaigns 
against machines, to give machines and feed machines 
disinformation to make machines do the wrong thing. That 
adversarial AI is a real threat.
    Mr. DeSaulnier. And, Mr. Louie, just following up, there is 
a section that the commissioner recommends us passing in 
Section 4 of the International Cyber Crime Prevention Act.
    To the degree you're capable, could you talk on my comments 
or the ability of that to help stem this threat--pardon the--
pardon the choice of words--and what might we do in addition to 
this?
    Mr. Louie. It's clear we have to defend forward, right. 
Playing a defensive posture waiting for the attack and trying 
to build a higher wall or a deeper moat is a failed strategy.
    Second is that we have to work with our allies. If there 
are criminal activities, we need to raise the costs of those 
activities to those individuals or the nation-states that are 
prosecuting the attack against us.
    Right now, the cost to attack is so low, the consequences 
are almost nonexistent, that we are inviting attack after 
attack after attack, and we will not stem it by simply having a 
higher wall of cyber defense.
    Mr. DeSaulnier. Thank you.
    Dr. Schmidt. Could I add----
    Mr. DeSaulnier. Sure.
    Dr. Schmidt. Could I add, Representative? The traditional 
framing of this is the Russia election attacks of 2016, which 
were done by humans, not by computers, as best we could tell.
    There's every reason to think that not only will a country 
like Russia try to do this, but that many nonaligned groups--
terrorist groups, unrelated groups--that are trying to disrupt 
the democratic processes of the democratic countries for 
whatever reason--economic, financial, political, just evil--
because the technologies are now so broadly available.
    So one of the things we talk about in the attack--in the 
book is that the software diffusion, right, the ability for 
people to access this is now--is the cat is out of the bag, 
whatever metaphor you care about.
    And we have got to get ourselves organized around the fact 
that there will be continuous attacks on our information space, 
which you are describing as bots. But they're really much more 
than that.
    It's attacks in terms of the quality of information, the 
target of information, attacks on the individuals involved, 
misusing their personal information, and this is also a 
national security issue coming forward.
    Mr. DeSaulnier. And I would add in the sophistication of 
neuroscience and targeting, not just to individuals and 
demographics, but as you know.
    Dr. Schmidt, if I could just--I want to take this 
opportunity because you're here, as somebody from the Bay Area 
who is very proud of our tech industry but recently has become 
critical and chagrined, in the sense of national security we 
often hear about scale. That's important to have scale, not 
just in tech but in finance we've heard it.
    So could you comment on that, just briefly, the importance 
of scale with some of the challenges we have had in regards to 
concentration, to Ms. Foxx's comments?
    Dr. Schmidt. I know here there is a concern about this. I 
will tell you that I would like to win the global competition 
and have America win in global competition.
    That is going to require large companies because of the 
scale issues, the economic issues, the number of people. The 
projects and products that we talk about in these economies--in 
these companies take thousands and thousands of people and 
many, many hundreds of millions of dollars to build.
    Those are problems of scale. We have an incredibly vibrant 
and diverse venture ecosystem with an awful lot of new 
startups, huge valuations. It's all incredibly exciting.
    My strong--and this is my recommendation and has nothing to 
do with the AI report--my strong recommendation is if you don't 
like what one of the big tech companies are doing, find a way 
to regulate their behavior through the normal mechanisms.
    I'm sure there are issues that need to be regulated. I 
would do it that way.
    Mr. DeSaulnier. Mr. Chairman, thank you. I'd love to follow 
up with you, Dr. Schmidt, because we want to get the right 
balance in here.
    Thank you, Mr. Chairman, for your indulgence.
    Mr. Langevin. Thank you, Mr. DeSaulnier.
    Now the chair recognizes Mr. Higgins for 5 minutes.
    Mr. Higgins. Thank you, Mr. Chairman. The timing and 
purpose of this joint hearing is crucial, as we discussed last 
month in our joint hearing with the House Homeland Security 
Committee.
    Our foreign adversaries have elevated the battlefields into 
the cyber realm, artificial intelligence, and let us not--let 
us not fail to observe the Chinese tremendous advancements in 
quantum technologies, which I believe is inseparable from the 
conversation regarding artificial intelligence.
    And I would like to ask the panel and want to get to my 
question rather quickly because I expect the answer to be 
complex. The theft of our technologies as of last year--as 
recent as last year the FBI [Federal Bureau of Investigation] 
advised that they investigated more than a thousand cases of 
Chinese theft of U.S. technology.
    That's inside our universities and our government research 
and development laboratories, and I believe--and I would like 
the panelists to consider how we take further action to protect 
our technologies.
    The report that we're discussing today is incredible work. 
I believe it'll be recognized as very significant work that you 
ladies and gentlemen have done, and we thank you for it.
    It's going to take us a while to get our head wrapped 
around this. A line struck me from your report. You stated that 
for the first time since World War II, America's technological 
predominance, the backbone of its economic and military power, 
is under threat, and that's a quote.
    I asked Dr. Schmidt to reflect upon that statement and tell 
America how the Federal Government and the private sector can 
work together to train the next generation of patriots, both 
