WEBVTT
00:00:17.679 --> 00:00:20.940
Hello, this is Paul Saunders. Thank you for joining
00:00:20.940 --> 00:00:24.500
us again for another episode of the Three Questions
00:00:24.500 --> 00:00:28.719
podcast. I'm really delighted to have with me
00:00:28.719 --> 00:00:33.560
today Fergus Hay. Fergus is the co -founder and
00:00:33.560 --> 00:00:37.200
chief executive officer of The Hacking Games,
00:00:37.340 --> 00:00:40.799
and I'll ask Fergus to describe that in a minute
00:00:40.799 --> 00:00:44.140
because I'm sure he'll do a better job than I,
00:00:44.259 --> 00:00:47.670
but it's very cool. And he's also quite active
00:00:47.670 --> 00:00:51.429
in the defense and technology space as director
00:00:51.429 --> 00:00:54.130
at the Defense and Security Forum, which is a
00:00:54.130 --> 00:00:58.990
defense think tank and a co -founder. of a firm
00:00:58.990 --> 00:01:03.450
called Elysian Fields, which is working as kind
00:01:03.450 --> 00:01:06.469
of an advisor in the venture capital and private
00:01:06.469 --> 00:01:09.689
equity area. So, Fergus, thank you so much for
00:01:09.689 --> 00:01:12.689
being with me today. And first, could I just
00:01:12.689 --> 00:01:15.969
ask you to explain the hacking games, which,
00:01:15.989 --> 00:01:18.930
you know, I was very excited to hear about when
00:01:18.930 --> 00:01:21.150
you first explained it to me, because it's like
00:01:21.150 --> 00:01:25.209
something you would see in a movie. Please tell
00:01:25.209 --> 00:01:28.060
me. Please tell me. Paul, first of all, it's
00:01:28.060 --> 00:01:30.859
a real privilege and delight to be invited, not
00:01:30.859 --> 00:01:32.599
only by you, but also by the national interest
00:01:32.599 --> 00:01:37.489
to share our mission and the major. global and
00:01:37.489 --> 00:01:40.049
social challenges that we think we are facing.
00:01:40.250 --> 00:01:42.829
So the Hacking Games was set up three years ago.
00:01:43.010 --> 00:01:46.230
Our vision is to create a generation of ethical
00:01:46.230 --> 00:01:48.969
hackers to make the world safer. And the reason
00:01:48.969 --> 00:01:51.549
why we think that's really important is that
00:01:51.549 --> 00:01:53.750
in the geopolitical world we're facing today,
00:01:53.989 --> 00:01:58.209
you've got the convergence of geopolitics, national
00:01:58.209 --> 00:02:02.090
security and defense threats. converging with
00:02:02.090 --> 00:02:05.629
cybercrime and the penetration of destabilization
00:02:05.629 --> 00:02:09.569
of governments, societies, companies, and citizens.
00:02:09.990 --> 00:02:11.430
And what's really, really interesting in that
00:02:11.430 --> 00:02:13.569
convergence is for so many years, we've been
00:02:13.569 --> 00:02:16.310
focused on cybercrime. And when you look at the
00:02:16.310 --> 00:02:19.129
scale of cybercrime, it's pretty spectacular.
00:02:19.270 --> 00:02:22.430
In fact, last year, it was forecasted to cost
00:02:22.430 --> 00:02:25.710
the world $10 .5 trillion. If that number is
00:02:25.710 --> 00:02:28.689
accurate at all, that is... the global cost of
00:02:28.689 --> 00:02:32.550
COVID according to the IMF. Next year, it's forecasted
00:02:32.550 --> 00:02:36.270
to cost the world $24 trillion. And if that number
00:02:36.270 --> 00:02:39.870
is accurate, that's the GDP of America. So what's
00:02:39.870 --> 00:02:44.050
really striking is that cybercrime is now converging
00:02:44.050 --> 00:02:46.229
with geopolitical and national security issues,
00:02:46.530 --> 00:02:49.729
because it's no longer a crime for personal cybergang
00:02:49.729 --> 00:02:53.370
benefit, but as you can see from the hack on
00:02:53.370 --> 00:02:56.050
Jaguar Land Rover. that cost two and a half billion
00:02:56.050 --> 00:02:58.569
pounds, or the hack on the largest retailers
00:02:58.569 --> 00:03:02.349
in Europe, which cost $600 million. These were
00:03:02.349 --> 00:03:05.210
all activated by hostile state actors, in those
00:03:05.210 --> 00:03:08.590
case, Russia, to destabilize NATO governments,
00:03:08.849 --> 00:03:12.389
to drive a fragmentation of the alliance, while
00:03:12.389 --> 00:03:14.409
governments focus on their own internal issues
00:03:14.409 --> 00:03:17.110
to restore stability and do their number one
00:03:17.110 --> 00:03:19.050
role, which is to protect their own citizens.
00:03:19.430 --> 00:03:23.039
So in this convergence of cyber attacks, with
00:03:23.039 --> 00:03:25.139
national security and defense, you have this
00:03:25.139 --> 00:03:28.800
new world of kind of hybrid warfare where cybercrime
00:03:28.800 --> 00:03:31.800
or cyber weaponization is the easiest thing and
00:03:31.800 --> 00:03:34.340
lowest cost thing to do than to launch a missile.
00:03:34.659 --> 00:03:38.620
And so we started an organization in order to
00:03:38.620 --> 00:03:41.219
identify who are the people being weaponized
00:03:41.219 --> 00:03:43.960
and what can we do to identify and harness the
00:03:43.960 --> 00:03:47.800
most talented hackers in the world to not destroy
00:03:47.800 --> 00:03:50.740
society, but... protect our national security,
00:03:51.080 --> 00:03:53.319
our private enterprises, our critical infrastructure,
00:03:53.500 --> 00:03:57.139
and our communities. And this is where the cut
00:03:57.139 --> 00:03:59.080
-through really comes, because when you dive
00:03:59.080 --> 00:04:02.060
into it, what you realize is the hackers who
00:04:02.060 --> 00:04:04.379
are committing these enormous destabilizations
00:04:04.379 --> 00:04:07.900
are children. In fact, the world's largest research
00:04:07.900 --> 00:04:11.620
report commissioned by the EU interviewed 9 ,000
00:04:11.620 --> 00:04:16.300
teenagers across the EU and the UK, and 69 %
00:04:16.300 --> 00:04:20.750
of them have committed a cybercrime. 69%, 45
00:04:20.750 --> 00:04:24.250
% of them have committed an act of serious financial
00:04:24.250 --> 00:04:27.129
crime. So that's like identity theft and crypto
00:04:27.129 --> 00:04:30.829
fraud and banking currency fraud. 45%, that's
00:04:30.829 --> 00:04:33.829
12 to 14 kids out of an average high school classroom.
00:04:34.089 --> 00:04:36.189
And so you've got the weaponization of young
00:04:36.189 --> 00:04:39.670
people by hostile state actors to commit cyber
00:04:39.670 --> 00:04:42.689
criminal acts. on behalf of destabilizing nations.
00:04:42.910 --> 00:04:45.170
And we see that as the major social problem,
00:04:45.310 --> 00:04:48.069
a major cyber problem, and a major national security
00:04:48.069 --> 00:04:50.769
problem. And so we're focusing our efforts on
00:04:50.769 --> 00:04:54.310
identifying elite Gen Z hackers and ensuring
00:04:54.310 --> 00:04:56.689
they use their skills to protect society, not
00:04:56.689 --> 00:04:59.420
damage society. Thanks very much, Fergus. That's
00:04:59.420 --> 00:05:02.040
definitely way better than I could have done.
00:05:02.180 --> 00:05:05.040
So, you know, as I'm kind of listening to you
00:05:05.040 --> 00:05:07.980
there and, you know, you touched on the geopolitics,
00:05:07.980 --> 00:05:11.279
you touched on some of the things that are happening
00:05:11.279 --> 00:05:15.019
that it kind of leads into my first question,
00:05:15.060 --> 00:05:18.240
which would be just as you look back yourself
00:05:18.240 --> 00:05:22.379
and kind of reflect on this field. What changes
00:05:22.379 --> 00:05:25.500
or evolution do you see? You know, we've got
00:05:25.500 --> 00:05:29.860
obviously AI, we've got the geopolitics again,
00:05:29.939 --> 00:05:32.319
which you touched on. There are a lot of other
00:05:32.319 --> 00:05:36.040
things happening globally, economically in society.
00:05:36.199 --> 00:05:38.740
I'm just wondering how you see hacking changing.
00:05:38.860 --> 00:05:41.800
And to me, it kind of feels like this very weird
00:05:41.800 --> 00:05:45.560
moment in world history where from a certain
00:05:45.560 --> 00:05:49.740
perspective, people globally are safer and better
00:05:49.740 --> 00:05:54.040
off and more secure and prosperous than maybe
00:05:54.040 --> 00:05:57.399
we've ever been in thousands of years of human
00:05:57.399 --> 00:06:00.339
civilization. But at the same time, there's like
00:06:00.339 --> 00:06:03.120
this incredible sense of vulnerability coming
00:06:03.120 --> 00:06:06.259
from the internet and some of these other tools.
