WEBVTT

00:01.039 --> 00:03.140
Say that . Good morning .

00:04.900 --> 00:08.739
Ladies and gentlemen , please

00:08.750 --> 00:10.861
remember to silence your phones . The

00:10.861 --> 00:12.917
event will begin shortly . Also as a

00:12.917 --> 00:14.917
reminder , we will not be following

00:14.917 --> 00:17.139
Chatham House rules today . If there is

00:17.139 --> 00:19.361
any press in the audience , please make

00:19.361 --> 00:21.472
your way to the press box in the rear

00:21.472 --> 00:22.917
of the room . Thank you .

00:27.139 --> 00:29.209
Yes . Yeah .

00:31.200 --> 00:32.200
Yes .

00:39.729 --> 00:40.439
What else ?

00:44.349 --> 00:45.349
Right ?

00:58.139 --> 00:59.139
Thank you .

01:12.379 --> 01:15.699
Yeah , thank you ,

01:23.889 --> 01:27.669
Mr Good

01:27.680 --> 01:31.160
morning . My name is Lieutenant Colonel

01:31.169 --> 01:33.610
J Gray Wall and I am a student from the

01:33.620 --> 01:35.860
National War College class of 2025 .

01:39.839 --> 01:41.561
It's my privilege and honor to

01:41.561 --> 01:43.561
introduce our guest speaker today .

01:43.949 --> 01:45.838
Before I get started , guests are

01:45.838 --> 01:48.005
reminded that we will not be following

01:48.005 --> 01:49.949
Chatham house rules today and this

01:49.949 --> 01:52.227
event is being broadcast to the public .

01:52.227 --> 01:54.171
We will have 20 minutes of remarks

01:54.171 --> 01:56.060
followed by 20 minutes of Q and A

01:56.060 --> 01:59.940
moderated by Doctor Rich Andres . Jake

01:59.949 --> 02:01.949
Sullivan is a 28th assistant to the

02:01.949 --> 02:04.339
President for National Security Affairs

02:04.860 --> 02:07.027
in the Obama Biden administration . He

02:07.027 --> 02:09.027
served as a Deputy assistant to the

02:09.027 --> 02:11.249
President and national security adviser

02:11.249 --> 02:13.471
to then Vice President Biden , Director

02:13.471 --> 02:15.416
of Policy Planning staff at the US

02:15.416 --> 02:17.360
Department of State and the Deputy

02:17.360 --> 02:19.360
Chief of staff for the Secretary of

02:19.360 --> 02:21.527
State Hillary Clinton . Please provide

02:21.527 --> 02:23.416
a warm welcome to Jake Sullivan .

02:38.020 --> 02:40.279
Good morning , everyone . And thanks so

02:40.289 --> 02:42.511
much for that introduction , Lieutenant

02:42.511 --> 02:44.678
Colonel Gray Wall . And I also want to

02:44.678 --> 02:46.622
thank the National War College for

02:46.622 --> 02:48.845
bringing us all together today . I want

02:48.845 --> 02:51.011
to thank my colleagues from across the

02:51.011 --> 02:53.178
intelligence community and dod as well

02:53.178 --> 02:55.345
as from the NSC who uh have really put

02:55.345 --> 02:57.949
their blood sweat toil , uh toil and

02:57.960 --> 03:00.080
tears into producing this National

03:00.089 --> 03:01.978
Security memorandum on artificial

03:01.978 --> 03:03.978
intelligence that we're rolling out

03:03.978 --> 03:05.978
today . Most importantly , though I

03:05.978 --> 03:08.145
wanna thank all of you for allowing me

03:08.145 --> 03:10.089
to be here to say a few words this

03:10.089 --> 03:12.460
morning . It's truly an honor for me to

03:12.470 --> 03:14.360
be here . And in fact , there's a

03:14.369 --> 03:16.160
reason I wanted to address this

03:16.169 --> 03:19.690
specific group of leaders more than 75

03:19.699 --> 03:21.921
years ago , just a few months after the

03:21.921 --> 03:24.032
second World War ended , then General

03:24.032 --> 03:26.255
Dwight Eisenhower wrote a letter to his

03:26.255 --> 03:29.539
fellow military leaders all around them .

03:29.550 --> 03:31.750
The world was changing . Nazi Germany

03:31.759 --> 03:34.610
had fallen nations were rebuilding the

03:34.619 --> 03:37.419
Cold War was just beginning and people

03:37.429 --> 03:39.373
everywhere were reckoning with the

03:39.373 --> 03:41.699
horrors of the holocaust . It was a new

03:41.710 --> 03:44.270
era one that demanded new strategies ,

03:44.279 --> 03:47.529
new thinking and new leadership . So

03:47.539 --> 03:49.729
General Eisenhower pitched an idea , a

03:49.740 --> 03:52.330
National War College . He didn't know

03:52.339 --> 03:54.395
where it would be or what exactly it

03:54.395 --> 03:56.506
would look like , but he knew America

03:56.506 --> 03:58.779
needed a school whose primary function

03:58.789 --> 04:02.009
would be as he wrote quote to develop

04:02.020 --> 04:04.470
doctrine rather than to accept and

04:04.479 --> 04:07.539
follow prescribed doctrine , develop ,

04:07.850 --> 04:11.009
not accept and follow . That idea has

04:11.020 --> 04:14.080
guided this institution ever since in

04:14.089 --> 04:16.311
the aftermath of the Second World War .

04:16.311 --> 04:18.367
It led your forebearers to reimagine

04:18.367 --> 04:20.589
our decision making apparatus including

04:20.589 --> 04:22.478
the establishment of the National

04:22.478 --> 04:24.533
Security Council . Thanks for that .

04:25.399 --> 04:27.455
During the Cold War , it led them to

04:27.455 --> 04:29.566
develop new strategies to advance our

04:29.566 --> 04:31.709
national security , including

04:31.720 --> 04:33.887
implementing containment , detente and

04:33.887 --> 04:36.859
beyond . And throughout the global war

04:36.869 --> 04:39.130
against terror , your predecessors have

04:39.140 --> 04:41.307
pioneered new thinking and new tactics

04:41.307 --> 04:43.529
that have helped keep our nation safe .

04:43.869 --> 04:46.329
Now , it's your turn , we're in an age

04:46.339 --> 04:48.061
of strategic competition in an

04:48.061 --> 04:50.450
interdependent world where we have to

04:50.459 --> 04:53.489
compete vigorously and also mobilize

04:53.500 --> 04:55.722
partners to solve great challenges that

04:55.722 --> 04:57.833
no one country can solve on its own .

04:58.510 --> 05:00.570
In this age , in this world , the

05:00.579 --> 05:02.746
application of artificial intelligence

05:02.746 --> 05:05.250
will define the future and our country

05:05.260 --> 05:06.816
must once again develop new

05:06.816 --> 05:08.890
capabilities , new tools . And as

05:08.899 --> 05:11.177
general Eisenhower said , new doctrine ,

05:11.570 --> 05:13.792
if we want to ensure that A I works for

05:13.792 --> 05:15.570
us , for our partners , for our

05:15.570 --> 05:17.626
interests and for our values and not

05:17.626 --> 05:20.059
against us , that's why I'm proud to

05:20.070 --> 05:22.320
announce that President Biden has

05:22.329 --> 05:24.899
signed a national security memorandum

05:24.910 --> 05:27.579
on artificial intelligence . This is

05:27.589 --> 05:30.190
our nation's first ever strategy for

05:30.200 --> 05:32.510
harnessing the power and managing the

05:32.519 --> 05:35.230
risks of A I to advance our national

05:35.239 --> 05:38.269
security . So today , I want to talk to

05:38.279 --> 05:40.279
you about what's brought us to this

05:40.279 --> 05:42.501
moment and how our country needs all of

05:42.501 --> 05:45.489
you to help us meet it . Like many of

05:45.500 --> 05:47.667
you here at the war college , I've had

05:47.667 --> 05:49.222
to grapple with A I and its

05:49.222 --> 05:51.167
implications for national security

05:51.167 --> 05:53.000
since I became national security

05:53.000 --> 05:54.722
advisor about what makes it so

05:54.722 --> 05:57.130
potentially transformative and about

05:57.140 --> 05:59.084
what makes it different from other

05:59.084 --> 06:01.196
technological leaps . Our country has

06:01.196 --> 06:03.700
navigated before , from electrification

06:03.709 --> 06:05.876
to nuclear weapons , to space flight ,

06:05.876 --> 06:08.380
to the internet . And I've seen three

06:08.390 --> 06:11.630
key things in particular . First , the

06:11.640 --> 06:15.079
sheer speed of the development of

06:15.089 --> 06:17.399
artificial intelligence , the technical

06:17.410 --> 06:19.519
frontier of A I continues to advance

06:19.529 --> 06:22.350
rapidly more rapidly than we've seen

06:22.609 --> 06:25.309
with other technologies . Let's just

06:25.320 --> 06:27.376
take protein folding as an example ,

06:27.920 --> 06:30.700
discovering a protein structure or how

06:30.709 --> 06:33.450
it folds is essential for understanding

06:33.459 --> 06:35.681
how it interacts with other molecules ,

06:35.700 --> 06:37.929
which can solve fundamental puzzles in

06:37.940 --> 06:40.540
medicine and accelerate the development

06:40.549 --> 06:44.359
of treatment and cures . Up until 2018 ,

06:44.730 --> 06:47.040
humanity had collectively discovered

06:47.049 --> 06:49.410
the structure of around 100 and 50,000

06:49.420 --> 06:52.440
proteins largely through manual efforts .

06:52.450 --> 06:54.709
Sometimes after years of painstaking

06:54.720 --> 06:56.940
work using advanced microscopes and X

06:56.950 --> 07:00.079
rays . Then Google Deep Mine showed

07:00.089 --> 07:02.311
that A I could predict the structure of

07:02.311 --> 07:05.910
a protein without any wet lab work . By

07:05.920 --> 07:09.709
2022 4 years later , that same team

07:09.720 --> 07:12.070
released predicted structures for

07:12.079 --> 07:14.670
almost every protein known to science .

07:14.890 --> 07:18.369
Hundreds of millions in all , just a

07:18.380 --> 07:21.299
few weeks ago , the scientists involved

07:21.309 --> 07:24.730
won a Nobel Prize . Now imagine that

07:24.739 --> 07:27.769
same pace of change in the realms of

07:27.779 --> 07:30.269
science that impact your work as

07:30.279 --> 07:32.390
national security leaders every day .

07:33.100 --> 07:35.859
Imagine how A I will impact areas where

07:35.869 --> 07:37.925
we're already seeing paradigm shifts

07:37.925 --> 07:39.925
from nuclear physics to rocketry to

07:39.925 --> 07:43.049
stealth or how it could impact areas of

07:43.059 --> 07:45.239
competition that may not have yet

07:45.250 --> 07:47.959
matured . That we actually can't even

07:47.970 --> 07:50.299
imagine just as the early Cold warriors

07:50.309 --> 07:52.476
could not really have imagined today's

07:52.476 --> 07:54.760
cyber operations . Put simply a

07:54.769 --> 07:56.869
specific A I application that we're

07:56.880 --> 07:59.010
trying to solve for today in the

07:59.019 --> 08:01.609
intelligence or military or commercial

08:01.619 --> 08:03.649
domains could look fundamentally

08:03.660 --> 08:06.209
different six weeks from now , let

08:06.220 --> 08:08.799
alone six a month from now or a year

08:08.809 --> 08:11.230
from now or six years from now . The

08:11.239 --> 08:13.609
speed of change in this area is

08:13.619 --> 08:16.929
breathtaking . This is compounded by

08:16.940 --> 08:19.529
huge uncertainty around A I's growth

08:19.540 --> 08:21.373
trajectory , which is the second

08:21.373 --> 08:23.989
distinctive trait . Over the last four

08:24.000 --> 08:26.056
years , I've met with scientists and

08:26.056 --> 08:28.167
entrepreneurs , lab , Ceo S and hyper

08:28.167 --> 08:30.829
scalar researchers and engineers and

08:30.839 --> 08:33.318
civil society advocates . And

08:33.328 --> 08:35.550
throughout all of those conversations ,

08:35.550 --> 08:37.161
there's clear agreement that

08:37.161 --> 08:39.384
developments in artificial intelligence

08:39.384 --> 08:41.384
are having a profound impact on our

08:41.384 --> 08:43.606
world . But opinions diverge when I ask

08:43.606 --> 08:45.679
them , what exactly should we expect

08:45.690 --> 08:49.539
next ? There's a spectrum of views . At

08:49.549 --> 08:51.438
one end , some experts believe we

08:51.438 --> 08:53.493
barely kicked off the A I revolution

08:53.809 --> 08:55.976
that A I capabilities will continue to

08:55.976 --> 08:58.020
grow exponentially building on

08:58.030 --> 09:00.086
themselves to unlock paths we didn't

09:00.086 --> 09:02.369
know existed and that this could happen

09:02.380 --> 09:05.190
fast . Well , within this decade . And

09:05.200 --> 09:07.311
if they're right , we could be on the

09:07.311 --> 09:09.311
cusp of one of the most significant

09:09.311 --> 09:11.969
technological shifts in human history .