civilian and military, to protect our research from theft and 
to gain dominance in these fields.
    And as related to that question, then I'm going to turn it 
to you, Dr. Schmidt, until recently, every nation in our--in 
our military academies we teach that each nation enjoys a 
certain degree of elements of power, that being the military, 
geographic, economic, cultural, and political.
    It seems to me and many of us that a new element of power 
must be considered as we balance our own strengths against that 
of the world.
    That would be artificial intelligence and the quantum era. 
And it seems to me that China appears to be leading the world 
in artificial intelligence theft and quantum technology.
    Dr. Schmidt, will you address that, please? Based upon your 
background, I believe you can give us a solid answer or at 
least guide us.
    Dr. Schmidt. Thank you. Two years ago, China announced a 
strategy with a goal to dominate the following industries: 
software, AI, semiconductors, energy, robotics, and high-speed 
transportation and biotech and quantum.
    Well, that's my whole world. That's everything I care 
about. It's furthermore everything that is driving the 
renaissance in America in American manufacturing, American 
leadership, American global platforms.
    We lack a strategy as a country to work in those areas. We 
must organize it.
    Mr. Higgins. Exactly.
    Dr. Schmidt. We must, must, must. It's a huge issue. You 
highlighted the quantum issue. It should upset us that quantum 
leadership in China is ahead of America in certain aspects of 
the quantum work. We need to get our act together.
    In order to do that, we need to do a number of things. The 
first is we need to work on our own workforce, our own STEM 
education, as we highlighted. We also need to recognize that we 
are critically dependent upon foreign researchers and foreign 
graduate students.
    One of the things we did in the AI report is we studied 
where were the top researchers coming from, and many of the top 
researchers are, in fact, graduate students who are coming from 
China, who are learning and researching in our universities.
    So we're empathetic that we have to keep them coming and 
keep them in the country and keep working on these things to 
help our Nation. Those are the best solutions that I have for 
you.
    There's, obviously, a concern about intellectual property 
theft. To the degree that it occurs and we know it occurs, it 
should be prosecuted to the fullest possibility of the law.
    And you can imagine that for Chinese students, for example, 
you could do investigations in that matter. You could also have 
all sorts of other ways about validating.
    But we need these people in America working on these hard 
problems and we need a national strategy to win.
    Mr. Louie. And we need to have--give the tools to our 
universities and research centers and companies to protect 
themselves. Well, it's really hard to protect yourself against 
a nation-state.
    We tend to look at cybersecurity breaches as an IT 
[information technology] problem. Our adversaries look at it as 
a domain of warfare, and you're not going to win in that 
strategy. We need to better share information, create useful 
interchanges between enterprises, academia, research centers, 
and the government.
    We can protect not everything but we can protect those 
things that are most vital to us. What's the point of leading 
in research if the other guy can just simply steal it?
    We have to put up the appropriate protections. We need to 
reform our IP laws and we need to have partnerships with our 
allies to make it expensive for those acts to continue to go 
unchecked.
    Ms. Clyburn. And, sir, we need to give the tools to law 
enforcement in order to recognize and take action.
    Mr. Higgins. I thank the chairman and the panelists.
    Mr. Chairman, God bless you for allowing us some indulgence 
with time, and I yield, good sir.
    Mr. Langevin. Very good. Thank you. Mr. Higgins.
    Mr. Kim is now recognized for 5 minutes.
    Mr. Kim. Thank you, Chairman. Thank you everybody for 
coming together here.
    And I've had the chance to be able to go through the report 
at length and then sat down with your staff. So I've gone 
through a lot of that and feel very good about the 
recommendations and I'm grateful for your work on that.
    But I wanted us to kind of take a step back here. As we're 
looking at something that would be potentially just a major 
undertaking in terms of structuring our national security, our 
defense innovation efforts, and potentially billions of dollars 
to be able to jettison, you know, some of the old standing ways 
in which we have been doing this, the question that I struggle 
with is, how do we best explain this to the American people?
    You know, how do we talk about this complexity in a human 
way to people in my district? And, you know, so I want to just 
ask Dr. Schmidt and then Secretary Work, when it comes to 
clearly articulating in a very understandable way how the 
threat will manifest and what people can understand and wrap 
their heads around, both in terms of the threats and the 
opportunities.
    I would appreciate your perspective on that because I think 
people in my district, they understand the threats of 
transnational terrorism. They understand the threats of 
conventional warfare, that kind of way. It's something that is 
visible and tangible to them.
    But they struggle, and, frankly, I struggle and others 
struggle to really understand what exactly are we talking about 
in terms of a threat here?
    Dr. Schmidt. Well, I always start from the standpoint of 
America as a place of great freedom and our values, and I am 
concerned that if we lose leadership in this area, the 
information freedom we have, the free speech we have, all of 
the things that have made us as a great country will be 
materially affected by those changes.
    And the way that would occur is because AI is, 
fundamentally, software and software can be deployed to change 