00:06:06.420 --> 00:06:09.019
Could you help me? How do you see all of that?
00:06:09.279 --> 00:06:11.600
Well, according to the hyperscalers over the
00:06:11.600 --> 00:06:13.740
weekend, we understand we've got a 10 % chance
00:06:13.740 --> 00:06:18.480
of total annihilation of civilization. So I think
00:06:18.480 --> 00:06:20.980
if you were to believe that... That's nice. That's
00:06:20.980 --> 00:06:25.439
nice. Good. Yeah. If you were to believe that,
00:06:25.579 --> 00:06:28.560
and let's suspend disbelief for a second and
00:06:28.560 --> 00:06:30.100
not think that people are just trying to pump
00:06:30.100 --> 00:06:33.079
their IPO price, you would argue that we've never
00:06:33.079 --> 00:06:35.699
been more vulnerable than we are today. I think
00:06:35.699 --> 00:06:38.160
the first thing we should do is dissect what
00:06:38.160 --> 00:06:41.519
hacking is, because already there you've used
00:06:41.519 --> 00:06:43.959
it as a negative term. But I think we have to
00:06:43.959 --> 00:06:45.560
understand there is a difference between hacking
00:06:45.560 --> 00:06:49.560
and hacking for harm. Now, hacking in its original
00:06:49.560 --> 00:06:53.000
definition is evaluating a construct, breaking
00:06:53.000 --> 00:06:55.879
it down into its component pieces, and identifying
00:06:55.879 --> 00:06:58.959
either where it's vulnerable or how it can be
00:06:58.959 --> 00:07:02.439
reused for another use. Now, based on that definition,
00:07:02.879 --> 00:07:05.939
It's actually the original source of human creativity.
00:07:06.079 --> 00:07:08.740
Without a hacking mindset, we'd be riding horses
00:07:08.740 --> 00:07:11.439
to work, or we'd be reading by candlelight, we'd
00:07:11.439 --> 00:07:14.139
be feeding ourselves by coal. So the question
00:07:14.139 --> 00:07:18.360
is, in the hands of whom to do what? Now, the
00:07:18.360 --> 00:07:20.720
evolution of hacking is worth understanding.
00:07:21.240 --> 00:07:25.240
Back in the 1970s, the original hackers were
00:07:25.240 --> 00:07:29.279
incredibly bright, highly niche, highly focused
00:07:29.279 --> 00:07:32.360
people who were... dustbin diving, you know,
00:07:32.379 --> 00:07:35.500
diving into dustbins to build computers, to understand
00:07:35.500 --> 00:07:38.680
constructs, to kind of find weaknesses and evolve
00:07:38.680 --> 00:07:41.259
them. Tim Berners -Lee, the founder of the World
00:07:41.259 --> 00:07:44.100
Wide Web, is a self -processed hacker. And this
00:07:44.100 --> 00:07:46.480
was often the baseline of amazing innovation.
00:07:47.000 --> 00:07:51.370
What grew out of that in the 90s was... bad actors
00:07:51.370 --> 00:07:54.649
understanding that they could use this to violate
00:07:54.649 --> 00:07:57.250
the global connectivity that was happening through
00:07:57.250 --> 00:08:00.209
the internet. And that's where you started to
00:08:00.209 --> 00:08:03.750
see a kind of nihilistic, anachronistic rise
00:08:03.750 --> 00:08:07.649
of kind of breaking things. And then in the 2000s,
00:08:07.649 --> 00:08:10.350
as the kind of one banking system hit the world,
00:08:10.490 --> 00:08:13.220
people understood, oh, wow. I can hack into this
00:08:13.220 --> 00:08:16.759
nascent banking system and steal. And so that's
00:08:16.759 --> 00:08:20.019
where you get the rise of, as always, early adopters
00:08:20.019 --> 00:08:22.639
of new techniques on the criminal world because
00:08:22.639 --> 00:08:25.120
they understand that they don't need ethics committees
00:08:25.120 --> 00:08:27.459
to make decisions on the adoption of technology.
00:08:27.740 --> 00:08:30.000
They just apply it really quickly for their own
00:08:30.000 --> 00:08:32.600
personal gain. So you saw money laundering and
00:08:32.600 --> 00:08:36.240
theft and fraud and scamming. What's changed
00:08:36.240 --> 00:08:41.610
now is two things. Number one, you have a generation
00:08:41.610 --> 00:08:44.730
who are natural -born technologists. I'm technically
00:08:44.730 --> 00:08:47.750
a millennial, so I was a natural -born internet
00:08:47.750 --> 00:08:50.289
user. I don't remember the world pre -internet.
00:08:50.870 --> 00:08:53.610
There's this current young generation, Gen Z,
00:08:53.669 --> 00:08:57.250
age 12 to 27. They don't know the world really
00:08:57.250 --> 00:09:01.309
before AI. They're so native that they see AI
00:09:01.309 --> 00:09:04.269
as just a natural vein in their body. They only
00:09:04.269 --> 00:09:06.789
know how to adopt it and utilize it, and I just
00:09:06.789 --> 00:09:09.230
see it in my children. So what that means is
00:09:09.230 --> 00:09:13.370
the gap, between human creativity and the technology
00:09:13.370 --> 00:09:17.750
to deploy it and implement it is zero. The adoption
00:09:17.750 --> 00:09:22.529
levels are 100%. There isn't a barrier to entry
00:09:22.529 --> 00:09:25.570
for the deployment of these products. And so
00:09:25.570 --> 00:09:28.610
we see this kind of convergence of two behaviors.
00:09:28.789 --> 00:09:32.769
At the very top tier, we see elite -level computer
00:09:32.769 --> 00:09:35.720
scientists who have been working with data. When
00:09:35.720 --> 00:09:38.039
it was cool, it was called big data. Then it
00:09:38.039 --> 00:09:41.500
became machine learning. Now it becomes AI. And
00:09:41.500 --> 00:09:43.720
what you're seeing is the maturity development
00:09:43.720 --> 00:09:46.259
and deployment of those models at enormous scale
00:09:46.259 --> 00:09:49.220
with the horsepower of the capital markets to
00:09:49.220 --> 00:09:51.720
drive it. So that's anthropic. Level down, you're
00:09:51.720 --> 00:09:53.860
seeing these insanely powerful models that are
00:09:53.860 --> 00:09:57.399
being deployed in the cybersecurity sector to
00:09:57.399 --> 00:10:01.360
do cybersecurity penetration testing at enormous
00:10:01.360 --> 00:10:04.379
scale. And that is worrying. We see that in the
00:10:04.379 --> 00:10:06.720
media. You're seeing a whole generation of young
00:10:06.720 --> 00:10:12.139
people adopting and absorbing AI tools for their
00:10:12.139 --> 00:10:14.399
basic daily lives. And that's going to create
00:10:14.399 --> 00:10:19.240
a generation of innovators and AI users to create
00:10:19.240 --> 00:10:22.779
further innovations and a leapfrog of innovation,
00:10:22.960 --> 00:10:25.720
technology, and wealth creation. And Elon Musk
00:10:25.720 --> 00:10:28.340
thinks it'll add 30 % to the global GDP, $30
00:10:28.340 --> 00:10:31.960
trillion of opportunity, which could may well
00:10:31.960 --> 00:10:36.990
be the case. hyperbolic headlines and such as
00:10:36.990 --> 00:10:40.570
we've got a 10 chance of total nihilism um they
00:10:40.570 --> 00:10:43.990
latch on to a tropics model and escaping the
00:10:43.990 --> 00:10:47.730
sandbox and hacking organizations but the truth
00:10:47.730 --> 00:10:51.110
is as always not as extreme on either side what
00:10:51.110 --> 00:10:54.610
we do know for sure is that side the cyber world
00:10:54.970 --> 00:10:58.309
is really the second category that has deployed
00:10:58.309 --> 00:11:01.909
cyber at such enormous scale, deployed AI at
00:11:01.909 --> 00:11:03.929
such enormous scale. And that will have both
00:11:03.929 --> 00:11:06.850
a threat to society, but ultimately, once it's
00:11:06.850 --> 00:11:08.629
adopted and deployed in an intelligent fashion,
00:11:08.750 --> 00:11:12.090
will enable stronger defenses of society. You're
00:11:12.090 --> 00:11:14.789
working a lot, you know, with governments and
00:11:14.789 --> 00:11:18.169
certainly also with companies. And I guess I
00:11:18.169 --> 00:11:21.370
wonder if you try to put on, you know, the hat
00:11:21.370 --> 00:11:26.169
of a government official or a corporate executive,
00:11:26.509 --> 00:11:29.309
you know, trying to navigate all of this. What
00:11:29.309 --> 00:11:32.029
do you think are the biggest challenges for people
00:11:32.029 --> 00:11:34.610
like that? The biggest problems they face? What
00:11:34.610 --> 00:11:37.129
are the biggest threats they see? Well, I think
00:11:37.129 --> 00:11:39.570
that, as always, the biggest problems are lurching
00:11:39.570 --> 00:11:43.090
into hyperbole. So, you know, actions without
00:11:43.090 --> 00:11:47.190
strategy cause reduction in effectiveness. So
00:11:47.190 --> 00:11:51.629
I think as leaders of organizations, before leaping
00:11:51.629 --> 00:11:54.029
into adopting AI blindly, it needs to be within
00:11:54.029 --> 00:11:56.669
strategy for productive, sustainable growth of
00:11:56.669 --> 00:12:00.850
the organization. I think that people have moved
00:12:00.850 --> 00:12:03.970
on now from it being a cost -cutting exercise.