09:13.140 --> 09:15.251
At the other end of the spectrum is a

09:15.251 --> 09:17.307
view that A I is in a growth spurt ,

09:17.307 --> 09:20.640
but it has or soon will plateau or at

09:20.650 --> 09:22.594
least the pace of change will slow

09:22.594 --> 09:24.428
considerably . And more dramatic

09:24.428 --> 09:26.761
breakthroughs are further down the road .

09:27.119 --> 09:29.640
Experts who believe this aren't saying

09:29.650 --> 09:31.919
A I won't be consequential . But they

09:31.929 --> 09:33.762
argue that the last mile work of

09:33.762 --> 09:37.010
applying A I that are already here ,

09:37.020 --> 09:39.187
the capabilities that are already here

09:39.187 --> 09:41.650
is what will matter most , not just now ,

09:41.659 --> 09:44.400
but for the foreseeable future , these

09:44.409 --> 09:46.576
views are vastly different with vastly

09:46.576 --> 09:48.880
different implications . Now ,

09:49.200 --> 09:51.780
innovation has never been predictable .

09:52.030 --> 09:54.419
But the degree of uncertainty in A I

09:54.429 --> 09:57.570
development is unprecedented . The size

09:57.580 --> 09:59.747
of the question mark distinguishes A I

09:59.747 --> 10:02.119
for many other technological challenges

10:02.130 --> 10:04.352
our government has had to face and make

10:04.352 --> 10:07.479
policy around and that is our

10:07.489 --> 10:09.489
responsibility as National security

10:09.489 --> 10:11.378
advisor . I have to make sure our

10:11.378 --> 10:13.545
government is ready for every scenario

10:13.545 --> 10:15.767
along the spectrum . We have to build a

10:15.767 --> 10:17.711
national security policy that will

10:17.711 --> 10:19.711
protect the American people and the

10:19.711 --> 10:21.433
American innovation innovation

10:21.433 --> 10:23.656
ecosystem , which is so critical to our

10:23.656 --> 10:25.767
advantage . Even if the opportunities

10:25.767 --> 10:27.878
and challenges we face could manifest

10:27.878 --> 10:29.933
in fundamentally different ways . We

10:29.933 --> 10:31.878
have to be prepared for the entire

10:31.878 --> 10:34.045
spectrum of possibilities of where A I

10:34.045 --> 10:37.140
is headed in 2025 2027 2030 beyond

10:38.609 --> 10:40.387
now , what makes this even more

10:40.387 --> 10:42.880
difficult is that private companies are

10:42.890 --> 10:45.057
leading the development of A I not the

10:45.057 --> 10:46.779
government . This is the third

10:46.779 --> 10:48.919
distinctive feature . Many of the

10:48.929 --> 10:51.039
technological leaps of the last 80

10:51.049 --> 10:53.229
years emerged from public research ,

10:53.289 --> 10:55.559
public funding , public procurement .

10:55.820 --> 10:57.653
Our government took an early and

10:57.653 --> 10:59.764
critical role in shaping developments

10:59.764 --> 11:01.487
from nuclear physics and space

11:01.487 --> 11:04.070
exploration to personal computing to

11:04.080 --> 11:06.979
the internet . That's not been the case

11:06.989 --> 11:09.267
with most of the recent A I revolution .

11:09.780 --> 11:11.780
While the Department of Defense and

11:11.780 --> 11:13.947
other agencies funded a large share of

11:13.947 --> 11:15.919
A I work in the 20th century , the

11:15.929 --> 11:17.985
private sector has propelled much of

11:17.985 --> 11:20.669
the last decade of progress and in many

11:20.679 --> 11:22.846
ways , that's something to celebrate .

11:23.020 --> 11:25.500
It's a testament to American ingenuity

11:25.640 --> 11:28.669
to the American innovation system that

11:28.679 --> 11:30.735
American companies lead the world in

11:30.735 --> 11:33.159
frontier A I . It's America's special

11:33.169 --> 11:35.679
sauce and it's a good thing . The

11:35.690 --> 11:37.801
taxpayers don't have to foot the full

11:37.801 --> 11:39.912
bill for A I training costs which can

11:39.912 --> 11:43.080
be staggeringly high . But those of us

11:43.090 --> 11:45.090
in government have to be clear eyed

11:45.090 --> 11:47.257
about the implications of this dynamic

11:47.257 --> 11:49.312
as both stewards and deploys of this

11:49.312 --> 11:51.890
technology here . Two things can be

11:51.900 --> 11:54.760
true . At the same time on the one hand ,

11:54.890 --> 11:57.112
major technology companies that develop

11:57.112 --> 11:59.750
and deploy A I systems by virtue of

11:59.760 --> 12:02.130
being American have given America a

12:02.140 --> 12:05.280
real national security lead a lead that

12:05.289 --> 12:08.429
we want to extend and they're also

12:08.440 --> 12:10.607
going head to head with PR C companies

12:10.607 --> 12:12.829
like Huawei to provide digital services

12:12.829 --> 12:14.773
to people around the world . We're

12:14.773 --> 12:16.773
supporting those efforts because we

12:16.773 --> 12:18.607
want the United States to be the

12:18.607 --> 12:20.440
technology partner of choice for

12:20.440 --> 12:22.440
countries around the world . On the

12:22.440 --> 12:24.051
other hand , we need to take

12:24.051 --> 12:26.030
responsible steps to ensure fair

12:26.039 --> 12:28.349
competition and open markets , to

12:28.359 --> 12:30.526
protect privacy , human rights , civil

12:30.526 --> 12:33.030
rights , civil liberties , to make sure

12:33.039 --> 12:35.206
that advanced A I systems are safe and

12:35.206 --> 12:38.169
trustworthy to implement safeguards so

12:38.179 --> 12:40.179
that A I isn't used to undercut our

12:40.179 --> 12:43.179
national security . The US government

12:43.190 --> 12:45.159
is fully capable of managing this

12:45.169 --> 12:47.391
healthy tension as long as we're honest

12:47.391 --> 12:49.558
and clear eyed about it and we have to

12:49.558 --> 12:51.336
get this right because there is

12:51.336 --> 12:53.502
probably no other technology that will

12:53.502 --> 12:55.336
be more critical to our national

12:55.336 --> 12:59.000
security in the years ahead . Now ,

12:59.010 --> 13:01.121
when it comes to A I and our national

13:01.121 --> 13:03.177
security , I have both good news and

13:03.177 --> 13:05.750
bad news . The good news is that thanks

13:05.760 --> 13:07.871
to President Biden and Vice President

13:07.871 --> 13:09.704
Harris's leadership , America is

13:09.704 --> 13:11.760
continuing to build a meaningful A I

13:11.760 --> 13:13.919
advantage here at home . President

13:13.929 --> 13:16.096
Biden signed an executive order on the

13:16.096 --> 13:18.207
development and use of A I , the most

13:18.207 --> 13:20.318
comprehensive action that any country

13:20.318 --> 13:22.373
in the world has ever taken on A I .

13:22.373 --> 13:24.318
We've worked to strengthen our A I

13:24.318 --> 13:26.690
talent , hardware , infrastructure and

13:26.700 --> 13:28.840
governance . We've attracted leading

13:28.849 --> 13:30.960
researchers and entrepreneurs to move

13:30.960 --> 13:33.016
to and remain in the United States .

13:33.260 --> 13:35.260
We've unleashed tens of billions of

13:35.260 --> 13:37.149
dollars in incentives to catalyze

13:37.149 --> 13:39.371
domestic leading edge chip production .

13:39.489 --> 13:41.711
We've led the world in issuing guidance

13:41.711 --> 13:43.822
to make sure that A I development and

13:43.822 --> 13:45.989
use is safe , secure and trustworthy .

13:46.030 --> 13:48.141
And as we've done all of this , we've

13:48.141 --> 13:50.030
scrutinized A I trends , not just

13:50.030 --> 13:52.500
frontier A I , but also the A I models

13:52.510 --> 13:54.621
that will proliferate most widely and

13:54.621 --> 13:56.677
rapidly around the world . And we're

13:56.677 --> 13:58.843
working to enhance American advantages

13:58.843 --> 14:00.969
across the board . But here's the bad

14:00.979 --> 14:03.859
news . Our lead is not guaranteed , it

14:03.880 --> 14:06.549
is not preordained and it is not enough

14:06.559 --> 14:08.670
to just guard the progress we've made

14:08.670 --> 14:11.010
as historic as it's been . We have to

14:11.020 --> 14:13.630
be faster in deploying A I in our

14:13.640 --> 14:15.950
national security enterprise than

14:15.960 --> 14:19.150
America's rivals are in theirs . They

14:19.159 --> 14:21.270
are in a persistent quest to leapfrog

14:21.270 --> 14:22.937
our military and intelligence

14:22.937 --> 14:24.937
capabilities . And the challenge is

14:24.937 --> 14:26.770
even more acute because they are

14:26.770 --> 14:28.603
unlikely to be bound by the same

14:28.603 --> 14:30.603
principles and responsibilities and

14:30.603 --> 14:33.729
values that we are . The stakes are

14:33.739 --> 14:35.890
high . If we don't act more

14:35.900 --> 14:38.250
intentionally to seize our advantages .

14:38.570 --> 14:41.549
If we don't deploy A I more quickly and

14:41.559 --> 14:43.726
more comprehensively to strengthen our

14:43.726 --> 14:46.309
national security , we risk squandering

14:46.320 --> 14:49.530
our hard earned league even if we have

14:49.539 --> 14:51.700
the best A I models . But our

14:51.710 --> 14:54.780
competitors are faster to deploy . We

14:54.789 --> 14:56.900
could see them seize the advantage in

14:56.900 --> 14:58.845
using A I capabilities against our

14:58.845 --> 15:00.900
people , our forces and our partners

15:00.900 --> 15:03.849
and allies . We could have the best

15:03.859 --> 15:06.026
team but lose because we didn't put it

15:06.026 --> 15:08.830
on the field . We could see advantages

15:08.840 --> 15:11.007
we built over decades in other domains

15:11.007 --> 15:13.419
like space and undersea operations be

15:13.429 --> 15:16.789
reduced or eroded entirely with A I

15:16.799 --> 15:20.140
enabled technology . And for all our

15:20.150 --> 15:22.549
strengths , there remains a risk of

15:22.559 --> 15:25.450
strategic surprise , we have to guard

15:25.460 --> 15:27.516
against that , which is why I'm here

15:27.516 --> 15:30.159
today . Our new National security

15:30.169 --> 15:32.409
memorandum on A I seeks to address

15:32.419 --> 15:34.559
exactly this set of challenges and as

15:34.570 --> 15:36.900
rising national security leaders , you

15:36.909 --> 15:38.965
will be charged with implementing it

15:38.965 --> 15:41.890
with no time to lose . So in the

15:41.900 --> 15:44.122
balance of my remarks , I want to spend

15:44.122 --> 15:45.733
a few minutes explaining the

15:45.733 --> 15:48.067
memorandum's three main lines of effort ,

15:48.067 --> 15:50.178
securing American leadership in A I ,

15:50.178 --> 15:52.233
harnessing A I for national security

15:52.233 --> 15:54.233
and strengthening international A I

15:54.233 --> 15:57.030
partnerships . First , we have to

15:57.039 --> 15:59.559
ensure the United States continues to

15:59.570 --> 16:02.630
lead the world in developing A I . Our

16:02.640 --> 16:04.880
competitors also know how important A I

16:04.890 --> 16:07.280
leadership is in today's age of

16:07.289 --> 16:09.440
geopolitical competition and they are

16:09.450 --> 16:11.506
investing huge resources to seize it

16:11.506 --> 16:13.561
for themselves . So we have to start

16:13.561 --> 16:15.617
upping our game and that starts with

16:15.617 --> 16:18.200
people . America has to continue to be

16:18.210 --> 16:20.432
a magnet for global scientific and tech

16:20.432 --> 16:22.599
talent . As I noted , we've already

16:22.609 --> 16:24.831
taken major steps to make it easier and

16:24.831 --> 16:26.640
faster for top A I scientists ,

16:26.650 --> 16:28.817
engineers and entrepreneurs to come to

16:28.817 --> 16:30.650
the United States , including by

16:30.650 --> 16:33.070
removing friction in our visa rules to

16:33.080 --> 16:35.247
attract talent from around the world .

16:35.690 --> 16:37.912
And through this new memorandum , we're

16:37.912 --> 16:40.023
taking more steps , streamlining visa

16:40.023 --> 16:41.746
processing wherever we can for

16:41.746 --> 16:43.690
applicants , working with emerging

16:43.690 --> 16:45.690
technologies . And we're calling on

16:45.690 --> 16:47.801
Congress to get in the game with us .