the way you perceive the world, as we have discussed in the 
testimony so far.
    There are plenty of military comments that Bob should make. 
But I think the impact on our society when a targeted opponent 
comes into our networks and our information space and begins to 
screw around could lead to a real increase in distrust in our 
government, a lack of patriotism, and a lack of belief in our 
country.
    Secretary Work. I would have started exactly where Chairman 
Schmidt did. A technological competition is a values 
competition at its core. The way these applications will be 
used will reflect the governance system of the country that is 
pursuing them.
    For our American citizens, all we have to do is say look at 
how these technologies are being used in China: population 
surveillance, lack of privacy, lack of civil liberties, 
minority suppression.
    We do not want a world in which these values are reflected 
through technology and the infrastructures that support them. 
And it is important for the United States, as the greatest 
democracy in the world, to apply these applications in a way 
that are consistent with privacy, civil liberties, and law.
    Depending on who you listen to, either a McKinsey or a BCG 
[Boston Consulting Group], the winner of the AI competition 
will accrue a $13 to $15 trillion economic advantage. They also 
say same things for, like, 5G. 5G might be $5 to $7 billion--
trillion, excuse me.
    So these technologies will affect our lives in ways that 
will make our citizens healthier, live longer, have better 
lives, be able to do their work better, be able to have just 
new ways of entertainment.
    This is a technology competition that is very important for 
us to win.
    Mr. Kim. Yeah. Well, thank you.
    Ms. Clyburn. Sir.
    Mr. Kim. Go ahead.
    Ms. Clyburn. I'm sorry. Sir, when I got my first coupon in 
the checkout line, I was excited. Oh, my gosh, I'm saving money 
here. Then I started noticing my social interactions online and 
seeing these ads pop up, and then I started wondering and 
learning more about algorithms.
    Then I started figuring certain things out, that my digital 
footprint, my information, my pattern, can not only be 
monetized but can be used against me.
    So when it comes to what we're speaking of today and 
explaining to everyday--your constituents, everyday people who 
are trying to save money and trying to make their lives 
easier--that convenience, used in the wrong way in the wrong 
hands, could make us the most vulnerable people ever, and 
that's why this is important.
    Mr. Kim. Thank you. Well, look, I'd like to build out that 
story and that narrative with you all. So let's keep working on 
that.
    Chairman, I yield back.
    Mr. Langevin. Thank you, Mr. Kim.
    Mr. Johnson is now recognized for 5 minutes.
    Mr. Johnson of Georgia. Johnson from Georgia?
    Mr. Langevin. That's you, sir.
    Mr. Johnson of Georgia. All right, thank you. I want to 
thank you, Mr. Chairman, and also Chairman Lynch for holding 
this hearing.
    China has embarked on what the Commission described in its 
report as a, quote, ``multi-pronged campaign of licit and 
illicit transfer--technology transfer. In effect,'' the 
Commission argues, ``China is using American taxpayers' dollars 
to fund its military and economic modernization,'' end quote.
    By some estimates, China's technology theft costs the 
United States between $300 billion and $600 billion a year. One 
of the several ways that China is seeking to gain a competitive 
edge is through venture capital investments in U.S.-based AI 
startups.
    In response, the Commission recommends in its final report 
that Congress require investors from U.S. competitors to 
disclose transactions in a broader set of sensitive 
technologies to the Committee on Foreign Investment in the 
United States.
    Mr. Louie, how broad is Chinese investment in U.S.-based AI 
startups?
    Mr. Louie. The Chinese are--both companies as well as 
regional and university organizations are active throughout the 
entire United States doing investments.
    Some of it's benign. Some of it's just because they want to 
make money. Some of it's suspect. But here's our challenge. 
CFIUS still remains fundamentally a voluntary series of 
regulations.
    We need to make sure that it is no longer voluntary. If 
you're taking money from potential adversaries or competitors, 
like China and Russia, it needs to be disclosed.
    Second, we have to make--we have been waiting for years now 
in the technology community for the list of the critical 
technologies that will be deemed critical for the United 
States.
    We still have not produced that list. Technology companies 
are guessing on whether or not something requires disclosure or 
not. We need to make it clear that these kinds of technologies 
like AI, like microelectronics, like quantum computing, any of 
these critical--biotechnologies--any of these critical areas, 
if you're taking direct foreign investment or indirect 
investments from these nation-states, they need to be 
disclosed.
    Most will be fine. But a few of them may not be. But we 
need to know and companies have a responsibility to disclose 
that.
    Mr. Johnson of Georgia. Thank you.
    What regulatory framework for disclosure of transactions is 
currently in place for venture capital investments and how 
would requiring disclosure to CFIUS help to protect sensitive 
AI technologies?
    Mr. Louie. It's hard to protect what you don't know it's 
happening. So the first one is just bring their knowledge up.
    Second is for those critical pieces of technology that we 
deemed as critical, we need to make sure that State is 
empowered and has the skill sets to review those technologies 
aggressively.
    You have this thing called a short form and a long form 
that you have to fill out, right? There's a lot of paperwork in 
the long form, and the short form is a little bit easier.
    I think we should take a look at the filing requirements, 