00:12:05.039 --> 00:12:07.340
and understanding that it's a productivity exercise.
00:12:07.840 --> 00:12:10.320
And people are also understanding the limitations
00:12:10.320 --> 00:12:13.000
of AI. We have a really strong view on this.
00:12:13.139 --> 00:12:15.259
You know, we think that the deployment of AI
00:12:15.259 --> 00:12:18.419
is most visible, sadly, in media and marketing.
00:12:18.830 --> 00:12:21.029
Just cruise your LinkedIn channel and you will
00:12:21.029 --> 00:12:24.309
see what happens when people deploy mediocre
00:12:24.309 --> 00:12:28.289
AI at scale without any genuine human insight
00:12:28.289 --> 00:12:31.309
and ingenuity. You just simply get mass mediocrity
00:12:31.309 --> 00:12:33.950
or slot. And I think the world is kind of overrun
00:12:33.950 --> 00:12:36.389
with it at the moment. And it's super boring.
00:12:36.789 --> 00:12:40.769
I mean, it's super boring and not engaging or
00:12:40.769 --> 00:12:44.009
interesting. There's no content. It's very vacuous.
00:12:44.210 --> 00:12:47.370
I mean, I completely agree. It's reductive. We're
00:12:47.370 --> 00:12:51.330
all becoming more mediocre. We couldn't be converging
00:12:51.330 --> 00:12:53.950
under the middle of the curve faster in the deployment
00:12:53.950 --> 00:12:57.789
of mediocre AI. However, what's really interesting
00:12:57.789 --> 00:13:01.129
is when you take a brilliant mind, what we call
00:13:01.129 --> 00:13:03.450
a corkscrew mind, the kind of minds that Winston
00:13:03.450 --> 00:13:07.789
Churchill in 1940 asked Alan Turing to find to
00:13:07.789 --> 00:13:10.090
break the Enigma code, people who think differently.
00:13:10.409 --> 00:13:12.450
And as you know, Alan Turing put cryptographic
00:13:12.450 --> 00:13:15.110
puzzles into the newspaper, said if you can crack
00:13:15.110 --> 00:13:18.200
this... And you can join Bletchley Park and these
00:13:18.200 --> 00:13:21.000
elite level of thinkers to break this enormous
00:13:21.000 --> 00:13:23.539
cryptographic puzzle in the Enigma code. And
00:13:23.539 --> 00:13:25.860
that was the basis of cryptography and cybersecurity
00:13:25.860 --> 00:13:29.340
today. So if you take a corkscrew mind, a mind
00:13:29.340 --> 00:13:31.220
that thinks differently, what today we would
00:13:31.220 --> 00:13:34.019
call a neurodivergent or neurobrilliant mind,
00:13:34.200 --> 00:13:36.559
and you have them develop incredibly insightful
00:13:36.559 --> 00:13:40.960
concepts and you use AI to scale that, you get
00:13:40.960 --> 00:13:43.480
brilliant outcomes. And I think the world has
00:13:43.480 --> 00:13:46.720
leapt into... mediocrity at scale through deployment
00:13:46.720 --> 00:13:50.340
of AI. And I think it'll start to revert to having
00:13:50.340 --> 00:13:53.440
really brilliant conceptual thinking that is
00:13:53.440 --> 00:13:56.019
then using AI to deploy at scale. And the example
00:13:56.019 --> 00:13:58.460
of that is if you ask the founders of Anthropic
00:13:58.460 --> 00:14:00.960
who they're hiring, they're hiring philosophers.
00:14:01.500 --> 00:14:04.639
They're hiring political theorists. They're hiring
00:14:04.639 --> 00:14:07.559
English lit majors, people who can think conceptually
00:14:07.559 --> 00:14:09.620
and broadly and have critical analysis and critical
00:14:09.620 --> 00:14:12.899
thinking and have originality of ideas and then
00:14:12.899 --> 00:14:16.019
use AI as a deployment. product. And I think
00:14:16.019 --> 00:14:20.080
if business leaders can find their strategy for
00:14:20.080 --> 00:14:22.899
sustainable growth, understand how deployment
00:14:22.899 --> 00:14:27.000
at scale can build productivity, but value human
00:14:27.000 --> 00:14:29.860
ingenuity and conceptual thinking and align it
00:14:29.860 --> 00:14:32.840
with the right deployment AI tools, then I think
00:14:32.840 --> 00:14:35.000
you've got a sustainable growth for a business
00:14:35.000 --> 00:14:38.100
that can have a competitive advantage while others
00:14:38.100 --> 00:14:40.519
perhaps fall into the mediocrity at scale space.
00:14:41.140 --> 00:14:45.500
Companies, I mean, companies have their challenges,
00:14:45.600 --> 00:14:48.220
but, you know, often when you're talking about
00:14:48.220 --> 00:14:51.500
a company, whether it's sort of big or small,
00:14:52.080 --> 00:14:56.519
you know, the decision making systems are fundamentally
00:14:56.519 --> 00:14:59.480
authoritarian. You know, I don't want to use
00:14:59.480 --> 00:15:02.639
that. I hate to use that word, but the structure
00:15:02.639 --> 00:15:05.379
is such that you have one group or maybe a small
00:15:05.379 --> 00:15:08.740
group of people. kind of making decisions and
00:15:08.740 --> 00:15:11.740
they can act pretty quickly. For governments
00:15:11.740 --> 00:15:15.500
and especially for our democratic societies,
00:15:16.200 --> 00:15:20.059
kind of seems like we're not able to act as quickly
00:15:20.059 --> 00:15:23.679
and seems certainly here in the United States
00:15:23.679 --> 00:15:27.340
and perhaps also in some other places that governments
00:15:27.340 --> 00:15:30.279
and security agencies and others, they're kind
00:15:30.279 --> 00:15:34.909
of really struggling to. keep up with all of
00:15:34.909 --> 00:15:37.070
this? What would you recommend to government
00:15:37.070 --> 00:15:40.909
officials or, you know, legislators, parliamentarians
00:15:40.909 --> 00:15:44.690
who need to think about how to govern all of
00:15:44.690 --> 00:15:47.210
this? You know, certainly in the U .S., we're
00:15:47.210 --> 00:15:51.370
starting to have a debate in a way that we haven't
00:15:51.370 --> 00:15:53.490
in recent years. You know, what kind of what
00:15:53.490 --> 00:15:56.610
should be the limits on AI? What thoughts do
00:15:56.610 --> 00:15:59.669
you have on that? And obviously, especially as
00:15:59.669 --> 00:16:02.269
it connects to hacking and security. Yeah, well,
00:16:02.330 --> 00:16:04.730
I think what's really interesting, if we draw
00:16:04.730 --> 00:16:07.230
a circle around hacking and cybersecurity and
00:16:07.230 --> 00:16:09.230
the role of AI and government's approaches to
00:16:09.230 --> 00:16:12.289
that, you know, the great philosopher Mike Tyson
00:16:12.289 --> 00:16:16.350
had a saying that everyone has a plan until they
00:16:16.350 --> 00:16:19.730
get punched in the face. And the truth is, my
00:16:19.730 --> 00:16:21.470
experience, and we work with a lot of governments
00:16:21.470 --> 00:16:23.289
around the world on both the defense side, the
00:16:23.289 --> 00:16:26.490
intelligence side, and the civic side, the truth
00:16:26.490 --> 00:16:29.919
is... For those who are in often liberal democracies
00:16:29.919 --> 00:16:33.480
and are living in a slightly delusional state
00:16:33.480 --> 00:16:37.220
of potential of conflict, there is an awful lot
00:16:37.220 --> 00:16:40.519
of committees making very few decisions at a
00:16:40.519 --> 00:16:44.080
snail's pace. However, for those governments
00:16:44.080 --> 00:16:48.740
who are proximate to conflict, so in there I
00:16:48.740 --> 00:16:52.899
would include the GCC post the American -Israel
00:16:52.899 --> 00:16:55.960
-Iran conflict. I would include the Balkans.