16:47.801 --> 16:49.700
Staple more green cards to stem

16:49.710 --> 16:51.766
diplomas as President Biden has been

16:51.766 --> 16:54.659
pushing to do for years . So that's the

16:54.669 --> 16:56.447
people part of the qua . Uh The

16:56.447 --> 16:58.599
equation next is hardware and power

16:59.500 --> 17:01.719
developing advanced A I systems

17:01.729 --> 17:03.820
requires large volumes of advanced

17:03.830 --> 17:06.928
chips and keeping those A I systems

17:06.938 --> 17:09.598
humming requires large amounts of power

17:10.348 --> 17:12.749
on chips . We've taken really

17:12.759 --> 17:15.680
significant steps forward . We passed

17:15.689 --> 17:17.633
the chips and Science Act making a

17:17.633 --> 17:19.356
generational investment in our

17:19.356 --> 17:21.578
semiconductor manufacturing , including

17:21.578 --> 17:23.633
the leading edge logic chips and the

17:23.633 --> 17:25.800
high bandwidth memory chips needed for

17:25.800 --> 17:28.650
A I . We've also taken decisive action

17:28.660 --> 17:31.030
to limit strategic competitors access

17:31.260 --> 17:33.910
to the most advanced chips necessary to

17:33.920 --> 17:36.260
train and use frontier A I systems with

17:36.270 --> 17:39.000
national security implications as well

17:39.010 --> 17:41.343
as the tools needed to make those chips .

17:42.109 --> 17:44.331
The National security memorandum builds

17:44.331 --> 17:46.387
on this progress by directing all of

17:46.387 --> 17:48.553
our national security agencies to make

17:48.553 --> 17:50.442
sure that those vital chip supply

17:50.442 --> 17:53.390
chains are secure and free from foreign

17:53.400 --> 17:57.069
interference on power . The memorandum

17:57.079 --> 17:59.589
recognizes the importance of designing

17:59.630 --> 18:02.140
permitting and constructing clean

18:02.150 --> 18:04.261
energy generation facilities that can

18:04.261 --> 18:06.569
serve A I data centers so that the

18:06.579 --> 18:09.069
companies building world leading A I

18:09.079 --> 18:10.857
infrastructure build as much as

18:10.857 --> 18:13.359
possible here in the United States in a

18:13.369 --> 18:15.591
way that is consistent with our climate

18:15.591 --> 18:18.689
goals . One thing is for certain , if

18:18.699 --> 18:20.500
we don't rapidly build out this

18:20.510 --> 18:22.677
infrastructure in the next few years ,

18:23.359 --> 18:25.300
adding tens or even hundreds of

18:25.310 --> 18:27.477
gigawatts of clean power to the grid .

18:27.920 --> 18:31.569
We will risk falling behind . Finally ,

18:31.579 --> 18:33.939
there's funding for innovation this

18:33.949 --> 18:36.140
fiscal year federal funding for

18:36.150 --> 18:37.761
non-defense R and D declined

18:37.761 --> 18:40.109
significantly and Congress still hasn't

18:40.119 --> 18:42.339
appropriated the science part of the

18:42.349 --> 18:44.540
Chips and Science Act . Even while

18:44.550 --> 18:46.550
China is increasing its science and

18:46.550 --> 18:48.910
technology budget 10% year over year .

18:49.969 --> 18:52.191
That can mean critical gaps in A Ir and

18:52.191 --> 18:54.449
D . We want to work with Congress to

18:54.459 --> 18:56.070
make sure this and the other

18:56.070 --> 18:58.126
requirements within the A I National

18:58.126 --> 19:00.369
Security memorandum are funded and

19:00.380 --> 19:02.213
we've received strong bipartisan

19:02.213 --> 19:04.269
signals of support for this from the

19:04.269 --> 19:05.936
hill . So it's time for us to

19:05.936 --> 19:08.047
collectively roll up our sleeves on a

19:08.047 --> 19:10.269
bicameral bipartisan basis and get this

19:10.269 --> 19:12.280
done . And we also have to be aware

19:12.290 --> 19:14.179
that our competitors are watching

19:14.179 --> 19:16.346
closely . Not least because they would

19:16.346 --> 19:18.660
love to depose our A I leadership one

19:18.670 --> 19:20.892
playbook . We've seen them deploy again

19:20.892 --> 19:23.329
and again is theft and espionage . So

19:23.339 --> 19:25.506
the National Security memorandum takes

19:25.506 --> 19:27.750
this head on it establishes addressing

19:27.760 --> 19:29.649
adversary threats against our A I

19:29.649 --> 19:31.719
sector as a top tier intelligence

19:31.729 --> 19:33.660
priority . A move that means more

19:33.670 --> 19:35.726
resources and more personnel will be

19:35.726 --> 19:38.239
devoted to combating this threat . It

19:38.250 --> 19:40.361
also directs people across government

19:40.361 --> 19:42.194
like so many of you to work more

19:42.194 --> 19:43.972
closely with private sector A I

19:43.972 --> 19:46.160
developers to provide them with timely

19:46.170 --> 19:48.337
cyber security and counterintelligence

19:48.337 --> 19:50.392
information to keep their technology

19:50.400 --> 19:52.880
secure just as we've already worked

19:52.890 --> 19:55.112
with , to protect other elements of the

19:55.112 --> 19:57.279
US private sector from threats to them

19:57.279 --> 20:00.079
to our national security . The second

20:00.089 --> 20:03.219
pillar focuses on how we harness our

20:03.229 --> 20:06.430
advantage and our enduring advantage to

20:06.439 --> 20:09.550
advance national security as national

20:09.560 --> 20:11.560
security advisor . I see how A I is

20:11.560 --> 20:13.338
already poised to transform the

20:13.338 --> 20:15.449
national security landscape and where

20:15.449 --> 20:17.780
you sit as war fighters , as diplomats ,

20:17.790 --> 20:19.920
as intelligence officers . I'm sure

20:19.930 --> 20:22.390
you're seeing it too . Some change is

20:22.400 --> 20:24.810
already here A I is reshaping our

20:24.819 --> 20:26.819
logistics , our cyber vulnerability

20:26.819 --> 20:28.790
detection , how we analyze and

20:28.949 --> 20:32.069
synthesize intelligence . Some change

20:32.079 --> 20:34.301
we see looming on the horizon including

20:34.301 --> 20:36.420
A I enabled applications that will

20:36.430 --> 20:38.829
transform the way our military trains

20:38.839 --> 20:41.560
and fights . But some change . As I

20:41.569 --> 20:43.736
said earlier , we truly cannot predict

20:43.760 --> 20:45.871
in both the form it will take and how

20:45.871 --> 20:47.900
fast it will come . Bottom line

20:48.040 --> 20:50.619
opportunities are already at hand and

20:50.630 --> 20:52.630
more soon will be . So we've got to

20:52.630 --> 20:54.741
seize them quickly and effectively or

20:54.741 --> 20:57.369
our competitors will first . That means

20:57.380 --> 20:59.500
all of us in the national security

20:59.510 --> 21:01.510
enterprise have to become much more

21:01.510 --> 21:04.050
adept users of A I . It means we need

21:04.060 --> 21:06.010
to make significant technical

21:06.489 --> 21:08.760
organizational and policy changes to

21:08.770 --> 21:10.992
ease collaboration with the actors that

21:10.992 --> 21:13.159
are driving this development . And the

21:13.159 --> 21:15.326
National Security memorandum does just

21:15.326 --> 21:17.369
that it directs agencies to propose

21:17.380 --> 21:19.047
ways to enable more effective

21:19.047 --> 21:20.936
collaboration with nontraditional

21:20.936 --> 21:23.219
vendors such as leading A I companies

21:23.229 --> 21:25.469
and cloud computing providers . In

21:25.479 --> 21:27.590
practice , that means quickly putting

21:27.599 --> 21:29.821
the most advanced systems to use in our

21:29.821 --> 21:32.469
national security enterprise just after

21:32.479 --> 21:35.479
they're developed , like how many in

21:35.489 --> 21:37.711
private industry are doing ? We need to

21:37.711 --> 21:39.600
be getting fast adoption of these

21:39.600 --> 21:41.479
systems which are iterating and

21:41.489 --> 21:44.670
advancing as we see every few months .

21:46.380 --> 21:48.680
Next . Today's A I systems are more

21:48.689 --> 21:52.119
generally capable than the bespoke and

21:52.130 --> 21:54.352
narrow tools that dominated prior A I .

21:54.352 --> 21:56.463
And this general capability is a huge

21:56.463 --> 21:58.989
advantage , but the flip side is they

21:59.000 --> 22:01.650
cost much more to train and run . So

22:01.660 --> 22:03.716
we're pushing agencies to use shared

22:03.716 --> 22:05.827
computing resources to accelerate A I

22:05.827 --> 22:08.119
adoption , lower costs and learn from

22:08.130 --> 22:11.099
one another as they responsibly address

22:11.109 --> 22:13.165
a wide range of threats from nuclear

22:13.165 --> 22:14.665
security to biosecurity to

22:14.665 --> 22:16.849
cybersecurity . And I emphasize that

22:16.859 --> 22:19.930
word responsibly developing and

22:19.939 --> 22:22.630
deploying A I safely securely and yes ,

22:22.640 --> 22:24.584
responsibly is the backbone of our

22:24.584 --> 22:27.040
strategy that includes ensuring that A

22:27.050 --> 22:29.520
I systems are free of bias and

22:29.530 --> 22:32.599
discrimination . This is profoundly in

22:32.609 --> 22:35.609
our self interest . One reason is that

22:35.619 --> 22:38.010
even if we can attract A I talent or

22:38.020 --> 22:39.964
foster A I development here in the

22:39.964 --> 22:41.964
United States , we won't be able to

22:41.964 --> 22:44.076
lead the world if people do not trust

22:44.076 --> 22:46.298
our systems . And that means developing

22:46.298 --> 22:48.076
standards for A I evaluations ,

22:48.076 --> 22:50.298
including what makes those systems work

22:50.298 --> 22:52.298
and how they might fail in the real

22:52.298 --> 22:54.409
world . It means running tests on the

22:54.409 --> 22:56.298
world's most advanced A I systems

22:56.298 --> 22:58.520
before they're released to the public .

22:58.520 --> 23:00.631
And it means leading the way in areas

23:00.631 --> 23:02.409
like content authentication and

23:02.409 --> 23:04.631
watermark . So people know when they're

23:04.631 --> 23:06.579
interacting with A I as opposed to

23:06.589 --> 23:08.811
interacting with , for example , a real

23:08.811 --> 23:11.550
human to do all of that , we have to

23:11.560 --> 23:14.150
empower and learn from a full range of

23:14.160 --> 23:16.271
A I firms , experts and entrepreneurs

23:16.271 --> 23:18.479
which our A I Safety Institute is now

23:18.489 --> 23:22.170
doing on a daily basis . Another reason

23:22.180 --> 23:23.736
we need to focus so much on

23:23.736 --> 23:25.609
responsibility safety and and

23:25.619 --> 23:27.397
trustworthiness is a little bit

23:27.397 --> 23:29.469
counterintuitive , ensuring security

23:29.479 --> 23:31.368
and trustworthiness will actually

23:31.368 --> 23:33.535
enable us to move faster , not slow us

23:33.535 --> 23:36.979
down . Put simply uncertainty breeds

23:36.989 --> 23:39.699
caution . When we lack confidence about

23:39.709 --> 23:41.820
safety and reliability , we're slower

23:41.820 --> 23:44.540
to experiment to adopt , to use new

23:44.550 --> 23:46.661
capabilities and we just can't afford

23:46.661 --> 23:48.439
to do that in today's strategic

23:48.439 --> 23:50.589
landscape . That's why our memorandum

23:50.599 --> 23:52.766
directs the first ever government wide

23:52.766 --> 23:55.089
framework on A I risk management

23:55.099 --> 23:57.155
commitments in the national security

23:57.155 --> 23:59.155
space commitments , like refraining

23:59.155 --> 24:01.377
from uses that depart from our nation's

24:01.377 --> 24:03.849
core values , avoiding harmful bias and

24:03.859 --> 24:05.415
discrimination , maximizing

24:05.420 --> 24:07.800
accountability , ensuring effective and

24:07.810 --> 24:11.609
appropriate human oversight . As I

24:11.619 --> 24:13.730
said , preventing misuse and ensuring

24:13.730 --> 24:15.841
high standards of accountability will

24:15.841 --> 24:18.008
not slow us down . It will actually do

24:18.008 --> 24:19.952
the opposite . And we've seen this

24:19.952 --> 24:22.175
before with technological change during

24:22.175 --> 24:24.286
the early days of the railroads , for

24:24.286 --> 24:26.397
example , the establishment of safety

24:26.397 --> 24:28.675
standards enabled trains to run faster ,

24:28.675 --> 24:30.452
thanks to increased certainty ,

24:30.452 --> 24:33.310
confidence and compatibility . And I

24:33.319 --> 24:35.375
also want to know we're gonna update

24:35.375 --> 24:37.597
this framework regularly this goes back

24:37.597 --> 24:39.875
to the uncertainty I mentioned earlier ,

24:39.875 --> 24:41.819
there may be capabilities or novel

24:41.819 --> 24:43.986
legal issues that just haven't emerged

24:43.986 --> 24:46.041
yet . We must and we will ensure our

24:46.041 --> 24:47.986
governance and our guard rails can

24:47.986 --> 24:50.208
adapt to meet the moment no matter what

24:50.208 --> 24:52.430
it looks like or how quickly it comes .