particularly for venture capital, and to make sure that people 
are well educated in the regulatory regime.
    The good news is most tech companies and early startups use 
very competent law firms. Having the law firms be a partner in 
this matter is going to be really critical for us to be able to 
not only protect the technologies, but quite frankly, protect 
those entrepreneurs' technologies.
    The last thing an entrepreneur wants to do is invest it, 
take some technology that they put their hearts and lives in 
and have a competitor overseas suddenly show up with that 
technology and compete against them.
    Mr. Johnson of Georgia. Thank you. Would any of the other 
panelists like to comment?
    [No response.]
    Mr. Johnson of Georgia. Okay. In a July 2020 report, the 
Center for Security and Emerging Technology at Georgetown found 
that out of 208 global Chinese professional associations, or 
CPAs, more than half advertised on their websites that they, 
quote, ``exchange technical information, bring scientists to 
China, or contributed to specific Chinese talent plans,'' end 
quote.
    Interestingly, however, the report also found that CPAs 
that advertise the transfer of technology in Chinese also are 
more likely to omit this information about that aspect of their 
missions from the English-language versions of their websites.
    Mr. Louie, why might CPAs hide this information from 
English-speaking members?
    Mr. Langevin. If you could answer briefly, please, Mr. 
Louie, because the gentleman's time has expired.
    Mr. Louie. I think it's difficult to read people's minds. 
But I would say that they are, clearly, attempting to encourage 
American companies to put technologies overseas.
    And my biggest point is the relationship between technology 
transfer from U.S. to China and China to U.S. is asymmetric. 
For Americans to invest in Chinese companies require huge 
amounts of regulatory--Chinese regulatory hurdles that you have 
to come, whereas Chinese investments in the U.S. has almost 
none, and we've got to fix that asymmetry. It doesn't help us 
on either--in either direction.
    Mr. Johnson of Georgia. Thank you. I yield back.
    Mr. Langevin. Thank you, Mr. Johnson.
    Ms. Houlahan is now recognized for 5 minutes.
    Ms. Houlahan. Thank you, Chairman, and I hope you all can 
hear me. I have a lot of questions and so I'll try and move 
through them quickly.
    My first set of questions is for Chairman Schmidt. I really 
appreciate the commissioners'--the Commission's conclusion and 
recommendations that help us find and build a full-time cadre 
for public servants with digital experience, and our success 
and failure in this definitely will hinge on the Federal 
Government's ability to compete with the private sector. And 
I'm really enthusiastic about the idea and concept of a Digital 
Service Academy.
    But I'd like to hear more about your cost estimates for 
that, time and other resources, what it takes to get this kind 
of an undertaking off the ground, and could you also talk us 
through your cost-benefit analysis, if there is one, of 
alternatives that might be less expensive like, perhaps, 
dramatically scaling up STEM-focused ROTC programs such as the 
one that I participated in?
    Dr. Schmidt. Perhaps Commissioner Clyburn would like to 
help me. These--it's a false equivalency to say that these 
somehow are related to the military and ROTC activities.
    We strongly support the military and ROTC activities, and 
they are phenomenal and we should invest more in them, 
especially with respect to specialized skills--the digital 
skills.
    In addition, the civilian workforce needs upgrading and the 
two are separate. The economics around the civilian workforce 
one are pretty straightforward.
    The cost of hiring the people, the cost of paying them, and 
so forth and so on, is much less expensive if they're going 
through a 4-year program which has been subsidized to some 
degree by the government and where they have a 5-year 
commitment to work.
    So the economics actually work.
    Mignon.
    Ms. Clyburn. Yes, I will have to get back in touch with you 
in terms of the actual amount of shoring up a full academy. But 
the per student--the last figure I remember was a $50,000 per, 
you know, life cycle in terms of the actual expenditure per 
student.
    But my answer to you is what's the cost of not doing 
anything? That cost is--cannot be quantified. It is a negative 
and a burden on our system, and we must address this 
immediately.
    Ms. Houlahan. Yeah, I completely agree with that. In fact, 
that's kind of part of my follow-on question, which is that I 
believe, Chairman Schmidt, you mentioned something about the 
universities having offered to help get the DSA, the Digital 
Service Academy, stood up.
    Could you maybe explain how that might happen, how they 
might plan to do that? And how do universities foresee being 
able to be supportive of this kind of initiative?
    Secretary Work. Well, ma'am, just to follow up, Appendix E 
in the report outlines what we believe are the recommended 
investments and all of our recommendations. Our estimate for 
the Digital Service Academy is that a $40 million initial 
investment would get us on the way.
    We did not make a calculation on how many--how much per 
year. But, for example, the STEM Corps or the Digital Corps in 
the Department of Defense, we thought $5 million in FY 2022 and 
$5 million in FY 2023 would allow you to essentially start the 
framework, and the National Reserve Digital Corps, managed 
through the Office of Management and Budget, about $16 million.
    Now, these----
    Ms. Houlahan. And that--and that actually leads, and I'm 
sorry to interrupt. I'm just trying to make sure that I ask, 
you know, probative questions that help kind of get us to where 
we need to be, which is, I believe, in support of these kinds 
of ideas.
    This last question is for Commissioner Clyburn and it has 
to do with the Digital Corps that we're just talking about.
    I'm also really interested in the idea of a Reserve 