00:16:56.600 --> 00:17:02.279
Poland, the states of Estonia, Lithuania, who
00:17:02.279 --> 00:17:06.160
are ordering with Russia. I would now also include
00:17:06.160 --> 00:17:10.480
most Western European states, particularly Germany,
00:17:10.720 --> 00:17:14.140
France, are pending threat of Russia and boldness
00:17:14.140 --> 00:17:17.400
of Russia with incursions and disruptions on
00:17:17.400 --> 00:17:20.839
mainland Europe. But those governments, they
00:17:20.839 --> 00:17:24.119
are firmly on a war footing. In fact, General
00:17:24.119 --> 00:17:26.500
Sir Richard Barron, who's a four -star general
00:17:26.500 --> 00:17:29.119
in the british armed forces who wrote the strategic
00:17:29.119 --> 00:17:31.460
defense review which is the largest defense strategy
00:17:31.460 --> 00:17:34.180
review of the uk army since the second world
00:17:34.180 --> 00:17:37.240
war what he said very simply is wars are won
00:17:37.240 --> 00:17:41.099
by will and you only have will in a nation to
00:17:41.099 --> 00:17:43.200
defend itself when you have awareness of the
00:17:43.200 --> 00:17:46.160
impending threat and my experience is those nations
00:17:46.160 --> 00:17:49.579
are proximate to an antagonistic hostile active
00:17:49.579 --> 00:17:55.470
enemy such as russia iran or or china are ramping
00:17:55.470 --> 00:17:59.750
up their expenditure, their adoption, their procurement,
00:18:00.029 --> 00:18:04.509
and their implementation of defenses. Now, it
00:18:04.509 --> 00:18:09.029
takes three years to procure a missile. It takes
00:18:09.029 --> 00:18:11.849
an awful long time to target, launch, and deploy,
00:18:12.029 --> 00:18:14.609
and it's very visible. However, a cyber attack
00:18:14.609 --> 00:18:18.289
can be deployed instantaneously. And so I see
00:18:18.289 --> 00:18:22.309
actually a really hard adoption of cyber defenses.
00:18:23.039 --> 00:18:25.720
because the scale of the cyber attacks is spectacular.
00:18:25.940 --> 00:18:28.000
And I'll just give you an example. The Jaguar
00:18:28.000 --> 00:18:30.960
Land Rover hack cost 2 .5 billion pounds, so
00:18:30.960 --> 00:18:36.880
$3 billion. It caused 240 companies on the supply
00:18:36.880 --> 00:18:39.680
chain to go bankrupt, and it necessitated a $2
00:18:39.680 --> 00:18:42.859
billion government bailout from the UK government,
00:18:43.059 --> 00:18:45.500
which the government equated to two months of
00:18:45.500 --> 00:18:48.960
stagnant GDP. So there you have a kind of $1,
00:18:49.079 --> 00:18:52.579
$2 billion damage. from an attack by a hostile
00:18:52.579 --> 00:18:56.059
state actor in Russia on our critical infrastructure.
00:18:56.440 --> 00:19:00.099
If Russia had sent a drone to bomb one of our
00:19:00.099 --> 00:19:03.359
oil refineries in southwest UK, it may have cost
00:19:03.359 --> 00:19:05.660
us 100 million. And that would have triggered
00:19:05.660 --> 00:19:09.440
Article 5. But yet this attack on our critical
00:19:09.440 --> 00:19:12.599
infrastructure costing $2 billion is not seen.
00:19:13.049 --> 00:19:15.549
as kinetic, so it's not seen as triggering article
00:19:15.549 --> 00:19:20.109
5. But within the security industry, it is very
00:19:20.109 --> 00:19:22.529
much seen as a highly aggressive attack. So what's
00:19:22.529 --> 00:19:25.710
happening is mobilization on the cyber side.
00:19:25.869 --> 00:19:28.210
And I'll give you an example. The Hacking Games
00:19:28.210 --> 00:19:31.150
works with a series of governments across NATO
00:19:31.150 --> 00:19:35.849
and the GCC, where we use our AI models to do
00:19:35.849 --> 00:19:39.470
two things. One, we use them to find elite Gen
00:19:39.470 --> 00:19:42.009
Z hackers in their countries based on how they
00:19:42.009 --> 00:19:44.089
play computer games, hack computer games in their
00:19:44.089 --> 00:19:47.150
levels of autism so that the country can recruit
00:19:47.150 --> 00:19:50.269
elite level hackers to protect their critical
00:19:50.269 --> 00:19:53.150
infrastructure. And then they use our AI models
00:19:53.150 --> 00:19:58.250
to replicate how hostile Gen Z hackers would
00:19:58.250 --> 00:20:00.630
be deployed against defenses and critical infrastructure
00:20:00.630 --> 00:20:04.549
to destabilize the nation. So the AI models can
00:20:04.549 --> 00:20:07.130
replicate those attacks. so that they can calibrate
00:20:07.130 --> 00:20:10.730
their defenses appropriately. So we see fast
00:20:10.730 --> 00:20:14.789
adoption by nations proximate to conflict, citing
00:20:14.789 --> 00:20:17.289
cyber as the first line of defense for critical
00:20:17.289 --> 00:20:19.890
infrastructure. However, we see those nations
00:20:19.890 --> 00:20:23.269
who might have, for example, a big Atlantic separating
00:20:23.269 --> 00:20:26.890
them from hostile state act moving slower with
00:20:26.890 --> 00:20:29.670
slightly less urgency. But we do see them also
00:20:29.670 --> 00:20:32.769
adopting speed. And you'll note that the White
00:20:32.769 --> 00:20:34.829
House announced a new policy just last month.
00:20:35.240 --> 00:20:38.319
empowering cybersecurity businesses in private
00:20:38.319 --> 00:20:41.720
sector in America to hack hostile state actors
00:20:41.720 --> 00:20:43.859
on behalf of the states. And that is a hugely
00:20:43.859 --> 00:20:46.839
controversial policy for lots of reasons. But
00:20:46.839 --> 00:20:49.720
that is a sign that governance is waking up fast
00:20:49.720 --> 00:20:52.960
to the critical threat of cybersecurity to their
00:20:52.960 --> 00:20:54.400
critical infrastructure and national security.
00:20:54.970 --> 00:20:58.630
Since you quoted, as you said, the great philosopher
00:20:58.630 --> 00:21:02.390
Mike Tyson, you're kind of reminding me as you're
00:21:02.390 --> 00:21:07.549
talking here now of the English writer Samuel
00:21:07.549 --> 00:21:12.789
Johnson, who very famously said, when a man knows
00:21:12.789 --> 00:21:16.049
he's to be hanged in a fortnight, it concentrates
00:21:16.049 --> 00:21:19.130
the mind wonderfully. That seems to be kind of
00:21:19.130 --> 00:21:21.890
the message that you're delivering there. You
00:21:21.890 --> 00:21:24.680
know, not a surprising one, really. that the
00:21:24.680 --> 00:21:27.819
governments that feel a really urgent threat
00:21:27.819 --> 00:21:31.519
would also feel a much more urgent need to respond.
00:21:31.859 --> 00:21:33.859
There are an enormous amount of philosophers
00:21:33.859 --> 00:21:38.619
who have also identified the folly of not respecting
00:21:38.619 --> 00:21:42.279
youth. And in fact, Einstein himself stated that
00:21:42.279 --> 00:21:45.150
he didn't have a new idea. from the age of 24
00:21:45.150 --> 00:21:47.930
year olds onwards. So I do think that the next
00:21:47.930 --> 00:21:51.789
wave of national security defense and preservation
00:21:51.789 --> 00:21:56.049
of our democratic societies is about not infantilizing
00:21:56.049 --> 00:21:58.930
the young generation, but in fact, understanding
00:21:58.930 --> 00:22:01.309
the incredible horsepower that they have at their
00:22:01.309 --> 00:22:03.650
fingertips between their creativity and their
00:22:03.650 --> 00:22:06.710
access to technology and harnessing that and
00:22:06.710 --> 00:22:09.380
respecting that and empowering that. for the
00:22:09.380 --> 00:22:11.660
defense of liberal democracy. I don't think we
00:22:11.660 --> 00:22:14.140
do a great job of that. I think we've slid into
00:22:14.140 --> 00:22:17.720
not recognizing the potential there. And in fact,
00:22:17.740 --> 00:22:21.259
if you go to Ukraine, which many of us have done,
00:22:21.420 --> 00:22:24.380
you will see that it is the young who are driving
00:22:24.380 --> 00:22:27.440
innovation of the most innovative defense force
00:22:27.440 --> 00:22:30.079
on the planet. And it is those innovations that
00:22:30.079 --> 00:22:32.119
the rest of us are now procuring in order to
00:22:32.119 --> 00:22:35.559
protect our citizens. So the one thing I would
00:22:35.559 --> 00:22:38.299
say is leaders would do well to identify the
00:22:38.299 --> 00:22:41.859
horsepower of Gen Z generation, identify, empower,
00:22:42.180 --> 00:22:44.180
and mobilize it for the protection of society.
00:22:44.730 --> 00:22:47.230
I think that's kind of a great thought to end
00:22:47.230 --> 00:22:50.410
on and certainly something pretty important for
00:22:50.410 --> 00:22:53.829
our leaders here in the US and elsewhere to think
00:22:53.829 --> 00:22:56.230
about. For everyone listening, thank you again
00:22:56.230 --> 00:22:59.150
for joining. This is Paul Saunders with the Center
00:22:59.150 --> 00:23:00.910
for the National Interest. I've been talking
00:23:00.910 --> 00:23:06.250
to Fergus Hay, the co -founder and CEO of The
00:23:06.250 --> 00:23:09.470
Hacking Games. Thank you so much, Fergus. Pleasurable.
00:23:15.589 --> 00:23:18.390
Thank you.