24:52.430 --> 24:54.597
Finally , we need to do all of this in

24:54.597 --> 24:56.708
lockstep with our partners , which is

24:56.708 --> 24:58.763
the third pillar of our memorandum .

24:58.763 --> 25:00.930
President Biden often says we're gonna

25:00.930 --> 25:02.986
see more technological change in the

25:02.986 --> 25:05.097
next 10 years than we saw in the last

25:05.097 --> 25:07.097
50 . He's right and it doesn't just

25:07.097 --> 25:08.875
apply to our country but to all

25:08.875 --> 25:10.930
countries . And when it comes to A I

25:10.930 --> 25:13.041
specifically , we need to ensure that

25:13.041 --> 25:15.041
people around the world are able to

25:15.041 --> 25:17.041
seize the benefits and mitigate the

25:17.041 --> 25:18.597
risks . That means building

25:18.597 --> 25:20.652
international norms and partnerships

25:20.652 --> 25:22.541
around A I . Over the last year ,

25:22.541 --> 25:24.652
thanks to the leadership of President

25:24.652 --> 25:26.875
Biden and Vice President Harris , we've

25:26.875 --> 25:29.097
laid that foundation . We developed the

25:29.097 --> 25:30.930
first ever international code of

25:30.930 --> 25:33.041
conduct on A I with our G7 partners .

25:33.041 --> 25:35.152
We joined more than two dozen nations

25:35.152 --> 25:37.319
at the Bletchley and Seoul A I summits

25:37.319 --> 25:39.640
to outline clear A I principles . We

25:39.650 --> 25:41.761
released our political declaration on

25:41.761 --> 25:43.983
the military use of A I which more than

25:43.983 --> 25:46.206
50 countries have endorsed , to outline

25:46.206 --> 25:48.372
what constitutes responsible practices

25:48.380 --> 25:50.547
for using A I in the military domain .

25:51.000 --> 25:53.020
And we sponsored the first ever UN

25:53.030 --> 25:55.079
General Assembly resolution on A I

25:55.089 --> 25:57.500
which passed unanimously including with

25:57.510 --> 25:59.939
the Pr C I might add as a co sponsor .

26:00.959 --> 26:03.126
It makes clear that as I said , we can

26:03.126 --> 26:05.292
both seize the benefits of A I for the

26:05.292 --> 26:09.040
world and advance A I safety . Let me

26:09.050 --> 26:11.161
take just a moment to speak about the

26:11.161 --> 26:13.430
PR C specifically almost a year ago

26:13.439 --> 26:15.550
when President Biden and President Xi

26:15.550 --> 26:17.772
met in San Francisco , they agreed to a

26:17.772 --> 26:19.995
dialogue between our two countries on A

26:19.995 --> 26:22.469
I risk and safety . And this past May ,

26:22.479 --> 26:24.312
some of our government's top A I

26:24.312 --> 26:26.550
experts met PR C officials in Geneva

26:26.869 --> 26:28.980
for a candid and constructive initial

26:28.980 --> 26:31.530
conversation . I strongly believe that

26:31.540 --> 26:33.651
we should always be willing to engage

26:33.651 --> 26:35.818
in dialogue about this technology with

26:35.818 --> 26:37.939
the PR C and with others to better

26:37.949 --> 26:39.560
understand risks and counter

26:39.560 --> 26:42.560
misperceptions . But those meetings do

26:42.569 --> 26:44.625
not diminish our deep concerns about

26:44.625 --> 26:46.847
the ways in which the PR C continues to

26:46.847 --> 26:48.847
use A I to repress its population ,

26:48.847 --> 26:51.109
spread misinformation and undermine the

26:51.119 --> 26:53.230
security of the United States and our

26:53.230 --> 26:56.060
allies and partners A I should be used

26:56.069 --> 26:58.339
to unleash possibilities and empower

26:58.349 --> 27:00.709
people and nations around the world ,

27:00.719 --> 27:03.369
especially developing economies want to

27:03.380 --> 27:05.650
know how to do that . They don't want

27:05.660 --> 27:07.930
to be left behind and we don't want

27:07.939 --> 27:10.680
that either . Our national security has

27:10.689 --> 27:12.800
always been stronger when we extend a

27:12.800 --> 27:15.089
hand to partners around the world . So

27:15.099 --> 27:17.210
we need to get the balance right . We

27:17.210 --> 27:19.266
need to balance protecting , cutting

27:19.266 --> 27:21.488
edge A I technologies on the one hand ,

27:21.488 --> 27:23.488
while also promoting A I technology

27:23.488 --> 27:25.880
adoption around the world , protect and

27:25.890 --> 27:28.890
promote we can and must and are doing

27:28.900 --> 27:32.849
both . So let me briefly preview for

27:32.859 --> 27:34.800
you a new global approach to A I

27:34.810 --> 27:38.540
diffusion , how A I can spread around

27:38.550 --> 27:40.550
the world in a responsible way that

27:40.550 --> 27:43.089
allows A I for good while protecting

27:43.099 --> 27:46.199
against downside risk . This new global

27:46.209 --> 27:48.209
approach complements the memorandum

27:48.209 --> 27:50.153
that has come out and comes out of

27:50.153 --> 27:52.376
extended conversations in the situation

27:52.376 --> 27:54.420
room and with allies , industry and

27:54.430 --> 27:57.390
partners over the last year , the finer

27:57.400 --> 27:59.511
print will come out later . But I can

27:59.511 --> 28:01.622
say now that it will give the private

28:01.622 --> 28:03.789
sector more clarity and predictability

28:03.789 --> 28:05.733
as they plan to invest hundreds of

28:05.733 --> 28:07.733
billions of dollars globally . This

28:07.733 --> 28:09.956
includes how our government will manage

28:09.956 --> 28:12.067
the export of the most advanced chips

28:12.067 --> 28:14.233
necessary to develop frontier models .

28:14.660 --> 28:16.719
How we will ensure broad access to

28:16.729 --> 28:18.785
substantial A I computing power that

28:18.785 --> 28:21.007
lies behind the bleeding edge but could

28:21.007 --> 28:22.729
nonetheless transform health ,

28:22.729 --> 28:24.785
agriculture and manufacturing around

28:24.785 --> 28:27.007
the world . How we will help facilitate

28:27.007 --> 28:29.229
partnerships between leading American A

28:29.229 --> 28:31.396
I firms and countries around the world

28:31.396 --> 28:33.173
that want to be part of the A I

28:33.173 --> 28:35.569
revolution and how we will set safety

28:35.579 --> 28:37.412
and security standards for these

28:37.412 --> 28:39.523
partnerships to ensure we effectively

28:39.523 --> 28:41.800
protect against risks while unleashing

28:41.810 --> 28:45.239
new opportunities . These partnerships

28:45.479 --> 28:47.930
are critical , they're fundamental to

28:47.939 --> 28:50.209
our leadership . We know that China is

28:50.219 --> 28:51.941
building its own technological

28:51.941 --> 28:54.052
ecosystem with digital infrastructure

28:54.052 --> 28:56.219
that won't protect sensitive data that

28:56.219 --> 28:58.052
can enable mass surveillance and

28:58.052 --> 28:59.780
censorship that can spread

28:59.790 --> 29:01.849
misinformation and that can make

29:01.859 --> 29:04.869
countries vulnerable to coercion . So

29:04.880 --> 29:06.936
we have to compete to provide a more

29:06.936 --> 29:09.430
attractive path ideally before

29:09.439 --> 29:11.606
countries go too far down an untrusted

29:11.606 --> 29:13.828
road from which it can be expensive and

29:13.828 --> 29:16.119
difficult to return . And that's what

29:16.130 --> 29:18.241
we're doing . We've already developed

29:18.241 --> 29:20.186
new partnerships that will support

29:20.186 --> 29:22.109
economic progress , technological

29:22.119 --> 29:24.089
innovation and indigenous A I

29:24.099 --> 29:26.750
ecosystems from Africa to Asia , to the

29:26.760 --> 29:28.949
Middle East and beyond . And we're

29:28.959 --> 29:30.848
gonna keep at it with a clear and

29:30.848 --> 29:32.903
rigorous approach to A I diffusion .

29:33.680 --> 29:35.810
Now , I do want to make sure I leave

29:35.819 --> 29:37.930
time for our conversation . So let me

29:37.930 --> 29:39.930
just close with this . Everything I

29:39.930 --> 29:42.041
just laid out is a plan , but we need

29:42.041 --> 29:44.152
all of you to turn it into progress .

29:44.152 --> 29:46.208
We need you and leaders across every

29:46.208 --> 29:48.263
state and every sector to adopt this

29:48.263 --> 29:50.319
technology , to advance our national

29:50.319 --> 29:52.760
security and to do it fast . We need

29:52.770 --> 29:54.992
you to ensure that our work aligns with

29:54.992 --> 29:56.826
the core values that have always

29:56.826 --> 29:58.937
underpinned American leadership . And

29:58.937 --> 30:01.770
as President Eisenhower said , we need

30:01.780 --> 30:04.599
you to constantly update and develop

30:04.609 --> 30:07.089
our A I doctrine in the years ahead .

30:09.160 --> 30:11.579
It will be hard , it will require

30:11.589 --> 30:14.199
constant thinking , constant rethinking ,

30:14.209 --> 30:16.069
constant innovation , constant

30:16.079 --> 30:18.301
collaboration and constant leadership .

30:18.869 --> 30:21.091
But with the past is our proof . I know

30:21.091 --> 30:23.036
that everyone in this room and all

30:23.036 --> 30:25.036
across our country is up for it and

30:25.036 --> 30:27.091
together we will win the competition

30:27.091 --> 30:29.313
for the 21st century . So , thank you .

30:29.313 --> 30:31.591
And I look forward to the conversation .

30:39.729 --> 30:42.069
Good morning . My name is Professor

30:42.079 --> 30:44.079
Richard Anderson , the National War

30:44.079 --> 30:46.246
College . And I will be moderating the

30:46.246 --> 30:48.579
question and answer portion of the talk .

30:48.579 --> 30:50.523
But before we started , I just , I

30:50.523 --> 30:52.523
wanted to say that uh it's , it's a

30:52.523 --> 30:54.523
little bit sobering and quite a bit

30:54.523 --> 30:56.635
inspirational uh to hear this message

30:56.635 --> 30:58.801
this morning . And I , I couldn't help

30:58.801 --> 31:01.023
thinking about this . Historically , if

31:01.023 --> 31:03.190
100 years ago , your predecessors here

31:03.190 --> 31:04.968
hearing about air power and the

31:04.968 --> 31:07.079
potential role that was going to play

31:07.079 --> 31:09.246
might have felt uh the way that you do

31:09.246 --> 31:11.468
this morning , there's a lot of changes

31:11.468 --> 31:13.690
coming . So um with that , I would like

31:13.690 --> 31:15.912
to invite the students to uh to come up

31:15.912 --> 31:18.079
to the microphone to ask any questions

31:18.079 --> 31:20.301
that you might have . And two reminders

31:20.301 --> 31:22.523
uh this morning , uh The first one is ,

31:22.523 --> 31:24.635
please introduce yourself your name ,

31:24.635 --> 31:26.635
rank the service or agency and this

31:26.635 --> 31:28.746
college that you're with . And uh the

31:28.746 --> 31:31.199
second reminder is this is not repeat ,

31:31.209 --> 31:33.439
this is not Chatham House Rules . The

31:33.449 --> 31:35.560
press is in attendance this morning .