Component. I also was a reservist myself for many years and was 
never, frankly, called upon. I'm an engineer.
    And I was wondering if you might, you know, kind of think 
about how to imagine the opportunity for people to participate 
in that while coming in and out of private--the private sector. 
How do you recommend the idea and how do we address potential 
conflicts with that?
    Ms. Clyburn. Your last thing was potential--well, let me 
say that the benefits are many. These individuals could come 
in, triage, help, assist, augment, at critical points in the 
cycle. If they are dedicated to a particular agency, it could 
make a world of difference.
    And I struggled over the last part of your question.
    Ms. Houlahan. Just the potential conflicts. You know, if 
you're coming in and out of the private sector, maybe even the 
defense industrial sector, and you are coming in and out as a 
reservist into this Digital Corps concept, what kind of 
conflicts could we foresee and how would we be able to address 
them?
    Ms. Clyburn. Well, Dr. Schmidt might be able to speak on 
that head on. But there would not be a--there should be a 
vetting process. There should be, I believe, you know, ways to 
address that, at least initially, in terms of the skills--the 
digital skills and the opportunity to enhance that three P or 
P3 partnership is worth some of the risks, I believe, moving 
forward.
    Ms. Houlahan. Thank you. And I apologize. I've run out of 
time so I will yield back, Chairman.
    Dr. Schmidt. Just as a comment, Congresswoman, your service 
and in the Reserves was not called on because undoubtedly they 
didn't know how to find you and why you were so valuable.
    So there is a demand-side problem where the government 
doesn't know that you're available and they can't take 
advantage of your skills. We have got to get that fixed.
    Ms. Houlahan. Concur. Thank you, sir.
    Mr. Langevin. Yeah, completely concur. That was a great 
point to raise. Very good.
    Well, this has been a great discussion. Are there any 
members that have not asked a question that would like to 
recognized?
    [No response.]
    Mr. Langevin. Okay. Hearing none, I just want to thank our 
witnesses for your testimony today. This has been invaluable.
    The report that you've produced is going to be both 
foundational and enduring, I have no doubt, as we confront the 
challenges and opportunities presented before us in harnessing 
the power of AI, and I know it will be very informative for 
Members of Congress and staff as we draw upon your expertise 
and all the time and the effort that went into your hearings 
and putting the report together.
    Before I close out the hearing, I just wanted to yield to 
the ranking member in case you had any final thoughts or 
comments, Elise.
    Ms. Stefanik. Thank you so much, Chairman Langevin, and 
thank you, thank you, to the commissioners for their tremendous 
work. It is going to help guide us in the future for an 
effective whole-of-government approach.
    I also appreciate your focus on the workforce challenge 
that is so urgent that lies in front of us. There is 
significant bipartisan interest in tackling this workforce 
issue with alacrity.
    So thank you for the great work. We look forward to 
integrating many of your recommendations through the National 
Defense Authorization Act this year, as we did in the last 
Congress, working on a bipartisan basis with Jim and our 
colleagues on the subcommittee and full committee.
    And thanks for coming in today and dedicating so much time. 
I yield back.
    Mr. Langevin. Thank you. Yeah, well said.
    Dr. Schmidt. And, Mr. Chairman, on behalf of the 
Commission, I would like to say, once again, that it has been a 
privilege and an honor for the commissioners and Commission to 
serve you.
    We remain ready and able and willing to work on this to 
make sure that we get to the great outcome for America.
    Thank you so much.
    Mr. Langevin. Thank you.
    Secretary Work. And, Mr. Chairman, if you'd allow me----
    Mr. Langevin. Yes, go ahead, Mr. Grothman. Go ahead, 
Ranking Member Grothman, go ahead.
    Mr. Grothman. Right. I'd just like to thank you as well. It 
is a very important issue. It's an issue we cannot afford to 
fail on and can't afford to lose on. I hope this committee has 
other hearings on this topic as time goes on.
    I appreciate all the work that went into report. the 
report, and thank you for coming to Washington today.
    Mr. Langevin. Excellent.
    Secretary Work. Mr. Chairman, I just wanted to add that we 
can provide the committees with a classified briefing to give a 
fuller picture of how China and Russia are approaching this 
competition, and I feel like I'm probably as immersed in this 
as much as anyone and there are things in the intelligence 
record that, quite frankly, surprised me very much.
    So we stand ready to come over and give that briefing to 
either the committee or to members of the committee or however 
you would like to see it.
    Mr. Langevin. Very good. Yeah, and thank you, Secretary 
Work, for raising that point and we will certainly take you up 
on that. I look forward to a classified session where we can 
get into some of those more sensitive details.
    But, again, on behalf of all of my colleagues on both 
committees, I know Mr. Lynch had to leave but on his behalf as 
well, we just want to thank you for your extraordinary 
contributions to this area of artificial intelligence.
    As I said, it will be a foundational document and very 
instructive in helping us and guiding us as we develop policies 
and legislation, going forward, to maximize the opportunities 
of AI.
    So thank you all very much. I look forward to staying in 
touch. With that, I thank the members for their questions and, 
with that, this hearing stands adjourned.
    I thank staff for all their hard work in putting this 
together, too. Thank you.
    Hearing is adjourned.
    [Whereupon, at 1:35 p.m., the subcommittees were 
adjourned.]
    