00:00:17.679 --> 00:00:20.940
Hello, this is Paul Saunders. Thank you for joining
00:00:20.940 --> 00:00:24.500
us again for another episode of the Three Questions
00:00:24.500 --> 00:00:28.719
podcast. I'm really delighted to have with me
00:00:28.719 --> 00:00:33.560
today Fergus Hay. Fergus is the co -founder and
00:00:33.560 --> 00:00:37.200
chief executive officer of The Hacking Games,
00:00:37.340 --> 00:00:40.799
and I'll ask Fergus to describe that in a minute
00:00:40.799 --> 00:00:44.140
because I'm sure he'll do a better job than I,
00:00:44.259 --> 00:00:47.670
but it's very cool. And he's also quite active
00:00:47.670 --> 00:00:51.429
in the defense and technology space as director
00:00:51.429 --> 00:00:54.130
at the Defense and Security Forum, which is a
00:00:54.130 --> 00:00:58.990
defense think tank and a co -founder. of a firm
00:00:58.990 --> 00:01:03.450
called Elysian Fields, which is working as kind
00:01:03.450 --> 00:01:06.469
of an advisor in the venture capital and private
00:01:06.469 --> 00:01:09.689
equity area. So, Fergus, thank you so much for
00:01:09.689 --> 00:01:12.689
being with me today. And first, could I just
00:01:12.689 --> 00:01:15.969
ask you to explain the hacking games, which,
00:01:15.989 --> 00:01:18.930
you know, I was very excited to hear about when
00:01:18.930 --> 00:01:21.150
you first explained it to me, because it's like
00:01:21.150 --> 00:01:25.209
something you would see in a movie. Please tell
00:01:25.209 --> 00:01:28.060
me. Please tell me. Paul, first of all, it's
00:01:28.060 --> 00:01:30.859
a real privilege and delight to be invited, not
00:01:30.859 --> 00:01:32.599
only by you, but also by the national interest
00:01:32.599 --> 00:01:37.489
to share our mission and the major. global and
00:01:37.489 --> 00:01:40.049
social challenges that we think we are facing.
00:01:40.250 --> 00:01:42.829
So the Hacking Games was set up three years ago.
00:01:43.010 --> 00:01:46.230
Our vision is to create a generation of ethical
00:01:46.230 --> 00:01:48.969
hackers to make the world safer. And the reason
00:01:48.969 --> 00:01:51.549
why we think that's really important is that
00:01:51.549 --> 00:01:53.750
in the geopolitical world we're facing today,
00:01:53.989 --> 00:01:58.209
you've got the convergence of geopolitics, national
00:01:58.209 --> 00:02:02.090
security and defense threats. converging with
00:02:02.090 --> 00:02:05.629
cybercrime and the penetration of destabilization
00:02:05.629 --> 00:02:09.569
of governments, societies, companies, and citizens.
00:02:09.990 --> 00:02:11.430
And what's really, really interesting in that
00:02:11.430 --> 00:02:13.569
convergence is for so many years, we've been
00:02:13.569 --> 00:02:16.310
focused on cybercrime. And when you look at the
00:02:16.310 --> 00:02:19.129
scale of cybercrime, it's pretty spectacular.
00:02:19.270 --> 00:02:22.430
In fact, last year, it was forecasted to cost
00:02:22.430 --> 00:02:25.710
the world $10 .5 trillion. If that number is
00:02:25.710 --> 00:02:28.689
accurate at all, that is... the global cost of
00:02:28.689 --> 00:02:32.550
COVID according to the IMF. Next year, it's forecasted
00:02:32.550 --> 00:02:36.270
to cost the world $24 trillion. And if that number
00:02:36.270 --> 00:02:39.870
is accurate, that's the GDP of America. So what's
00:02:39.870 --> 00:02:44.050
really striking is that cybercrime is now converging
00:02:44.050 --> 00:02:46.229
with geopolitical and national security issues,
00:02:46.530 --> 00:02:49.729
because it's no longer a crime for personal cybergang
00:02:49.729 --> 00:02:53.370
benefit, but as you can see from the hack on
00:02:53.370 --> 00:02:56.050
Jaguar Land Rover. that cost two and a half billion
00:02:56.050 --> 00:02:58.569
pounds, or the hack on the largest retailers
00:02:58.569 --> 00:03:02.349
in Europe, which cost $600 million. These were
00:03:02.349 --> 00:03:05.210
all activated by hostile state actors, in those
00:03:05.210 --> 00:03:08.590
case, Russia, to destabilize NATO governments,
00:03:08.849 --> 00:03:12.389
to drive a fragmentation of the alliance, while
00:03:12.389 --> 00:03:14.409
governments focus on their own internal issues
00:03:14.409 --> 00:03:17.110
to restore stability and do their number one
00:03:17.110 --> 00:03:19.050
role, which is to protect their own citizens.
00:03:19.430 --> 00:03:23.039
So in this convergence of cyber attacks, with
00:03:23.039 --> 00:03:25.139
national security and defense, you have this
00:03:25.139 --> 00:03:28.800
new world of kind of hybrid warfare where cybercrime
00:03:28.800 --> 00:03:31.800
or cyber weaponization is the easiest thing and
00:03:31.800 --> 00:03:34.340
lowest cost thing to do than to launch a missile.
00:03:34.659 --> 00:03:38.620
And so we started an organization in order to
00:03:38.620 --> 00:03:41.219
identify who are the people being weaponized
00:03:41.219 --> 00:03:43.960
and what can we do to identify and harness the
00:03:43.960 --> 00:03:47.800
most talented hackers in the world to not destroy
00:03:47.800 --> 00:03:50.740
society, but... protect our national security,
00:03:51.080 --> 00:03:53.319
our private enterprises, our critical infrastructure,
00:03:53.500 --> 00:03:57.139
and our communities. And this is where the cut
00:03:57.139 --> 00:03:59.080
-through really comes, because when you dive
00:03:59.080 --> 00:04:02.060
into it, what you realize is the hackers who
00:04:02.060 --> 00:04:04.379
are committing these enormous destabilizations
00:04:04.379 --> 00:04:07.900
are children. In fact, the world's largest research
00:04:07.900 --> 00:04:11.620
report commissioned by the EU interviewed 9 ,000
00:04:11.620 --> 00:04:16.300
teenagers across the EU and the UK, and 69 %
00:04:16.300 --> 00:04:20.750
of them have committed a cybercrime. 69%, 45
00:04:20.750 --> 00:04:24.250
% of them have committed an act of serious financial
00:04:24.250 --> 00:04:27.129
crime. So that's like identity theft and crypto
00:04:27.129 --> 00:04:30.829
fraud and banking currency fraud. 45%, that's
00:04:30.829 --> 00:04:33.829
12 to 14 kids out of an average high school classroom.
00:04:34.089 --> 00:04:36.189
And so you've got the weaponization of young
00:04:36.189 --> 00:04:39.670
people by hostile state actors to commit cyber
00:04:39.670 --> 00:04:42.689
criminal acts. on behalf of destabilizing nations.
00:04:42.910 --> 00:04:45.170
And we see that as the major social problem,
00:04:45.310 --> 00:04:48.069
a major cyber problem, and a major national security
00:04:48.069 --> 00:04:50.769
problem. And so we're focusing our efforts on
00:04:50.769 --> 00:04:54.310
identifying elite Gen Z hackers and ensuring
00:04:54.310 --> 00:04:56.689
they use their skills to protect society, not
00:04:56.689 --> 00:04:59.420
damage society. Thanks very much, Fergus. That's
00:04:59.420 --> 00:05:02.040
definitely way better than I could have done.
00:05:02.180 --> 00:05:05.040
So, you know, as I'm kind of listening to you
00:05:05.040 --> 00:05:07.980
there and, you know, you touched on the geopolitics,
00:05:07.980 --> 00:05:11.279
you touched on some of the things that are happening
00:05:11.279 --> 00:05:15.019
that it kind of leads into my first question,
00:05:15.060 --> 00:05:18.240
which would be just as you look back yourself
00:05:18.240 --> 00:05:22.379
and kind of reflect on this field. What changes
00:05:22.379 --> 00:05:25.500
or evolution do you see? You know, we've got
00:05:25.500 --> 00:05:29.860
obviously AI, we've got the geopolitics again,
00:05:29.939 --> 00:05:32.319
which you touched on. There are a lot of other
00:05:32.319 --> 00:05:36.040
things happening globally, economically in society.
00:05:36.199 --> 00:05:38.740
I'm just wondering how you see hacking changing.
00:05:38.860 --> 00:05:41.800
And to me, it kind of feels like this very weird
00:05:41.800 --> 00:05:45.560
moment in world history where from a certain
00:05:45.560 --> 00:05:49.740
perspective, people globally are safer and better
00:05:49.740 --> 00:05:54.040
off and more secure and prosperous than maybe
00:05:54.040 --> 00:05:57.399
we've ever been in thousands of years of human
00:05:57.399 --> 00:06:00.339
civilization. But at the same time, there's like
00:06:00.339 --> 00:06:03.120
this incredible sense of vulnerability coming
00:06:03.120 --> 00:06:06.259
from the internet and some of these other tools.