31:35.560 --> 31:37.838
And so your comments are on the record ,

31:38.640 --> 31:40.473
sir . Thank you for that , sir ,

31:40.473 --> 31:42.584
Lieutenant Colonel Mike Gallucci , uh

31:42.584 --> 31:44.640
represent our United States Army and

31:44.640 --> 31:46.619
I'm a very proud member of the

31:46.630 --> 31:48.574
Eisenhower School and to the point

31:48.574 --> 31:50.797
about Town House Rules , all that great

31:50.797 --> 31:53.019
stuff . Uh We definitely uh some of the

31:53.019 --> 31:55.186
students who are more intelligent than

31:55.186 --> 31:57.186
I helped me make this question just

31:57.186 --> 31:59.408
right to get the essence of it and keep

31:59.408 --> 32:01.630
it in the box . So that being said , uh

32:01.630 --> 32:03.574
sir , for decades , the theocratic

32:03.574 --> 32:05.574
state regime in Iran played a now a

32:05.574 --> 32:07.352
significant role in the largely

32:07.352 --> 32:09.408
destabilized Middle East . And we've

32:09.408 --> 32:11.241
been hearing about over the last

32:11.241 --> 32:13.408
several years that the media has shown

32:13.408 --> 32:15.352
that Iranian youth are potentially

32:15.352 --> 32:17.400
ready for a paradigm shift into a

32:17.410 --> 32:20.579
different type of government . So that

32:20.589 --> 32:23.609
being said , uh perhaps if that were to

32:23.619 --> 32:25.339
happen , uh what would be the

32:25.349 --> 32:28.780
implications for our United States and

32:28.790 --> 32:30.734
how might those implications drive

32:30.734 --> 32:33.140
foreign policy uh in the Middle East

32:33.150 --> 32:35.317
moving forward again , sir ? Thank you

32:35.317 --> 32:37.539
very much . Appreciate the time . Thank

32:37.539 --> 32:39.761
you and thank you for your service . Um

32:40.880 --> 32:43.060
We start from the premise that Iran

32:43.069 --> 32:45.699
represents a destabilizing and

32:45.709 --> 32:47.900
dangerous threat to American interests

32:47.910 --> 32:49.910
and the interests of our allies and

32:49.910 --> 32:51.910
partners across the Middle East and

32:51.910 --> 32:54.132
beyond . Uh That's true in the way that

32:54.132 --> 32:56.369
they fund and fuel proxies who are

32:56.380 --> 32:59.650
attacking our troops ,

33:00.189 --> 33:03.359
attacking our friends . Uh and are

33:03.369 --> 33:05.599
undermining a vision of peace and

33:05.609 --> 33:07.553
stability for the Middle East that

33:07.589 --> 33:09.478
countries of good faith have been

33:09.478 --> 33:11.533
pursuing for a long time . So we are

33:11.533 --> 33:13.645
clear eyed about the threat that Iran

33:13.645 --> 33:15.867
poses and we have been working across a

33:15.867 --> 33:17.867
range of national security tools to

33:17.867 --> 33:19.922
push back against that threat and to

33:19.922 --> 33:21.978
support our partners in pushing back

33:21.978 --> 33:24.200
against that threat . And maybe there's

33:24.200 --> 33:26.145
no more real world in living color

33:26.145 --> 33:28.300
example of that than the fact that uh

33:28.310 --> 33:30.989
our brave men and women helped directly

33:31.000 --> 33:33.390
defend Israel against a rainfall of

33:33.400 --> 33:35.689
Iranian ballistic missiles just a few

33:35.699 --> 33:38.810
weeks ago , rendering that attack

33:38.819 --> 33:40.819
ineffective and helping protect our

33:40.819 --> 33:43.260
partner . And showing also that we have

33:43.270 --> 33:45.410
the capacity and the wherewithal to

33:45.420 --> 33:47.642
stand up and defend against the kind of

33:47.642 --> 33:49.587
threat that Iran poses . Iran also

33:49.587 --> 33:51.809
poses a threat in so far as we continue

33:51.809 --> 33:53.642
to have grave concerns about its

33:53.642 --> 33:55.864
nuclear program , and we are determined

33:55.864 --> 33:57.698
to prevent Iran from acquiring a

33:57.698 --> 33:59.698
nuclear weapon . But the essence of

33:59.698 --> 34:01.587
your question is really about the

34:01.587 --> 34:03.864
future of Iran's political system here .

34:03.864 --> 34:06.031
Our view is quite straightforward . We

34:06.031 --> 34:08.198
believe that the future of Iran should

34:08.198 --> 34:10.698
be up to the Iranian people and we do

34:10.708 --> 34:12.875
not believe that the United States has

34:12.875 --> 34:14.875
a particularly good track record of

34:14.875 --> 34:16.875
designing policy to try to generate

34:16.875 --> 34:19.097
regime change in other countries . That

34:19.097 --> 34:21.208
that should be what we do . What what

34:21.208 --> 34:23.389
we should do is design a strategy to

34:23.399 --> 34:25.121
push back against their malign

34:25.121 --> 34:27.066
activities and behavior across the

34:27.066 --> 34:29.232
region and then stand on principle and

34:29.232 --> 34:31.600
in support of universal human rights

34:31.610 --> 34:33.832
and values and the proposition that the

34:33.832 --> 34:35.999
Iranian people should be able to chart

34:35.999 --> 34:37.777
a course for their own future ,

34:37.777 --> 34:39.554
consistent with those univers ,

34:39.554 --> 34:41.888
universal rights and values . Thank you ,

34:42.989 --> 34:44.656
please . Good morning , sir ,

34:44.656 --> 34:46.822
Lieutenant Colonel Mike Walls , United

34:46.822 --> 34:49.156
States Marine Corps , Eisenhower School .

34:49.156 --> 34:51.267
So I'm wondering if you wouldn't mind

34:51.267 --> 34:50.709
speaking to , what are , what are some

34:50.719 --> 34:52.719
of the sources of our hesitation or

34:52.719 --> 34:55.060
resistance to allow wing Ukraine uh to

34:55.070 --> 34:56.903
employ American provided or NATO

34:56.903 --> 34:59.100
provided weapon systems against valid

34:59.110 --> 35:01.166
military targets on the territory of

35:01.166 --> 35:03.221
the Russian Federation ? Thank you ,

35:03.221 --> 35:05.570
sir . Uh Thank you for that . So first

35:05.639 --> 35:07.870
I want to start by saying and , and I

35:07.879 --> 35:09.823
know many people in this room have

35:09.823 --> 35:11.768
worked on it and I see some , some

35:11.768 --> 35:14.709
folks um in uniform who I've personally

35:14.719 --> 35:16.775
had the chance , who is sitting here

35:16.775 --> 35:18.997
today , who I personally had the chance

35:18.997 --> 35:21.479
to sit with and work through . How do

35:21.489 --> 35:23.929
we deliver for Ukraine , an immense

35:23.939 --> 35:26.179
array of capabilities to defend their

35:26.189 --> 35:28.411
sovereignty and territorial integrity .

35:28.411 --> 35:30.356
And I'm immensely proud of the 2.5

35:30.356 --> 35:32.356
years of effort that we have put in

35:32.356 --> 35:35.610
mobilizing a massive quantity and

35:36.100 --> 35:38.610
extent of capability for Ukraine .

35:39.320 --> 35:43.219
Second , um we do in fact allow the use

35:43.229 --> 35:45.320
of us systems directly across the

35:45.330 --> 35:47.709
border in NATO on the common sense

35:47.719 --> 35:49.540
principle that if you've got an

35:49.550 --> 35:51.550
artillery system firing at you from

35:51.550 --> 35:53.328
across the border in Kharkiv or

35:53.328 --> 35:55.272
elsewhere along that era , Russian

35:55.272 --> 35:57.494
frontier , you ought to be able to fire

35:57.494 --> 35:59.606
back . Uh The question has been about

35:59.810 --> 36:01.850
uh long range strike and the use of

36:01.860 --> 36:04.149
long range strike into Russia and the

36:04.159 --> 36:07.159
position uh of the administration to

36:07.169 --> 36:09.500
date has been not to permit the use of

36:09.510 --> 36:11.489
long range strike into Russia . Uh

36:11.500 --> 36:13.667
President Biden and President Zelensky

36:13.667 --> 36:15.667
have had the opportunity to discuss

36:15.667 --> 36:17.667
this , including most recently when

36:17.667 --> 36:19.611
President Zelensky and he spoke by

36:19.611 --> 36:23.270
telephone last week and um seeing as

36:23.280 --> 36:25.336
how we're not in Chatham House rules

36:25.336 --> 36:27.502
here today and seeing as how I want to

36:27.502 --> 36:29.724
allow those conversations to continue ,

36:29.724 --> 36:31.669
I would just say our policy hasn't

36:31.669 --> 36:33.724
changed at this point . Uh but those

36:33.724 --> 36:35.947
conversations will continue and I think

36:35.947 --> 36:38.560
that they are best uh conducted um in ,

36:38.570 --> 36:41.159
in closed , behind closed doors so that

36:41.169 --> 36:43.336
we can work through the puts and takes

36:43.336 --> 36:45.169
of this issue with our Ukrainian

36:45.169 --> 36:47.290
partners . Thank you . So , before we

36:47.300 --> 36:49.300
go to the next student question , I

36:49.300 --> 36:51.244
wanted to take prerogative here as

36:51.244 --> 36:53.189
moderator and , and ask if I could

36:53.189 --> 36:55.133
follow up , you had mentioned fast

36:55.133 --> 36:58.100
adoption and this has been a perennial

36:58.110 --> 37:00.277
problem with the Department of Defense

37:00.277 --> 37:02.443
in particular . So I thought with this

37:02.443 --> 37:04.666
audience , this might be an opportunity

37:04.666 --> 37:06.777
to address how , how might we do that

37:06.777 --> 37:08.832
in light of some of the delays we've

37:08.832 --> 37:10.943
had with weapons systems and adoption

37:10.943 --> 37:13.054
in the past , if you had any thoughts

37:13.054 --> 37:16.020
on that ? Uh When you say , how might

37:16.030 --> 37:18.252
we do that with when we've had delay in

37:18.252 --> 37:20.419
the past ? I'm , I'm thinking about uh

37:20.419 --> 37:23.060
weapon systems that take 2 to 3 decades

37:23.219 --> 37:25.219
from conception until field the the

37:25.219 --> 37:27.659
actual product . And I was wondering if

37:27.669 --> 37:29.870
with artificial intelligence , there ,

37:29.969 --> 37:32.025
there might be some way to make that

37:32.025 --> 37:33.858
work faster to get that into the

37:33.858 --> 37:36.025
battlefield to get that into the hands

37:36.025 --> 37:37.969
of work . Look , it means it means

37:37.969 --> 37:39.580
dramatically reimagining our

37:39.580 --> 37:41.636
acquisition and procurement system ,

37:41.636 --> 37:43.760
our field , our testing , validating

37:43.770 --> 37:46.639
and fielding system . And it will also

37:46.649 --> 37:49.159
mean what FDR once called bold and

37:49.169 --> 37:51.169
persistent experimentation and risk

37:51.169 --> 37:53.058
taking . And that means all of us

37:53.058 --> 37:55.336
getting a bit out of our comfort zones ,

37:55.336 --> 37:57.502
we have to make sure that we are doing

37:57.502 --> 37:57.110
things in a responsible way . And part

37:57.120 --> 37:59.231
of what the NSM does is set the guard

37:59.231 --> 38:02.330
rails uh and create a framework that

38:02.340 --> 38:04.159
then tells people go innovate

38:04.169 --> 38:07.770
experiment adopt . Um But it's been

38:07.780 --> 38:09.836
interesting just coming back to this

38:09.836 --> 38:12.113
last question about Ukraine to watch a ,

38:12.113 --> 38:16.030
a real live example of

38:16.040 --> 38:18.110
the rapid adoption and iteration of

38:18.120 --> 38:20.064
technology , including A I enabled

38:20.064 --> 38:22.929
technology on the battlefield . And

38:23.110 --> 38:25.830
that has required a constant give and

38:25.840 --> 38:29.250
take an iteration that um each side is

38:29.260 --> 38:32.250
learning in ways that force

38:32.790 --> 38:36.639
us to adopt and alter . For example ,

38:36.649 --> 38:39.949
UAB technology update software , uh

38:39.959 --> 38:42.739
adjust electronic warfare capabilities ,

38:42.879 --> 38:45.459
uh deal with jamming environments in

38:45.469 --> 38:48.750
new and , and distinct ways . And

38:48.760 --> 38:52.209
actually out of the last 2.5 years of

38:52.219 --> 38:54.340
the war against Ukraine , the United

38:54.350 --> 38:57.030
States military has learned an enormous

38:57.040 --> 38:59.096
amount about this and has started to

38:59.096 --> 39:01.262
put some of those lessons to work . So

39:01.262 --> 39:04.020
from my perspective , um this should be

39:04.209 --> 39:06.750
absolutely the kind of textbook we can

39:06.760 --> 39:09.739
then apply to reshape the US military

39:09.750 --> 39:11.583
across the board in terms of its

39:11.583 --> 39:13.639
ability to rapidly adopt and iterate

39:13.639 --> 39:15.639
this technology and what makes this

39:15.639 --> 39:17.750
technology different is it's not just

39:17.750 --> 39:19.750
about the long lead time for uh you

39:19.750 --> 39:21.861
know , a pre precision strike missile

39:21.861 --> 39:24.083
or an aircraft or a surface combatant .