=======================================================================

                            A P P E N D I X

                             March 12, 2021
     
=======================================================================


              PREPARED STATEMENTS SUBMITTED FOR THE RECORD

                             March 12, 2021

=======================================================================

[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]

      
=======================================================================


              QUESTIONS SUBMITTED BY MEMBERS POST HEARING

                             March 12, 2021

=======================================================================

      

                   QUESTIONS SUBMITTED BY MR. MOULTON

    Mr. Moulton. Dr. Schmidt, your report cites the Future of Defense 
Task Force report, which I co-led last year. Both our reports agree 
that the Department of Defense has a long way to go before it can 
achieve competitive advantage in the field of AI, and both reports 
recognize that achieving that competitive advantage will require 
significant investment from the Department and U.S. Government writ-
large. The Department pays a lot of lip service to technologies of the 
future like AI, but continues to pour billions of dollars into 
platforms of the past. Do you believe the Department can achieve its 
modernization goals, and the goals of your report, without reallocating 
investments that currently go to legacy systems, platforms, and 
concepts of operation?
    Dr. Schmidt. The AI Commission's Final Report does not identify 
specific systems, platforms, or concepts that the Department should 
retire, or suggest specific areas where the Department should reduce 
its investments. Instead, the report outlines how the Department should 
approach the development and management of competitive technology 
investments. As a first step, DOD should produce a Technology Annex to 
the National Defense Strategy (NDS). This annex should identify and 
prioritize emerging technologies and applications that can enable the 
capabilities and concepts that will be required to solve the 
operational challenges outlined in the NDS. The annex should be more 
than a simple list of technologies; it should be an integrated strategy 
that marshals resource allocation, R&D, and acquisition processes 
toward common ends. The development of such an annex should be overseen 
by a Tri-Chair Steering Committee on Emerging Technology, led by the 
Deputy Secretary of Defense, the Principal Deputy Director for National 
Intelligence, and the Vice Chairman of the Joint Chiefs of Staff, as 
our report recommends. In addition, DOD should institutionalize a 
Department-wide, enduring review and decision-making process to divest 
from legacy systems. Priority threats and challenges should guide 
decisions to divest or reallocate funds. The U.S. military needs to 
integrate AI across its missions. Our report argues that if it is too 
costly or ineffective to equip a platform or system with AI--or to make 
it compatible with AI-enabled systems--then DOD should divest from that 
platform or system.
    Our report also notes that there is an obvious use case for AI 
technologies to be leveraged as decision support tools in conducting 
this kind of analysis. In particular, AI should aid the Department in 
weighing data to compare the risk/reward tradeoffs between new versus 
old technologies and operating concepts. Finally, the report argues 
that efforts to develop and field next generation capabilities must be 
adequately resourced. We recommend that DOD should commit 3.4% of its 
annual budget to science and technology, which aligns with 
recommendations from the Future of Defense Task Force and the Defense 
Science Board. Based on our analysis of DOD spending on AI to date, we 
also recommend allocating $8 billion for R&D of core AI annually.
    Mr. Moulton. Mr. Work, your report repeatedly points to the 
benefits of international AI collaboration, both for the Department of 
Defense and U.S. Government writ large. I strongly agree that we should 
push for increased collaboration, for all the reasons you listed in the 
report and more: we need our technology to integrate with our allies' 
technologies, we should be able to leverage the research and 
development of our allies, and most importantly we need to set the tone 
for ethical and responsible global use of these technologies. At the 
same time, the Department has a reasonable instinct to protect its 
information and technologies, which sometimes creates a roadblock for 
collaboration. How can the Department better balance its 
responsibilities to protect certain information and technologies with 
this clear need for increased international cooperation and 
collaboration?
    Secretary Work. Extending U.S. and allied technology advantages 
requires establishing an effective protection regime that safeguards 
sensitive technologies and preserves the integrity of our research and 
commercial environment. It also requires cultivating international 
collaboration to accelerate innovation. We cannot do just one or the 
other; we must do both. The AI Commission's Final Report argues that 
the United States should take a judicious approach to export controls 
for emerging technologies. This approach would focus primarily on 
discrete chokepoints that the United States and our allies control and 
that have substantial downstream effects on technology development. 
High-end semiconductor manufacturing equipment, such as EUV and ArF 
immersion photolithography equipment, is a clear example of such a 
technology. We must also work with our allies to strengthen our 
collective ability to protect the integrity of our innovation 
environment. The report recommends building a coalition committed to 
research integrity, and sidelining those who do not abide by the values 
that underpin innovation and global science cooperation. It also 
recommends sharpening the focus of CFIUS and research protection 
efforts to address the concerning actions of U.S. competitors, while 
working to exempt trusted allies and facilitate closer cooperation. 
Additionally, the United States must work to strengthen allied 
capabilities on technology protection, particularly as we further 
integrate our national security innovation bases with our partners in 
Europe and Asia. Our collective technology protection efforts will only 
be as strong as our weakest link. Finally, the report recommends 
collaborating with like-minded allies and partners, through an Emerging 
Technology Coalition, in a number of concrete areas that would enable 
the U.S. government to increase collaboration that protects innovation 
and research. For example, the United States, with allies and partners, 
should work to advance R&D on privacy-preserving machine learning or 
develop data-sharing best practices. Likewise, we see the Coalition as 