00:06:06.420 --> 00:06:09.019
Could you help me? How do you see all of that?
00:06:09.279 --> 00:06:11.600
Well, according to the hyperscalers over the
00:06:11.600 --> 00:06:13.740
weekend, we understand we've got a 10 % chance
00:06:13.740 --> 00:06:18.480
of total annihilation of civilization. So I think
00:06:18.480 --> 00:06:20.980
if you were to believe that... That's nice. That's
00:06:20.980 --> 00:06:25.439
nice. Good. Yeah. If you were to believe that,
00:06:25.579 --> 00:06:28.560
and let's suspend disbelief for a second and
00:06:28.560 --> 00:06:30.100
not think that people are just trying to pump
00:06:30.100 --> 00:06:33.079
their IPO price, you would argue that we've never
00:06:33.079 --> 00:06:35.699
been more vulnerable than we are today. I think
00:06:35.699 --> 00:06:38.160
the first thing we should do is dissect what
00:06:38.160 --> 00:06:41.519
hacking is, because already there you've used
00:06:41.519 --> 00:06:43.959
it as a negative term. But I think we have to
00:06:43.959 --> 00:06:45.560
understand there is a difference between hacking
00:06:45.560 --> 00:06:49.560
and hacking for harm. Now, hacking in its original
00:06:49.560 --> 00:06:53.000
definition is evaluating a construct, breaking
00:06:53.000 --> 00:06:55.879
it down into its component pieces, and identifying
00:06:55.879 --> 00:06:58.959
either where it's vulnerable or how it can be
00:06:58.959 --> 00:07:02.439
reused for another use. Now, based on that definition,
00:07:02.879 --> 00:07:05.939
It's actually the original source of human creativity.
00:07:06.079 --> 00:07:08.740
Without a hacking mindset, we'd be riding horses
00:07:08.740 --> 00:07:11.439
to work, or we'd be reading by candlelight, we'd
00:07:11.439 --> 00:07:14.139
be feeding ourselves by coal. So the question
00:07:14.139 --> 00:07:18.360
is, in the hands of whom to do what? Now, the
00:07:18.360 --> 00:07:20.720
evolution of hacking is worth understanding.
00:07:21.240 --> 00:07:25.240
Back in the 1970s, the original hackers were
00:07:25.240 --> 00:07:29.279
incredibly bright, highly niche, highly focused
00:07:29.279 --> 00:07:32.360
people who were... dustbin diving, you know,
00:07:32.379 --> 00:07:35.500
diving into dustbins to build computers, to understand
00:07:35.500 --> 00:07:38.680
constructs, to kind of find weaknesses and evolve
00:07:38.680 --> 00:07:41.259
them. Tim Berners -Lee, the founder of the World
00:07:41.259 --> 00:07:44.100
Wide Web, is a self -processed hacker. And this
00:07:44.100 --> 00:07:46.480
was often the baseline of amazing innovation.
00:07:47.000 --> 00:07:51.370
What grew out of that in the 90s was... bad actors
00:07:51.370 --> 00:07:54.649
understanding that they could use this to violate
00:07:54.649 --> 00:07:57.250
the global connectivity that was happening through
00:07:57.250 --> 00:08:00.209
the internet. And that's where you started to
00:08:00.209 --> 00:08:03.750
see a kind of nihilistic, anachronistic rise
00:08:03.750 --> 00:08:07.649
of kind of breaking things. And then in the 2000s,
00:08:07.649 --> 00:08:10.350
as the kind of one banking system hit the world,
00:08:10.490 --> 00:08:13.220
people understood, oh, wow. I can hack into this
00:08:13.220 --> 00:08:16.759
nascent banking system and steal. And so that's
00:08:16.759 --> 00:08:20.019
where you get the rise of, as always, early adopters
00:08:20.019 --> 00:08:22.639
of new techniques on the criminal world because
00:08:22.639 --> 00:08:25.120
they understand that they don't need ethics committees
00:08:25.120 --> 00:08:27.459
to make decisions on the adoption of technology.
00:08:27.740 --> 00:08:30.000
They just apply it really quickly for their own
00:08:30.000 --> 00:08:32.600
personal gain. So you saw money laundering and
00:08:32.600 --> 00:08:36.240
theft and fraud and scamming. What's changed
00:08:36.240 --> 00:08:41.610
now is two things. Number one, you have a generation
00:08:41.610 --> 00:08:44.730
who are natural -born technologists. I'm technically
00:08:44.730 --> 00:08:47.750
a millennial, so I was a natural -born internet
00:08:47.750 --> 00:08:50.289
user. I don't remember the world pre -internet.
00:08:50.870 --> 00:08:53.610
There's this current young generation, Gen Z,
00:08:53.669 --> 00:08:57.250
age 12 to 27. They don't know the world really
00:08:57.250 --> 00:09:01.309
before AI. They're so native that they see AI
00:09:01.309 --> 00:09:04.269
as just a natural vein in their body. They only
00:09:04.269 --> 00:09:06.789
know how to adopt it and utilize it, and I just
00:09:06.789 --> 00:09:09.230
see it in my children. So what that means is
00:09:09.230 --> 00:09:13.370
the gap, between human creativity and the technology
00:09:13.370 --> 00:09:17.750
to deploy it and implement it is zero. The adoption
00:09:17.750 --> 00:09:22.529
levels are 100%. There isn't a barrier to entry
00:09:22.529 --> 00:09:25.570
for the deployment of these products. And so
00:09:25.570 --> 00:09:28.610
we see this kind of convergence of two behaviors.
00:09:28.789 --> 00:09:32.769
At the very top tier, we see elite -level computer
00:09:32.769 --> 00:09:35.720
scientists who have been working with data. When
00:09:35.720 --> 00:09:38.039
it was cool, it was called big data. Then it
00:09:38.039 --> 00:09:41.500
became machine learning. Now it becomes AI. And
00:09:41.500 --> 00:09:43.720
what you're seeing is the maturity development
00:09:43.720 --> 00:09:46.259
and deployment of those models at enormous scale
00:09:46.259 --> 00:09:49.220
with the horsepower of the capital markets to
00:09:49.220 --> 00:09:51.720
drive it. So that's anthropic. Level down, you're
00:09:51.720 --> 00:09:53.860
seeing these insanely powerful models that are
00:09:53.860 --> 00:09:57.399
being deployed in the cybersecurity sector to
00:09:57.399 --> 00:10:01.360
do cybersecurity penetration testing at enormous
00:10:01.360 --> 00:10:04.379
scale. And that is worrying. We see that in the
00:10:04.379 --> 00:10:06.720
media. You're seeing a whole generation of young
00:10:06.720 --> 00:10:12.139
people adopting and absorbing AI tools for their
00:10:12.139 --> 00:10:14.399
basic daily lives. And that's going to create
00:10:14.399 --> 00:10:19.240
a generation of innovators and AI users to create
00:10:19.240 --> 00:10:22.779
further innovations and a leapfrog of innovation,
00:10:22.960 --> 00:10:25.720
technology, and wealth creation. And Elon Musk
00:10:25.720 --> 00:10:28.340
thinks it'll add 30 % to the global GDP, $30
00:10:28.340 --> 00:10:31.960
trillion of opportunity, which could may well
00:10:31.960 --> 00:10:36.990
be the case. hyperbolic headlines and such as
00:10:36.990 --> 00:10:40.570
we've got a 10 chance of total nihilism um they
00:10:40.570 --> 00:10:43.990
latch on to a tropics model and escaping the
00:10:43.990 --> 00:10:47.730
sandbox and hacking organizations but the truth
00:10:47.730 --> 00:10:51.110
is as always not as extreme on either side what
00:10:51.110 --> 00:10:54.610
we do know for sure is that side the cyber world
00:10:54.970 --> 00:10:58.309
is really the second category that has deployed
00:10:58.309 --> 00:11:01.909
cyber at such enormous scale, deployed AI at
00:11:01.909 --> 00:11:03.929
such enormous scale. And that will have both
00:11:03.929 --> 00:11:06.850
a threat to society, but ultimately, once it's
00:11:06.850 --> 00:11:08.629
adopted and deployed in an intelligent fashion,
00:11:08.750 --> 00:11:12.090
will enable stronger defenses of society. You're
00:11:12.090 --> 00:11:14.789
working a lot, you know, with governments and
00:11:14.789 --> 00:11:18.169
certainly also with companies. And I guess I
00:11:18.169 --> 00:11:21.370
wonder if you try to put on, you know, the hat
00:11:21.370 --> 00:11:26.169
of a government official or a corporate executive,
00:11:26.509 --> 00:11:29.309
you know, trying to navigate all of this. What
00:11:29.309 --> 00:11:32.029
do you think are the biggest challenges for people
00:11:32.029 --> 00:11:34.610
like that? The biggest problems they face? What
00:11:34.610 --> 00:11:37.129
are the biggest threats they see? Well, I think
00:11:37.129 --> 00:11:39.570
that, as always, the biggest problems are lurching
00:11:39.570 --> 00:11:43.090
into hyperbole. So, you know, actions without
00:11:43.090 --> 00:11:47.190
strategy cause reduction in effectiveness. So
00:11:47.190 --> 00:11:51.629
I think as leaders of organizations, before leaping
00:11:51.629 --> 00:11:54.029
into adopting AI blindly, it needs to be within
00:11:54.029 --> 00:11:56.669
strategy for productive, sustainable growth of
00:11:56.669 --> 00:12:00.850
the organization. I think that people have moved
00:12:00.850 --> 00:12:03.970
on now from it being a cost -cutting exercise.