39:24.870 --> 39:27.489
This is about a combination of hardware

39:27.500 --> 39:31.239
and software where the weapon

39:31.250 --> 39:33.250
system as it looks today might look

39:33.250 --> 39:35.472
slightly different a month from now and

39:35.472 --> 39:37.472
then slightly different from that a

39:37.472 --> 39:39.639
month later and so forth . And we need

39:39.639 --> 39:41.694
an entire ecosystem of development ,

39:41.694 --> 39:44.570
procurement and deployment that is

39:44.580 --> 39:46.747
different from how we've thought about

39:46.747 --> 39:48.913
the more static systems that we put to

39:48.913 --> 39:51.024
work today . Uh The good news is that

39:51.024 --> 39:53.247
there is great work underway at the dod

39:53.247 --> 39:54.858
through replicator and other

39:54.858 --> 39:56.858
initiatives to try to do this . But

39:56.858 --> 39:56.600
it's gonna have to be thought of less

39:56.610 --> 39:58.388
as , hey , we have a particular

39:58.388 --> 40:00.443
initiative on technology and more as

40:00.443 --> 40:02.443
this is just the way we do business

40:02.443 --> 40:04.760
across every element of our acquisition

40:04.770 --> 40:07.010
and procurement system . Good . Thank

40:07.020 --> 40:10.729
you , me . Good morning and thank you

40:10.739 --> 40:12.795
very much for your time . My name is

40:12.795 --> 40:15.017
Fiona Saunders . I'm one of the private

40:15.017 --> 40:17.017
sector fellows um at the Eisenhower

40:17.017 --> 40:20.159
School and I am from IBM . Um like many

40:20.169 --> 40:22.113
of my esteemed colleagues from the

40:22.113 --> 40:25.000
private sector . Um We work with our

40:25.179 --> 40:28.209
dod partners and um I was wondering if

40:28.219 --> 40:31.449
you had put any thought into how we

40:31.459 --> 40:34.399
might work maybe through private sector ,

40:34.409 --> 40:36.298
um public private partnerships or

40:36.298 --> 40:38.520
something . So you don't end up with an

40:38.520 --> 40:41.909
instance of A I and um how it's

40:41.919 --> 40:43.959
implemented at the Navy that is

40:43.969 --> 40:45.747
different from the army that is

40:45.747 --> 40:47.913
different from the Marine Corps and um

40:47.913 --> 40:50.025
which wouldn't benefit the taxpayer .

40:50.025 --> 40:52.280
Thank you . Actually , if , if you

40:52.290 --> 40:54.346
think about the second pillar of the

40:54.346 --> 40:56.512
national security memorandum , what it

40:56.512 --> 40:58.734
actually tries to do is put down a road

40:58.734 --> 41:01.429
map that says here's how the national

41:01.439 --> 41:03.790
security enterprise , uh the joint

41:03.800 --> 41:05.744
force , the intelligence community

41:05.744 --> 41:07.522
should work with private sector

41:07.522 --> 41:09.689
partners . And here's how that you can

41:09.689 --> 41:12.229
work in a transparent um effective and

41:12.239 --> 41:15.639
yes legal way uh to make sure that we

41:15.649 --> 41:17.760
are adopting private sector , develop

41:17.760 --> 41:19.538
technologies , capabilities and

41:19.538 --> 41:22.929
solutions into the force into our

41:22.939 --> 41:25.106
intelligence community in a way that's

41:25.106 --> 41:27.870
also shared across the entire national

41:27.879 --> 41:30.429
security enterprise . So the whole uh

41:30.439 --> 41:33.229
design of this second pillar is about

41:33.239 --> 41:35.239
answering your question . How do we

41:35.239 --> 41:37.406
take a solution that IBM has developed

41:37.406 --> 41:39.949
that has applications for war fighting

41:39.959 --> 41:42.015
or for logistics or for intelligence

41:42.015 --> 41:45.270
analysis , incorporate it and then also

41:45.449 --> 41:49.010
make sure that it is um available on a

41:49.020 --> 41:51.500
consistent basis across the board and

41:51.510 --> 41:53.454
that we're not setting up multiple

41:53.454 --> 41:55.454
different competing or inconsistent

41:55.454 --> 41:57.566
solutions . Now , before this NSM was

41:57.566 --> 41:59.732
passed , some of this work was getting

41:59.732 --> 42:02.770
done in a patchwork way by um you know ,

42:02.909 --> 42:05.239
entrepreneurial people within the

42:05.250 --> 42:08.429
various services . But for the first

42:08.439 --> 42:11.030
time , we now have a framework to say

42:11.350 --> 42:14.090
here is a demand signal to industry we

42:14.100 --> 42:16.211
want what you are offering , we wanna

42:16.211 --> 42:18.389
incorporate it rapidly , effectively

42:18.399 --> 42:20.590
comprehensively and in a way that

42:20.600 --> 42:23.310
reduces overlap gaps and , and , and ,

42:23.320 --> 42:26.550
and conflicts . Now , I said at the end

42:26.560 --> 42:28.727
of my remarks , we've set out the plan

42:28.727 --> 42:30.949
now , it's gonna be up to all of you to

42:30.949 --> 42:30.820
implement it , right . The proof will

42:30.830 --> 42:32.663
be in the pudding and the actual

42:32.663 --> 42:34.497
execution of that . But the core

42:34.497 --> 42:36.663
insight that we've come to in the last

42:36.663 --> 42:38.719
couple of years is that we're ahead

42:38.979 --> 42:41.257
when it comes to the latent capability .

42:41.760 --> 42:43.871
The United States has the best latent

42:43.871 --> 42:46.093
capability on A I in the world . How do

42:46.093 --> 42:47.760
we transform that into actual

42:47.760 --> 42:51.010
application on the battlefield

42:51.969 --> 42:55.639
in our logistics in our um uh

42:55.649 --> 42:59.169
intelligence enterprise ? And there it

42:59.179 --> 43:01.401
is going to require that it is going to

43:01.401 --> 43:03.568
have to be , you use the word public ,

43:03.568 --> 43:05.623
private partnership . Often times we

43:05.623 --> 43:07.623
think of that as this kind of , you

43:07.623 --> 43:09.790
know , bespoke thing that's just gonna

43:09.790 --> 43:11.957
have to be the watchword of the entire

43:11.957 --> 43:14.229
effort because this technology has to

43:14.239 --> 43:16.110
such a significant extent , been

43:16.120 --> 43:18.231
developed and advanced in the private

43:18.231 --> 43:20.231
sector . And it's one of the things

43:20.231 --> 43:20.169
that makes A I different and it's the

43:20.179 --> 43:22.235
thing that is gonna have to make our

43:22.235 --> 43:24.290
whole approach different . Thank you

43:24.290 --> 43:26.401
very much for your time . All right .

43:26.401 --> 43:28.729
Thank you , sir . Good morning , sir .

43:28.739 --> 43:30.795
I'm Lieutenant Colonel Blake Jackson

43:30.795 --> 43:33.017
with the United States Marine Corps and

43:33.017 --> 43:35.295
I'm a student at the Eisenhower School .

43:35.295 --> 43:37.461
Uh I would like to further explore the

43:37.461 --> 43:39.406
juxtaposition of collaboration and

43:39.406 --> 43:41.628
competition in light of the US national

43:41.628 --> 43:43.461
security strategy , particularly

43:43.461 --> 43:46.300
regarding competition with China . Are

43:46.310 --> 43:48.489
there markets sectors or capabilities

43:48.500 --> 43:50.556
where cooper operation could enhance

43:50.556 --> 43:53.100
global security without compromising us

43:53.110 --> 43:55.750
strategic interests ? Conversely , in

43:55.760 --> 43:57.982
addition to what sectors you've already

43:57.982 --> 44:00.093
addressed , where else is competition

44:00.093 --> 44:02.149
essential to protect and advance our

44:02.149 --> 44:05.260
national security ? Thank you . So when

44:05.429 --> 44:07.540
President Biden and President Xi met

44:09.149 --> 44:12.530
um for a summit in , in Woodside in San

44:12.540 --> 44:15.250
Francisco about a year ago , they

44:15.260 --> 44:17.482
actually agreed on some areas where the

44:17.482 --> 44:19.482
US and China should work together ,

44:19.482 --> 44:21.760
where our interests align and converge .

44:21.760 --> 44:23.816
I'll give you a couple of examples .

44:23.816 --> 44:25.927
One of them is counter narcotics . Uh

44:25.927 --> 44:27.982
many of the fentaNYL precursors that

44:27.982 --> 44:30.409
flow into Mexico to be formed into the

44:30.419 --> 44:32.530
pills that come into this country and

44:32.530 --> 44:34.530
kill Americans emanate in the first

44:34.530 --> 44:36.810
place from China . And so we wanted to

44:36.820 --> 44:40.600
work with the PR C to place greater uh

44:40.610 --> 44:43.080
restraint and regulation on those

44:43.090 --> 44:46.179
precursors . So that flow was stemmed .

44:46.790 --> 44:48.957
And over the course of the last year ,

44:48.957 --> 44:50.846
we've made progress on that . And

44:50.846 --> 44:52.957
that's thanks to incredible work from

44:52.957 --> 44:55.169
DH S DOJ uh other agencies in the US

44:55.179 --> 44:57.530
government . Uh but there's more work

44:57.540 --> 44:59.484
to do on that front . But it is an

44:59.484 --> 45:01.596
example of where actually a threat to

45:01.596 --> 45:03.596
the United States but also a global

45:03.596 --> 45:05.818
threat . The global threat of synthetic

45:05.818 --> 45:09.320
opioids can be met in part by work

45:09.330 --> 45:11.497
together between the United States and

45:11.497 --> 45:13.719
China climate is another example , John

45:13.719 --> 45:15.590
Podesta , the president's senior

45:15.600 --> 45:17.822
climate envoy was recently in Beijing ,

45:18.300 --> 45:21.629
particularly focused on uh certain

45:21.639 --> 45:23.750
noxious gasses in the work the US and

45:23.750 --> 45:25.750
China can do together . And we have

45:25.750 --> 45:27.972
examples of over the course of the past

45:27.972 --> 45:30.139
decade where we have been able to make

45:30.139 --> 45:32.195
some progress in that area . But our

45:32.195 --> 45:35.290
view all along has been that there is

45:35.300 --> 45:37.522
nothing inconsistent with pursuing that

45:37.522 --> 45:39.300
kind of collaboration where our

45:39.300 --> 45:41.411
interests converge and also competing

45:41.411 --> 45:43.550
vigorously where they do not converge

45:43.860 --> 45:46.027
whether it's in the military domain or

45:46.027 --> 45:47.971
in the technology domain or in the

45:47.971 --> 45:50.027
diplomatic domain where we see China

45:50.027 --> 45:52.449
taking steps that are coercive and

45:52.459 --> 45:54.348
counter to what we see as kind of

45:54.348 --> 45:56.292
universal principles or the United

45:56.292 --> 45:58.515
Nations Charter . So we have to be able

45:58.515 --> 46:00.626
to do both at the same time . And the

46:00.626 --> 46:02.848
president has designed a strategy where

46:02.848 --> 46:05.015
we compete intensively , but we manage

46:05.015 --> 46:07.070
that competition responsibly . So it

46:07.070 --> 46:08.959
doesn't veer into conflict and it

46:08.959 --> 46:11.070
preserves this space to work together

46:11.070 --> 46:13.126
on certain issues . The last point I

46:13.126 --> 46:15.070
would make is all of this requires

46:15.070 --> 46:17.292
intense diplomacy at a leader level and

46:17.292 --> 46:19.459
at all levels below that . And it also

46:19.459 --> 46:21.681
requires extensive military to military

46:21.681 --> 46:24.000
communications . I was in China uh at

46:24.010 --> 46:25.679
the end of August and had the

46:25.689 --> 46:27.578
opportunity to meet with the vice

46:27.578 --> 46:29.489
chairman of the Chinese Military

46:29.500 --> 46:33.000
Commission . Um This is the first

46:33.010 --> 46:35.350
time that an American official has met

46:35.360 --> 46:37.527
with basically the top ranking uniform

46:37.629 --> 46:41.090
in China in nearly a decade . And it

46:41.100 --> 46:44.250
goes to show you that we don't have the

46:44.260 --> 46:46.489
the depth of engagement at a mill mill

46:46.500 --> 46:48.444
level we need . And that's because

46:48.444 --> 46:50.556
China has historically been resistant

46:50.556 --> 46:52.500
to that . The last couple of years

46:52.500 --> 46:52.389
that's opened up quite a bit .

46:52.399 --> 46:54.399
Secretary Austin , Chairman Brown ,

46:54.399 --> 46:56.810
others including Admiral PPAR have had

46:56.820 --> 46:58.598
the opportunity to engage their

46:58.598 --> 47:00.764
counterparts . We need to sustain that

47:00.800 --> 47:03.590
because um the only way that we are

47:03.600 --> 47:05.433
going to effectively manage this

47:05.433 --> 47:08.290
competition is to maintain a thickness

47:08.300 --> 47:10.467
and a frequency of engagement at every

47:10.467 --> 47:12.633
level and particularly with respect to

47:12.633 --> 47:14.800
military , to military communication .

47:15.030 --> 47:17.949
Thank you very much . Uh So before we

47:17.959 --> 47:19.760
go on , I I would ask , do we

47:19.770 --> 47:21.729
prioritize questions on artificial

47:21.739 --> 47:23.795
intelligence ? All of your questions

47:23.795 --> 47:26.090
are , are welcome . But if you have a

47:26.100 --> 47:28.211
question on artificial intelligence ,

47:28.211 --> 47:30.322
uh would you move to the front of the

47:30.322 --> 47:32.760
line , please ? Ok . Uh In that case ,

47:32.770 --> 47:33.800
sir , over to you

47:37.590 --> 47:41.479
what your , your , your all of your

47:41.489 --> 47:43.545
questions are welcome . And I'd just

47:43.545 --> 47:45.211
like to prioritize artificial

47:45.211 --> 47:47.322
intelligence if I could . Thank you .