a forum for nations to explore alignment on regulatory mechanisms to 
protect innovation--such as export controls, intellectual property, and 
trade. Collaboration in the context of military and intelligence 
activities presents a different set of challenges. In the NATO context, 
for example, differential technology adoption and expertise across the 
alliance present challenges to AI interoperability, which is 
fundamental to joint operations. The Commission has called for DOD to 
prioritize its efforts to accelerate adoption of AI and other emerging 
technologies at NATO. As this work is done, it is critical that AI 
systems are developed and adopted in a responsible manner. That means 
that they must comply with the rule of law and ethical principles, and 
that systems should be secure, reliable, and trustworthy.
    Mr. Moulton. Dr. Louie, this report makes it clear that both the 
government and private sector have important roles to play in the 
development and deployment of AI. However, balancing those roles is no 
easy task. How should the Department of Defense balance its reliance on 
private-sector capabilities with its efforts to build internal 
capability, both in terms of personnel training and infrastructure? 
What elements of AI development and deployment should the Department 
work to master internally, and for what elements should the Department 
expect to rely on the private sector, both in the near and long term?
    Mr. Louie. The Department of Defense (DOD) needs to shift from 
concepts of ``reliance'' and ``dependencies'' and to those of 
``partnerships'' and ``mutual interests.'' The U.S. industrial 
technology base is a significant source of this country's competitive 
advantage and innovation. The DOD should not duplicate U.S. private 
sector and industrial efforts. Instead, the federal government should 
use its buying power as this country's largest purchaser of 
capabilities and advanced technologies to speed the development, 
adoption, and deployment of cutting-edge technologies internally. This 
will drive transformation and deliver new and powerful capabilities to 
the DOD. The DOD should increase public-private partnerships with 
academia and industry to encourage new R&D efforts to benefit those 
institutions and the U.S. national security community mutually. To 
achieve the Commission's goal of an AI Ready DOD by 2025, the DOD 
requires a core group of people with deep technical expertise, as well 
as baseline digital literacy, across much of its workforce to 
understand how to use new emerging technologies effectively. Even 
military personnel who do not have deep technical expertise must 
develop core competencies in building, using, and responsibly teaming 
with machines. At a minimum, ensuring these competencies are present 
and prevalent throughout the force will help DOD personnel be better 
consumers of contractor-developed applications and technologies. 
However, it will also ensure our military can develop and update 
applications real-time to solve mission-specific challenges where it's 
most critical, such as the tactical edge. The way we manage the careers 
of our military personnel must change. The most significant hurdle to 
developing this technical expertise is the lack of career fields in 
software development, data science, and artificial intelligence. We 
must incentivize digital literacy like we incentivize joint 
warfighting, through a system of critical billets and an emerging 
technology certification process. Our acquisition workforce must also 
build digital literacy. Acquisition professionals must have sufficient 
understanding of digital and emerging technologies in order to 
thoughtfully apply the full breadth of acquisition pathways and 
contracting approaches. Our report recognizes that there are a number 
of acquisition workforce training initiatives underway across the 
Department related to digital and emerging technologies. These 
initiatives should be coordinated for maximum impact. It is also 
critical that acquisition personnel have common access to digital 
technology courses available across the enterprise, as well as best 
practices and a community of experts that illustrate how different 
acquisition and contracting approaches can be used to deliver best of 
breed technologies. It is important to note that our recommendations do 
not necessarily seek to promote internal development over private 
sector-led development. The private sector will continue to be a 
critical partner for the Department in the development and deployment 
of AI. To better leverage the limited STEM resources and talent 
available in the United States and increase the rate of adoption of 
critical new technologies, the DOD should purchase and implement 
commercial off-the-shelf (COTS) whenever possible, and only invest in 
government-only-solutions (GOTS) in areas that COTS cannot address. 
Applications already proven in the commercial sector can generate labor 
and cost savings, speed administrative actions, and inform decision-
making with superior insights if adopted by the Department. Off-the-
shelf technologies will be particularly useful to optimize core 
business and administrative processes, as well as logistics and 
sustainment systems. Other applications, including those fielded in the 
operational environment, will satisfy unique needs and use-cases of the 
Department. Using this approach would help strengthen our industrial 
base, focus our government spending and talent on applications that 
will give us unique capabilities, and a competitive advantage over our 
competitors and adversaries.
    There are several steps outlined in our report that are critical to 
accelerating AI adoption in DOD. Many of these same steps can lay the 
foundation for even more productive collaboration with the private 
sector. They include:
      Building an integrated technology scouting program that 
mobilizes a community of practice from across the DOD, the IC, federal, 
private, and international partners to constantly monitor emerging 
technology efforts across industry, the USG, adversaries, and allies.
      Based on inputs from the technology scouting program, 
producing a Technology Annex to the National Defense Strategy that 
identifies, prioritizes, and resources emerging technologies and 
applications, including AI applications, that solve the most critical 
operational challenges and drive new concepts of operations.
      Communicating technology priorities to the private sector 
by publicly publishing an unclassified complement to the Technology 
Annex that identifies specific operational challenges and capability 
gaps that the private sector could help solve.
      Establishing a common digital ecosystem to provide the 
technical foundation for ubiquitous AI development and fielding and 
provide access to AI software, trained models, data, and compute, as 
well as development environments.
                                 ______
                                 