00:12:05.039 --> 00:12:07.340
and understanding that it's a productivity exercise.
00:12:07.840 --> 00:12:10.320
And people are also understanding the limitations
00:12:10.320 --> 00:12:13.000
of AI. We have a really strong view on this.
00:12:13.139 --> 00:12:15.259
You know, we think that the deployment of AI
00:12:15.259 --> 00:12:18.419
is most visible, sadly, in media and marketing.
00:12:18.830 --> 00:12:21.029
Just cruise your LinkedIn channel and you will
00:12:21.029 --> 00:12:24.309
see what happens when people deploy mediocre
00:12:24.309 --> 00:12:28.289
AI at scale without any genuine human insight
00:12:28.289 --> 00:12:31.309
and ingenuity. You just simply get mass mediocrity
00:12:31.309 --> 00:12:33.950
or slot. And I think the world is kind of overrun
00:12:33.950 --> 00:12:36.389
with it at the moment. And it's super boring.
00:12:36.789 --> 00:12:40.769
I mean, it's super boring and not engaging or
00:12:40.769 --> 00:12:44.009
interesting. There's no content. It's very vacuous.
00:12:44.210 --> 00:12:47.370
I mean, I completely agree. It's reductive. We're
00:12:47.370 --> 00:12:51.330
all becoming more mediocre. We couldn't be converging
00:12:51.330 --> 00:12:53.950
under the middle of the curve faster in the deployment
00:12:53.950 --> 00:12:57.789
of mediocre AI. However, what's really interesting
00:12:57.789 --> 00:13:01.129
is when you take a brilliant mind, what we call
00:13:01.129 --> 00:13:03.450
a corkscrew mind, the kind of minds that Winston
00:13:03.450 --> 00:13:07.789
Churchill in 1940 asked Alan Turing to find to
00:13:07.789 --> 00:13:10.090
break the Enigma code, people who think differently.
00:13:10.409 --> 00:13:12.450
And as you know, Alan Turing put cryptographic
00:13:12.450 --> 00:13:15.110
puzzles into the newspaper, said if you can crack
00:13:15.110 --> 00:13:18.200
this... And you can join Bletchley Park and these
00:13:18.200 --> 00:13:21.000
elite level of thinkers to break this enormous
00:13:21.000 --> 00:13:23.539
cryptographic puzzle in the Enigma code. And
00:13:23.539 --> 00:13:25.860
that was the basis of cryptography and cybersecurity
00:13:25.860 --> 00:13:29.340
today. So if you take a corkscrew mind, a mind
00:13:29.340 --> 00:13:31.220
that thinks differently, what today we would
00:13:31.220 --> 00:13:34.019
call a neurodivergent or neurobrilliant mind,
00:13:34.200 --> 00:13:36.559
and you have them develop incredibly insightful
00:13:36.559 --> 00:13:40.960
concepts and you use AI to scale that, you get
00:13:40.960 --> 00:13:43.480
brilliant outcomes. And I think the world has
00:13:43.480 --> 00:13:46.720
leapt into... mediocrity at scale through deployment
00:13:46.720 --> 00:13:50.340
of AI. And I think it'll start to revert to having
00:13:50.340 --> 00:13:53.440
really brilliant conceptual thinking that is
00:13:53.440 --> 00:13:56.019
then using AI to deploy at scale. And the example
00:13:56.019 --> 00:13:58.460
of that is if you ask the founders of Anthropic
00:13:58.460 --> 00:14:00.960
who they're hiring, they're hiring philosophers.
00:14:01.500 --> 00:14:04.639
They're hiring political theorists. They're hiring
00:14:04.639 --> 00:14:07.559
English lit majors, people who can think conceptually
00:14:07.559 --> 00:14:09.620
and broadly and have critical analysis and critical
00:14:09.620 --> 00:14:12.899
thinking and have originality of ideas and then
00:14:12.899 --> 00:14:16.019
use AI as a deployment. product. And I think
00:14:16.019 --> 00:14:20.080
if business leaders can find their strategy for
00:14:20.080 --> 00:14:22.899
sustainable growth, understand how deployment
00:14:22.899 --> 00:14:27.000
at scale can build productivity, but value human
00:14:27.000 --> 00:14:29.860
ingenuity and conceptual thinking and align it
00:14:29.860 --> 00:14:32.840
with the right deployment AI tools, then I think
00:14:32.840 --> 00:14:35.000
you've got a sustainable growth for a business
00:14:35.000 --> 00:14:38.100
that can have a competitive advantage while others
00:14:38.100 --> 00:14:40.519
perhaps fall into the mediocrity at scale space.
00:14:41.140 --> 00:14:45.500
Companies, I mean, companies have their challenges,
00:14:45.600 --> 00:14:48.220
but, you know, often when you're talking about
00:14:48.220 --> 00:14:51.500
a company, whether it's sort of big or small,
00:14:52.080 --> 00:14:56.519
you know, the decision making systems are fundamentally
00:14:56.519 --> 00:14:59.480
authoritarian. You know, I don't want to use
00:14:59.480 --> 00:15:02.639
that. I hate to use that word, but the structure
00:15:02.639 --> 00:15:05.379
is such that you have one group or maybe a small
00:15:05.379 --> 00:15:08.740
group of people. kind of making decisions and
00:15:08.740 --> 00:15:11.740
they can act pretty quickly. For governments
00:15:11.740 --> 00:15:15.500
and especially for our democratic societies,
00:15:16.200 --> 00:15:20.059
kind of seems like we're not able to act as quickly
00:15:20.059 --> 00:15:23.679
and seems certainly here in the United States
00:15:23.679 --> 00:15:27.340
and perhaps also in some other places that governments
00:15:27.340 --> 00:15:30.279
and security agencies and others, they're kind
00:15:30.279 --> 00:15:34.909
of really struggling to. keep up with all of
00:15:34.909 --> 00:15:37.070
this? What would you recommend to government
00:15:37.070 --> 00:15:40.909
officials or, you know, legislators, parliamentarians
00:15:40.909 --> 00:15:44.690
who need to think about how to govern all of
00:15:44.690 --> 00:15:47.210
this? You know, certainly in the U .S., we're
00:15:47.210 --> 00:15:51.370
starting to have a debate in a way that we haven't
00:15:51.370 --> 00:15:53.490
in recent years. You know, what kind of what
00:15:53.490 --> 00:15:56.610
should be the limits on AI? What thoughts do
00:15:56.610 --> 00:15:59.669
you have on that? And obviously, especially as
00:15:59.669 --> 00:16:02.269
it connects to hacking and security. Yeah, well,
00:16:02.330 --> 00:16:04.730
I think what's really interesting, if we draw
00:16:04.730 --> 00:16:07.230
a circle around hacking and cybersecurity and
00:16:07.230 --> 00:16:09.230
the role of AI and government's approaches to
00:16:09.230 --> 00:16:12.289
that, you know, the great philosopher Mike Tyson
00:16:12.289 --> 00:16:16.350
had a saying that everyone has a plan until they
00:16:16.350 --> 00:16:19.730
get punched in the face. And the truth is, my
00:16:19.730 --> 00:16:21.470
experience, and we work with a lot of governments
00:16:21.470 --> 00:16:23.289
around the world on both the defense side, the
00:16:23.289 --> 00:16:26.490
intelligence side, and the civic side, the truth
00:16:26.490 --> 00:16:29.919
is... For those who are in often liberal democracies
00:16:29.919 --> 00:16:33.480
and are living in a slightly delusional state
00:16:33.480 --> 00:16:37.220
of potential of conflict, there is an awful lot
00:16:37.220 --> 00:16:40.519
of committees making very few decisions at a
00:16:40.519 --> 00:16:44.080
snail's pace. However, for those governments
00:16:44.080 --> 00:16:48.740
who are proximate to conflict, so in there I
00:16:48.740 --> 00:16:52.899
would include the GCC post the American -Israel
00:16:52.899 --> 00:16:55.960
-Iran conflict. I would include the Balkans.