47:47.322 --> 47:47.100
Good morning , sir , Lieutenant Colonel

47:47.110 --> 47:49.332
Axel Gonzalez , United States Air Force

47:49.332 --> 47:51.554
Intelligence with the mighty College of

47:51.554 --> 47:53.530
Information and cyberspace . In

47:53.540 --> 47:56.199
addition to A I , the holistic

47:56.209 --> 47:59.219
application of the information

47:59.229 --> 48:01.173
instrument of power , what are key

48:01.173 --> 48:03.173
opportunities you see to evolve its

48:03.173 --> 48:05.451
effectiveness in strategic competition ?

48:06.830 --> 48:08.941
A A at least the question said A I in

48:08.949 --> 48:12.600
it So , so that's good , good

48:12.610 --> 48:16.350
start . Um I was uh I was

48:16.360 --> 48:19.580
giving remarks yesterday uh on

48:20.239 --> 48:22.072
uh our foreign economic policy ,

48:22.072 --> 48:23.850
international economics and the

48:23.850 --> 48:25.517
questions were meant to be on

48:25.517 --> 48:27.572
international economics and , and an

48:27.572 --> 48:29.628
enterprising reporter managed to say

48:29.628 --> 48:31.795
you're talking about the international

48:31.795 --> 48:31.739
economy , the Middle East is part of

48:31.750 --> 48:33.861
the international economy . So let me

48:33.861 --> 48:35.861
ask you a question about the Middle

48:35.861 --> 48:38.028
East . It was quite , it was good , it

48:38.028 --> 48:40.520
was good . Um No , but II , I don't

48:40.530 --> 48:42.586
mean to make light of the question ,

48:42.586 --> 48:44.752
which is a really important question .

48:44.752 --> 48:46.974
And we are seeing in real time the ways

48:46.974 --> 48:48.586
in which our competitors and

48:48.586 --> 48:50.808
adversaries use a vertically integrated

48:50.808 --> 48:54.379
strategy of information warfare to

48:54.429 --> 48:56.919
shape global narratives to try to

48:56.929 --> 48:58.739
divide , distract and undermine

48:58.750 --> 49:00.750
confidence in our own country , our

49:00.750 --> 49:02.694
institutions , our democracy , our

49:02.694 --> 49:04.929
staying power , our capacity and we

49:04.939 --> 49:06.939
gotta be on the field to be able to

49:06.939 --> 49:09.229
effectively push back . And I think

49:09.239 --> 49:11.461
this does require a whole of government

49:11.461 --> 49:14.770
approach where uh I see two real

49:14.780 --> 49:17.090
challenges . One of them has to do with

49:17.100 --> 49:20.030
silos Dod has

49:20.040 --> 49:22.179
significant information operations ,

49:22.189 --> 49:24.411
enterprise through the MISO program and

49:24.411 --> 49:26.689
others . Uh uh The State Department has

49:26.699 --> 49:28.949
the GC , the Global Engagement Center ,

49:29.340 --> 49:31.451
other elements of our government have

49:31.451 --> 49:33.639
tools that they use and it's incumbent

49:33.649 --> 49:35.871
upon us . And I would say that we get a

49:35.871 --> 49:38.093
grade of incomplete for trying to bring

49:38.093 --> 49:40.260
those silos together so that we have a

49:40.260 --> 49:42.316
more whole of government approach on

49:42.316 --> 49:44.482
this . And the second piece is that we

49:44.482 --> 49:47.659
have to carry out this effort with

49:47.669 --> 49:49.669
respect to information warfare as a

49:49.669 --> 49:52.800
democracy that doesn't do state run

49:52.810 --> 49:55.060
propaganda and have just arms of

49:55.070 --> 49:57.181
misinformation spread everywhere . We

49:57.181 --> 49:59.348
gotta do it consistent with our values

49:59.348 --> 50:01.379
in our ways . And that may seem at

50:01.389 --> 50:04.419
first blush like that's a disadvantage .

50:05.080 --> 50:08.629
But if you take the capacity and soft

50:08.639 --> 50:11.790
power and communications platform of

50:11.800 --> 50:13.856
every element of the United States ,

50:13.856 --> 50:15.800
our government , our society , our

50:15.800 --> 50:17.744
people , we should be able to more

50:17.744 --> 50:20.750
effectively harness um innovative ways

50:20.760 --> 50:22.871
of communicating that aren't just the

50:22.871 --> 50:25.570
kind of top down info ops

50:26.050 --> 50:29.090
style of doing things , but really take

50:29.100 --> 50:32.729
advantage of um the

50:32.739 --> 50:35.570
creativity and capacity

50:35.679 --> 50:38.469
of uh you know , of the American

50:38.479 --> 50:41.850
information ecosystem . That's at

50:41.860 --> 50:43.790
30,000 ft . How that plays out in

50:43.800 --> 50:45.744
practice is again something that I

50:45.744 --> 50:47.911
think we deserve a great of incomplete

50:47.911 --> 50:50.070
on because um we are not where we

50:50.080 --> 50:51.969
should be on that front , but the

50:51.969 --> 50:53.913
combination of trying to make this

50:53.913 --> 50:56.080
whole of government and then trying to

50:56.080 --> 50:58.310
make it whole of society because we're

50:58.320 --> 51:00.264
just , we have a different mode of

51:00.264 --> 51:02.376
operating with information operations

51:02.376 --> 51:04.153
or information warfare than our

51:04.153 --> 51:06.264
adversaries who are not democracies .

51:06.264 --> 51:08.264
This for me is the way that we will

51:08.264 --> 51:10.598
make progress on this issue . All right ,

51:10.598 --> 51:12.542
ma'am . Good morning . My question

51:12.542 --> 51:14.820
actually relates to what you just said .

51:14.820 --> 51:16.598
Um an say , Chad , I'm with the

51:16.598 --> 51:18.709
National War College State Department

51:18.709 --> 51:18.639
and I come out of the public diplomacy

51:18.649 --> 51:21.560
cone . Uh In recent days , we've seen

51:21.570 --> 51:23.060
uh headlines that Russian

51:23.070 --> 51:25.126
disinformation is once again pouring

51:25.126 --> 51:28.050
into our information ecosystem . Uh and

51:28.060 --> 51:30.227
has the potential to upset the results

51:30.227 --> 51:32.969
of our 2024 presidential election . As

51:32.979 --> 51:36.070
you just said , um research , research

51:36.080 --> 51:37.969
that I've done at the war college

51:37.969 --> 51:40.191
indicates that our interagency response

51:40.191 --> 51:42.080
to information warfare by foreign

51:42.080 --> 51:45.580
actors is insufficient and uh not very

51:45.590 --> 51:48.929
well coordinate . So given

51:48.979 --> 51:51.090
state actors like rush are so good at

51:51.090 --> 51:52.757
this , how can we enhance our

51:52.757 --> 51:55.479
capabilities using influence A I to

51:55.489 --> 51:59.229
both identify disinformation and create

51:59.239 --> 52:01.510
our own narratives . And since you just

52:01.520 --> 52:03.742
talked about breaking down silos , what

52:03.742 --> 52:05.853
do you think about the idea of having

52:05.853 --> 52:07.909
sort of an information czar maybe at

52:07.909 --> 52:09.742
the National Security Council or

52:09.742 --> 52:11.964
another actor that will have sort of an

52:11.964 --> 52:14.131
overall view on protecting information

52:14.131 --> 52:16.409
resiliency in our domestic environment .

52:16.409 --> 52:18.631
It's a good question on the information

52:18.631 --> 52:20.242
Zar one challenge just to be

52:20.242 --> 52:22.409
straightforward is that a lot of these

52:22.409 --> 52:24.689
questions around information operations

52:25.360 --> 52:28.830
um around public diplomacy , around the

52:28.840 --> 52:31.139
voice that America uses to speak to the

52:31.149 --> 52:33.300
world , bleed over into questions of

52:33.310 --> 52:36.949
propaganda or politics ? Um And so the

52:36.959 --> 52:39.181
question is , should the White House be

52:39.181 --> 52:41.292
the central locust for this effort or

52:41.292 --> 52:43.515
should it actually be distributed to an

52:43.515 --> 52:45.570
agency ? Should this be a case where

52:45.570 --> 52:47.515
you have a lead agency model ? One

52:47.515 --> 52:49.737
agency is in charge and they're sort of

52:49.737 --> 52:51.737
running it , but it's kept one step

52:51.737 --> 52:53.903
removed from the White House , I think

52:53.903 --> 52:53.139
on the czar , that's something we've

52:53.149 --> 52:55.427
been grappling with and thinking about ,

52:55.427 --> 52:57.316
um you know , across the national

52:57.316 --> 53:00.070
security enterprise and , and under NSM

53:00.080 --> 53:02.080
two , most decision making does get

53:02.080 --> 53:04.520
kind of brought into the NSC process .

53:04.689 --> 53:06.856
But this may be an area where actually

53:06.856 --> 53:09.078
a lead agency model is a more effective

53:09.078 --> 53:11.719
way of setting up for long term success

53:11.770 --> 53:14.719
that insulates this from the Tos and

53:14.729 --> 53:16.951
fros of politics on both sides . That's

53:16.959 --> 53:19.209
not just by the way about the election

53:19.219 --> 53:21.389
interference issues . It's about the

53:21.399 --> 53:24.399
larger question of how we decide

53:24.409 --> 53:26.353
messages , we communicate , how we

53:26.353 --> 53:29.120
decide what messages we combat , um

53:29.129 --> 53:31.351
which are fraught . Uh It's , and , and

53:31.351 --> 53:33.518
that goes to the point about how we're

53:33.518 --> 53:35.573
a democracy and it's not a situation

53:35.573 --> 53:37.573
where , you know , a small group of

53:37.573 --> 53:39.685
unelected people sitting in a foreign

53:39.685 --> 53:41.796
capital are saying this is what we're

53:41.796 --> 53:43.962
gonna go drive , just go do it and lie

53:43.962 --> 53:46.830
and cheat and , you know , hide and so

53:46.840 --> 53:48.951
forth , we can't do that . So we have

53:48.951 --> 53:50.784
to have something that's kind of

53:50.784 --> 53:52.840
consistent with the political system

53:52.840 --> 53:55.007
that we're dealing with , with respect

53:55.007 --> 53:57.173
to how we get better . I think we have

53:57.173 --> 53:59.396
actually made substantial progress over

53:59.396 --> 53:59.379
the last eight years , but there's a

53:59.389 --> 54:01.620
long way to go from when we fir you

54:01.629 --> 54:03.685
know , the first major election , we

54:03.685 --> 54:07.469
saw significant uh Russian information

54:07.479 --> 54:10.169
operations to today where we see it

54:10.179 --> 54:12.346
from Russia , we see it from Iran , we

54:12.346 --> 54:15.979
see it from other countries and um we

54:15.989 --> 54:19.100
now have systems in place that actually

54:19.110 --> 54:22.409
can pretty rapidly identify deep

54:22.419 --> 54:26.179
fakes and um and call them

54:26.189 --> 54:28.189
out . And we just saw an example of

54:28.189 --> 54:30.245
this over the course of the last few

54:30.245 --> 54:32.550
days where the system worked . I'd like

54:32.560 --> 54:34.449
to shorten the string so it works

54:34.449 --> 54:38.030
faster and it works in more cases .