                    QUESTIONS SUBMITTED BY MRS. BICE
    Mrs. Bice. We know that investments in STEM education are needed to 
maintain our competitive advantage over our nation's adversaries in AI 
technology. What additional actions do you think the private sector 
could take to aid in growing a skilled workforce that aren't already 
being done? Are there incentives that Congress should consider 
providing to defense contractors to encourage additional engagement 
with students in the STEM fields?
    The Commissioners. The United States needs to make significant 
investments in STEM education. New investments could help address the 
lack of diversity in STEM programs and provide for more equitable 
access to STEM education for all Americans. Without accelerating the 
growth of a skilled technical workforce, the United States will fall 
behind in the global technology competition. Much of the best STEM 
talent is in the private sector. Many experts in the private sector 
would be willing to contribute to government work if there were more 
effective ways to do so. The AI Commission's Final Report suggests 
several ways for the government to improve its partnerships with the 
private sector; there are actions the private sector can take to 
reciprocate as well. Recommended actions include:
      Public-private partnerships at the state and local level 
for K-12, community college, and university education. Addressing the 
shortfall of digital talent in America will require public-private 
partnerships at all levels of education. High poverty and low-
performing schools in the K-12 academic system are ripe for 
opportunities for partnerships that will benefit American students and 
widen accessibility for a quality education. There are also pockets of 
private-public partnerships that have helped higher education 
institutions respond to the required digital talent needs of local 
areas, but more can be done to incentivize cooperation.
      Pass a National Defense Education Act II (NDEA II). The 
Commission believes the time is right for a second NDEA, one that 
mirrors the intent of the first legislation to increase the technology 
education of the nation's youth, but with important distinctions. NDEA 
II should focus on funding students acquiring digital skills, like 
mathematics, computer science, information science, data science, and 
statistics. NDEA II should include K-12 education and reskilling 
programs that address deficiencies across the spectrum of the American 
educational system, purposefully targeting under-resourced school 
districts. The Commission also recommends investments in university-
level STEM programs with 25,000 undergraduate, 5,000 graduate, and 500 
PhD-level scholarships. Undergraduate scholarships should include 
credit hours at community colleges to ensure more Americans have access 
to affordable STEM education. Ultimately, the goal of NDEA II is to 
widen the digital talent pool by incentivizing programs for 
underrepresented Americans.
      Direct participation on government advisory boards. There 
are several areas where public-private partnerships could allow for 
technology leaders and innovators to sit on advisory boards to assist 
with curriculum development or program execution, such as with the STEM 
Corp and the United States Digital Service Academy (USDSA) proposed by 
the Commission. Our STEM Corps recommendation includes a scholarship 
program, advisory board, private-sector partnership program, and STEM 
Corps member management program. Our USDSA recommendation includes a 
Federal Advisory Committee composed of private sector and academic 
technology leaders.
      Internships. Our report called for a National Reserve 
Digital Corps (NRDC) to allow for part-time support for the government 
as well as a United States Digital Service Academy (USDSA). Both of 
these programs would need private partners that would take in interns 
or fellows as part of their programs.
      Faculty exchanges and government online courses. There is 
a wide array of available online courses and training that already 
exists in the private sector. Much of this--and the content developers 
who produce it--could be used to augment existing programs and 
implement our recommendations. USDSA, for example, could recruit 
adjunct faculty, primarily from private-sector technology companies, to 
augment its tenure-tracked faculty, and ensure that it keeps relevant 
with industry best practices and commercial state-of-the-practice 
techniques. Additionally, the private sector could offer to the 
government already-developed content at low cost or lowered government 
rates.
      Short courses for leaders at all levels. Our 
recommendations called for the Department of Defense to establish a 
short course on emerging technologies for general and flag officers and 
senior executive-level civilian leaders. Industry could also offer 
other opportunities to take in government leaders at all levels for 
short courses where they can be exposed to technologies and software 
that they do not have readily available through their existing 
government infrastructure. They could be shown new and interesting ways 
to think through and solve problems using hardware and software that 
they might then take back to their government departments and agencies.

                                  [all]