00:16:56.600 --> 00:17:02.279
Poland, the states of Estonia, Lithuania, who
00:17:02.279 --> 00:17:06.160
are ordering with Russia. I would now also include
00:17:06.160 --> 00:17:10.480
most Western European states, particularly Germany,
00:17:10.720 --> 00:17:14.140
France, are pending threat of Russia and boldness
00:17:14.140 --> 00:17:17.400
of Russia with incursions and disruptions on
00:17:17.400 --> 00:17:20.839
mainland Europe. But those governments, they
00:17:20.839 --> 00:17:24.119
are firmly on a war footing. In fact, General
00:17:24.119 --> 00:17:26.500
Sir Richard Barron, who's a four -star general
00:17:26.500 --> 00:17:29.119
in the british armed forces who wrote the strategic
00:17:29.119 --> 00:17:31.460
defense review which is the largest defense strategy
00:17:31.460 --> 00:17:34.180
review of the uk army since the second world
00:17:34.180 --> 00:17:37.240
war what he said very simply is wars are won
00:17:37.240 --> 00:17:41.099
by will and you only have will in a nation to
00:17:41.099 --> 00:17:43.200
defend itself when you have awareness of the
00:17:43.200 --> 00:17:46.160
impending threat and my experience is those nations
00:17:46.160 --> 00:17:49.579
are proximate to an antagonistic hostile active
00:17:49.579 --> 00:17:55.470
enemy such as russia iran or or china are ramping
00:17:55.470 --> 00:17:59.750
up their expenditure, their adoption, their procurement,
00:18:00.029 --> 00:18:04.509
and their implementation of defenses. Now, it
00:18:04.509 --> 00:18:09.029
takes three years to procure a missile. It takes
00:18:09.029 --> 00:18:11.849
an awful long time to target, launch, and deploy,
00:18:12.029 --> 00:18:14.609
and it's very visible. However, a cyber attack
00:18:14.609 --> 00:18:18.289
can be deployed instantaneously. And so I see
00:18:18.289 --> 00:18:22.309
actually a really hard adoption of cyber defenses.
00:18:23.039 --> 00:18:25.720
because the scale of the cyber attacks is spectacular.
00:18:25.940 --> 00:18:28.000
And I'll just give you an example. The Jaguar
00:18:28.000 --> 00:18:30.960
Land Rover hack cost 2 .5 billion pounds, so
00:18:30.960 --> 00:18:36.880
$3 billion. It caused 240 companies on the supply
00:18:36.880 --> 00:18:39.680
chain to go bankrupt, and it necessitated a $2
00:18:39.680 --> 00:18:42.859
billion government bailout from the UK government,
00:18:43.059 --> 00:18:45.500
which the government equated to two months of
00:18:45.500 --> 00:18:48.960
stagnant GDP. So there you have a kind of $1,
00:18:49.079 --> 00:18:52.579
$2 billion damage. from an attack by a hostile
00:18:52.579 --> 00:18:56.059
state actor in Russia on our critical infrastructure.
00:18:56.440 --> 00:19:00.099
If Russia had sent a drone to bomb one of our
00:19:00.099 --> 00:19:03.359
oil refineries in southwest UK, it may have cost
00:19:03.359 --> 00:19:05.660
us 100 million. And that would have triggered
00:19:05.660 --> 00:19:09.440
Article 5. But yet this attack on our critical
00:19:09.440 --> 00:19:12.599
infrastructure costing $2 billion is not seen.
00:19:13.049 --> 00:19:15.549
as kinetic, so it's not seen as triggering article
00:19:15.549 --> 00:19:20.109
5. But within the security industry, it is very
00:19:20.109 --> 00:19:22.529
much seen as a highly aggressive attack. So what's
00:19:22.529 --> 00:19:25.710
happening is mobilization on the cyber side.
00:19:25.869 --> 00:19:28.210
And I'll give you an example. The Hacking Games
00:19:28.210 --> 00:19:31.150
works with a series of governments across NATO
00:19:31.150 --> 00:19:35.849
and the GCC, where we use our AI models to do
00:19:35.849 --> 00:19:39.470
two things. One, we use them to find elite Gen
00:19:39.470 --> 00:19:42.009
Z hackers in their countries based on how they
00:19:42.009 --> 00:19:44.089
play computer games, hack computer games in their
00:19:44.089 --> 00:19:47.150
levels of autism so that the country can recruit
00:19:47.150 --> 00:19:50.269
elite level hackers to protect their critical
00:19:50.269 --> 00:19:53.150
infrastructure. And then they use our AI models
00:19:53.150 --> 00:19:58.250
to replicate how hostile Gen Z hackers would
00:19:58.250 --> 00:20:00.630
be deployed against defenses and critical infrastructure
00:20:00.630 --> 00:20:04.549
to destabilize the nation. So the AI models can
00:20:04.549 --> 00:20:07.130
replicate those attacks. so that they can calibrate
00:20:07.130 --> 00:20:10.730
their defenses appropriately. So we see fast
00:20:10.730 --> 00:20:14.789
adoption by nations proximate to conflict, citing
00:20:14.789 --> 00:20:17.289
cyber as the first line of defense for critical
00:20:17.289 --> 00:20:19.890
infrastructure. However, we see those nations
00:20:19.890 --> 00:20:23.269
who might have, for example, a big Atlantic separating
00:20:23.269 --> 00:20:26.890
them from hostile state act moving slower with
00:20:26.890 --> 00:20:29.670
slightly less urgency. But we do see them also
00:20:29.670 --> 00:20:32.769
adopting speed. And you'll note that the White
00:20:32.769 --> 00:20:34.829
House announced a new policy just last month.
00:20:35.240 --> 00:20:38.319
empowering cybersecurity businesses in private
00:20:38.319 --> 00:20:41.720
sector in America to hack hostile state actors
00:20:41.720 --> 00:20:43.859
on behalf of the states. And that is a hugely
00:20:43.859 --> 00:20:46.839
controversial policy for lots of reasons. But
00:20:46.839 --> 00:20:49.720
that is a sign that governance is waking up fast
00:20:49.720 --> 00:20:52.960
to the critical threat of cybersecurity to their
00:20:52.960 --> 00:20:54.400
critical infrastructure and national security.
00:20:54.970 --> 00:20:58.630
Since you quoted, as you said, the great philosopher
00:20:58.630 --> 00:21:02.390
Mike Tyson, you're kind of reminding me as you're
00:21:02.390 --> 00:21:07.549
talking here now of the English writer Samuel
00:21:07.549 --> 00:21:12.789
Johnson, who very famously said, when a man knows
00:21:12.789 --> 00:21:16.049
he's to be hanged in a fortnight, it concentrates
00:21:16.049 --> 00:21:19.130
the mind wonderfully. That seems to be kind of
00:21:19.130 --> 00:21:21.890
the message that you're delivering there. You
00:21:21.890 --> 00:21:24.680
know, not a surprising one, really. that the
00:21:24.680 --> 00:21:27.819
governments that feel a really urgent threat
00:21:27.819 --> 00:21:31.519
would also feel a much more urgent need to respond.
00:21:31.859 --> 00:21:33.859
There are an enormous amount of philosophers
00:21:33.859 --> 00:21:38.619
who have also identified the folly of not respecting
00:21:38.619 --> 00:21:42.279
youth. And in fact, Einstein himself stated that
00:21:42.279 --> 00:21:45.150
he didn't have a new idea. from the age of 24
00:21:45.150 --> 00:21:47.930
year olds onwards. So I do think that the next
00:21:47.930 --> 00:21:51.789
wave of national security defense and preservation
00:21:51.789 --> 00:21:56.049
of our democratic societies is about not infantilizing
00:21:56.049 --> 00:21:58.930
the young generation, but in fact, understanding
00:21:58.930 --> 00:22:01.309
the incredible horsepower that they have at their
00:22:01.309 --> 00:22:03.650
fingertips between their creativity and their
00:22:03.650 --> 00:22:06.710
access to technology and harnessing that and
00:22:06.710 --> 00:22:09.380
respecting that and empowering that. for the
00:22:09.380 --> 00:22:11.660
defense of liberal democracy. I don't think we
00:22:11.660 --> 00:22:14.140
do a great job of that. I think we've slid into
00:22:14.140 --> 00:22:17.720
not recognizing the potential there. And in fact,
00:22:17.740 --> 00:22:21.259
if you go to Ukraine, which many of us have done,
00:22:21.420 --> 00:22:24.380
you will see that it is the young who are driving
00:22:24.380 --> 00:22:27.440
innovation of the most innovative defense force
00:22:27.440 --> 00:22:30.079
on the planet. And it is those innovations that
00:22:30.079 --> 00:22:32.119
the rest of us are now procuring in order to
00:22:32.119 --> 00:22:35.559
protect our citizens. So the one thing I would
00:22:35.559 --> 00:22:38.299
say is leaders would do well to identify the
00:22:38.299 --> 00:22:41.859
horsepower of Gen Z generation, identify, empower,
00:22:42.180 --> 00:22:44.180
and mobilize it for the protection of society.
00:22:44.730 --> 00:22:47.230
I think that's kind of a great thought to end
00:22:47.230 --> 00:22:50.410
on and certainly something pretty important for
00:22:50.410 --> 00:22:53.829
our leaders here in the US and elsewhere to think
00:22:53.829 --> 00:22:56.230
about. For everyone listening, thank you again
00:22:56.230 --> 00:22:59.150
for joining. This is Paul Saunders with the Center
00:22:59.150 --> 00:23:00.910
for the National Interest. I've been talking
00:23:00.910 --> 00:23:06.250
to Fergus Hay, the co -founder and CEO of The
00:23:06.250 --> 00:23:09.470
Hacking Games. Thank you so much, Fergus. Pleasurable.
00:23:15.589 --> 00:23:18.390
Thank you.