54:38.040 --> 54:40.250
But I believe that as we exercise this

54:40.260 --> 54:42.427
muscle going forward , we can continue

54:42.427 --> 54:44.593
to make progress . Um We are not where

54:44.593 --> 54:46.649
we need to be . I acknowledge that ,

54:46.649 --> 54:48.760
but I think we are continuing to move

54:48.760 --> 54:50.816
forward in a way where this is a top

54:50.816 --> 54:53.038
flight priority uh for our intelligence

54:53.038 --> 54:55.370
community , for our law enforcement ,

54:55.510 --> 54:59.310
for our NSA and cyber com officials who

54:59.320 --> 55:01.487
are looking at foreign actors engaging

55:01.487 --> 55:03.739
in disruptive cyber attacks as part of

55:03.750 --> 55:06.139
this effort uh to undermine the US

55:06.149 --> 55:08.419
election . And regardless of whether

55:08.429 --> 55:10.540
you're a Republican or Democrat , you

55:10.540 --> 55:12.429
should look at this as a national

55:12.429 --> 55:14.373
security issue as an attack on our

55:14.373 --> 55:16.040
country and think what can we

55:16.040 --> 55:17.870
collectively do to harden our

55:17.879 --> 55:19.823
resilience ? The last point that I

55:19.823 --> 55:21.546
would make is I just think the

55:21.546 --> 55:23.768
awareness a big part of success here is

55:23.768 --> 55:27.260
not just stopping every tweet

55:27.270 --> 55:29.820
or video we should do as much of that

55:29.830 --> 55:31.941
as possible . It's raising resilience

55:31.941 --> 55:33.997
by raising awareness , having people

55:33.997 --> 55:35.941
understand when they're looking at

55:35.941 --> 55:38.219
something or consuming something . Hey ,

55:38.219 --> 55:40.386
this very well could be have a foreign

55:40.386 --> 55:42.659
actor provenance to it . I think we've

55:42.669 --> 55:44.780
made progress on that , but there's a

55:44.780 --> 55:46.947
long way to go to get to where we need

55:46.947 --> 55:49.058
to be . Do you think we have time for

55:49.060 --> 55:51.282
one or two more questions ? Sure . Ok ,

55:51.282 --> 55:55.020
excellent . Uh please . Uh

55:55.030 --> 55:57.308
morning sir , Colonel Paul Albuquerque ,

55:57.308 --> 55:59.419
United States Air Force from National

55:59.419 --> 56:01.949
War College . Uh on the topic of A I

56:01.959 --> 56:04.070
today , our digital shadow , the data

56:04.070 --> 56:06.126
that we just put out from using apps

56:06.270 --> 56:08.760
and hardware is collected by companies .

56:09.050 --> 56:11.669
Uh And it's a major industry for them

56:12.110 --> 56:15.840
um with the abilities of A I to uh

56:15.850 --> 56:18.429
collect that data and then potentially

56:18.439 --> 56:21.129
use it to influence customers . The

56:21.139 --> 56:23.083
American population . What sort of

56:23.083 --> 56:25.370
guardrails or restrictions do you think

56:25.379 --> 56:27.323
need to be put in place with those

56:27.323 --> 56:31.030
corporations to ensure privacy and uh

56:31.790 --> 56:34.360
minimizing undue influence on the US

56:34.370 --> 56:36.426
population ? Uh Look , I think there

56:36.426 --> 56:38.481
has to be an element of transparency

56:38.481 --> 56:40.537
and choice , transparency about what

56:40.780 --> 56:43.600
the data privacy uh

56:43.610 --> 56:46.770
parameters are with a given A I model .

56:46.780 --> 56:49.879
Customers should understand what their

56:49.889 --> 56:53.459
data is being used for and how uh it is

56:53.469 --> 56:56.110
being used with respect to them or

56:56.449 --> 56:59.010
bundled up and sold off or for that

56:59.020 --> 57:01.850
matter , sold to foreign competitors .

57:01.860 --> 57:03.971
And in fact , we have a data security

57:03.971 --> 57:07.179
rule making its way uh into final form

57:07.189 --> 57:08.911
right now to deal with certain

57:08.911 --> 57:12.070
sensitive data that historically has

57:12.080 --> 57:14.302
been able to go to countries of concern

57:14.302 --> 57:16.469
like the PR C that should not be going

57:16.469 --> 57:19.120
there . But on the more day to day

57:19.139 --> 57:20.917
manifestation of this as you're

57:20.917 --> 57:22.806
describing , we should start with

57:22.806 --> 57:24.861
people understanding and then having

57:24.861 --> 57:27.649
choices about what data that they wanna

57:27.659 --> 57:30.050
share and not share . And we have with

57:30.060 --> 57:33.370
respect to um cookies on

57:33.379 --> 57:36.699
websites or um other data being

57:36.709 --> 57:39.540
collected by uh technology companies

57:39.550 --> 57:41.717
some experience now over the course of

57:41.717 --> 57:43.883
the past few years and we just need to

57:43.883 --> 57:46.409
adapt that in my view from kind of

57:46.419 --> 57:50.409
existing online platforms to frontier

57:50.419 --> 57:54.239
A I models or the API S of A of a chat

57:54.250 --> 57:56.489
GP T or what have you . And I think

57:56.500 --> 57:58.667
that there is a common sense way to do

57:58.667 --> 58:00.556
this where you can take advantage

58:00.556 --> 58:02.611
effectively of this technology , but

58:02.611 --> 58:04.778
also have some agency in ensuring your

58:04.778 --> 58:06.833
data is not used in ways that you're

58:06.833 --> 58:09.870
not comfortable with . So we're running

58:09.879 --> 58:11.879
low on time , but I think I'm gonna

58:11.879 --> 58:13.990
cheat and I'm gonna ask if the next ,

58:13.990 --> 58:16.157
uh , the both of you could please give

58:16.157 --> 58:18.157
your questions and this will be our

58:18.157 --> 58:19.990
last question for the day . Good

58:19.990 --> 58:22.212
morning . I'm Lieutenant Colonel Kelsey

58:22.212 --> 58:24.435
Rouse , the United States Army from the

58:24.435 --> 58:24.330
Eisenhower school and pulling again on

58:24.340 --> 58:27.040
the ethics piece on A I . Uh you

58:27.050 --> 58:29.106
briefly touch it on the previous two

58:29.106 --> 58:31.629
questions . So how does the American

58:31.639 --> 58:33.750
public along with strategic leaders

58:33.760 --> 58:35.649
properly had the discussion about

58:35.649 --> 58:37.709
ethics when A I right now , when we

58:37.719 --> 58:39.886
talk about certain types of controls ,

58:39.886 --> 58:41.441
it can be seen as an actual

58:41.441 --> 58:43.552
infringement on civil rights within a

58:43.552 --> 58:47.280
free democracy . Uh That is actually

58:47.290 --> 58:49.401
very similar to my question as well .

58:49.401 --> 58:51.568
So we'll just go with that one . Thank

58:51.568 --> 58:53.623
you . Oh , wow . Ok , that's great .

58:53.689 --> 58:55.467
Well , look , a big part of the

58:55.467 --> 58:57.245
National Security memorandum is

58:57.245 --> 58:59.219
actually um setting out some basic

58:59.229 --> 59:01.396
principles that , you know , we didn't

59:01.396 --> 59:03.618
pull out of thin air that have been the

59:03.618 --> 59:07.270
product of years of work now on bias

59:07.280 --> 59:10.280
and discrimination on privacy , on

59:10.379 --> 59:13.199
civil liberties , on safety and

59:13.209 --> 59:15.760
trustworthiness . And that work now

59:15.770 --> 59:18.189
manifests itself uh in a number of

59:18.199 --> 59:20.032
different guidelines that the US

59:20.032 --> 59:22.729
government has put forward . Um and not

59:22.739 --> 59:24.739
just put forward ourselves but then

59:24.739 --> 59:26.860
tried to work with other countries in

59:26.870 --> 59:29.199
kind of concentric circles , the G7 ,

59:29.209 --> 59:31.739
our closest partners , um you know , a

59:31.750 --> 59:33.917
larger group that came together at the

59:33.917 --> 59:35.917
Bletchley and Seoul A I summits and

59:35.917 --> 59:38.028
ultimately the UN General Assembly to

59:38.028 --> 59:40.320
try to set out a series of principles

59:40.330 --> 59:42.629
for how this should be uh effectively

59:42.639 --> 59:44.861
used . And then there's the question of

59:44.861 --> 59:46.972
accountability and oversight and part

59:46.972 --> 59:46.550
of what the National Security

59:46.560 --> 59:49.489
memorandum does is say it's not good

59:49.500 --> 59:51.556
enough to say this is how A I should

59:51.556 --> 59:53.333
and shouldn't be used . We need

59:53.333 --> 59:55.850
mechanisms to enforce that within our

59:55.860 --> 59:57.709
government , whether it's the

59:57.719 --> 59:59.719
responsible use of A I for military

59:59.719 --> 01:00:02.040
purposes or for intelligence gathering

01:00:02.050 --> 01:00:04.389
purposes or intelligence analysis

01:00:04.399 --> 01:00:07.370
purposes or uh whether it's something

01:00:07.379 --> 01:00:09.212
as simple as just the commercial

01:00:09.212 --> 01:00:11.323
application of A I and how each of us

01:00:11.323 --> 01:00:13.046
should feel safe and secure as

01:00:13.046 --> 01:00:15.212
Americans that we are not being harmed

01:00:15.212 --> 01:00:18.770
through its use . So basically , we

01:00:18.780 --> 01:00:20.947
have built over the course of the last

01:00:20.947 --> 01:00:23.300
couple of years , I think a pretty good

01:00:23.310 --> 01:00:26.659
set of markers , right and left rules

01:00:26.669 --> 01:00:28.725
of the road that try to get at these

01:00:28.725 --> 01:00:30.919
questions of fundamental values . And

01:00:30.929 --> 01:00:32.707
then what the National Security

01:00:32.707 --> 01:00:34.899
memorandum tries to do is say , um

01:00:35.080 --> 01:00:39.040
here's our guidelines . Uh Now here's

01:00:39.050 --> 01:00:41.106
how they have to be implemented with

01:00:41.106 --> 01:00:43.217
special care in the national security

01:00:43.217 --> 01:00:45.383
space because of the unique risks that

01:00:45.383 --> 01:00:47.606
come with the national security space .

01:00:47.606 --> 01:00:47.209
And here's how we need to set up

01:00:47.219 --> 01:00:49.108
oversight and accountability in a

01:00:49.108 --> 01:00:51.163
transparent way to make sure that we

01:00:51.163 --> 01:00:53.386
are holding ourselves accountable . And

01:00:53.386 --> 01:00:55.552
I can tell you when we were talking to

01:00:55.552 --> 01:00:57.386
the President about the National

01:00:57.386 --> 01:00:59.608
Security memorandum and we had multiple

01:00:59.608 --> 01:00:59.520
meetings with him about this because he

01:00:59.530 --> 01:01:01.863
didn't just take it and say , ok , fine ,

01:01:01.863 --> 01:01:04.086
we went back and forth on this over the

01:01:04.086 --> 01:01:06.363
course of the past two or three months .

01:01:06.363 --> 01:01:08.530
This was one of his questions . How do

01:01:08.530 --> 01:01:10.863
we really make sure on a rigorous basis ?

01:01:10.863 --> 01:01:13.086
Not just that we've set forth standards

01:01:13.086 --> 01:01:15.252
that make sense to us , but how can we

01:01:15.252 --> 01:01:15.090
have confidence through thick and

01:01:15.100 --> 01:01:17.070
through thin through crisis and

01:01:17.080 --> 01:01:19.302
contingency , they're actually going to

01:01:19.302 --> 01:01:21.580
be enforced and not run roughshod over .

01:01:21.580 --> 01:01:23.691
And that is a big part of what this N

01:01:23.691 --> 01:01:25.747
SS M is trying to achieve . The last

01:01:25.747 --> 01:01:29.300
point is I talked a lot about the

01:01:29.310 --> 01:01:31.659
uncertainty of this technology . So you

01:01:31.669 --> 01:01:35.419
can have principles around human

01:01:35.429 --> 01:01:38.870
oversight , around privacy , around

01:01:39.389 --> 01:01:43.100
uh bias or discrimination

01:01:43.350 --> 01:01:45.840
today . And the technology can involve

01:01:45.850 --> 01:01:48.600
in ways that make what you've said

01:01:48.610 --> 01:01:50.709
today seems somewhat anachronistic

01:01:50.719 --> 01:01:53.149
tomorrow . So the other thing that

01:01:53.159 --> 01:01:55.550
we've done is built in the need to

01:01:55.560 --> 01:01:58.060
constantly update and adapt this as we

01:01:58.070 --> 01:02:01.300
go , just as we need to adapt a drone

01:02:01.310 --> 01:02:03.532
on a battlefield , we need to adapt the

01:02:03.532 --> 01:02:05.643
values and guardrails side of this as

01:02:05.643 --> 01:02:07.754
well to keep pace with the technology

01:02:07.754 --> 01:02:09.643
and the lessons we learn from the

01:02:09.643 --> 01:02:12.229
evolution of that technology . Well ,

01:02:12.239 --> 01:02:14.461
thank you . Thank you very much and I'd

01:02:14.461 --> 01:02:16.350
like to close by inviting Admiral

01:02:16.350 --> 01:02:18.850
Garvin to uh closing remarks . Thank

01:02:18.860 --> 01:02:19.860
you . Thank you .

01:02:28.409 --> 01:02:30.631
I'll be exceptionally brief . Uh Just a

01:02:30.631 --> 01:02:32.853
quick , thank you again , sir . We , we

01:02:32.853 --> 01:02:35.076
all can imagine what your schedule must

01:02:35.076 --> 01:02:37.076
be like and so the ability to spend

01:02:37.076 --> 01:02:39.020
some time with you this morning is

01:02:39.020 --> 01:02:38.949
absolutely priceless . So thank you

01:02:39.080 --> 01:02:40.969
another round of applause for the

01:02:40.969 --> 01:02:42.080
national security .

01:02:56.510 --> 01:02:59.270
Ok . There's a

01:03:04.000 --> 01:03:05.000
thank you .

01:03:09.020 --> 01:03:09.399
Why was that ?