رونوشت 🔗
1
00:00:00,413 --> 00:00:02,996
(upbeat music)
2
00:00:05,400 --> 00:00:06,720
- Welcome to another episode
3
00:00:06,720 --> 00:00:10,380
of "Supply Chain Frontiers,"
the MIT CTL podcast,
4
00:00:10,380 --> 00:00:11,730
where we explore the trends,
5
00:00:11,730 --> 00:00:14,430
technologies, and innovations
shaping the future
6
00:00:14,430 --> 00:00:16,230
of supply chain management.
7
00:00:16,230 --> 00:00:18,600
I'm your host, Mackenzie Berry.
8
00:00:18,600 --> 00:00:20,610
Today, we're discussing cybersecurity
9
00:00:20,610 --> 00:00:22,680
and how it impacts supply chains.
10
00:00:22,680 --> 00:00:24,120
Joining me are researchers
11
00:00:24,120 --> 00:00:27,090
from the MIT Center for
Transportation and Logistics,
12
00:00:27,090 --> 00:00:30,660
co-founders of its new Supply
Chain Cybersecurity Lab,
13
00:00:30,660 --> 00:00:34,080
Doctors Miguel Rodriguez
Garcia and Camilo Mora.
14
00:00:34,080 --> 00:00:36,750
Thank you both for joining me here.
15
00:00:36,750 --> 00:00:38,400
First question for you all,
16
00:00:38,400 --> 00:00:41,100
what is the Supply
Chain Cybersecurity Lab,
17
00:00:41,100 --> 00:00:43,023
and why launch it right now?
18
00:00:43,920 --> 00:00:46,440
- Let me tell you a
quick story, Mackenzie.
19
00:00:46,440 --> 00:00:50,190
So, in my case, about like
a year and a half ago,
20
00:00:50,190 --> 00:00:54,480
I was at an industry conference
when a colleague of mine,
21
00:00:54,480 --> 00:00:56,640
a supply chain professional,
22
00:00:56,640 --> 00:01:01,140
came up to me and said,
"Camilo, we were attacked."
23
00:01:01,140 --> 00:01:04,680
Of course, my first reaction
was like, "What do you mean,
24
00:01:04,680 --> 00:01:06,420
like, you were attacked?"
25
00:01:06,420 --> 00:01:08,550
And then, this person explained
26
00:01:08,550 --> 00:01:12,390
that their company had
experienced a ransomware attack.
27
00:01:12,390 --> 00:01:15,810
The incident disrupted
operations nationwide,
28
00:01:15,810 --> 00:01:18,450
and it took them, you know, roughly,
29
00:01:18,450 --> 00:01:20,760
about, like three months to fully restore
30
00:01:20,760 --> 00:01:24,600
and recover their systems and operations.
31
00:01:24,600 --> 00:01:27,210
That moment, that
conversation really stayed
32
00:01:27,210 --> 00:01:32,160
with me because it made me
realize, like the enormous impact
33
00:01:32,160 --> 00:01:34,140
that a cyberattack can have beyond
34
00:01:34,140 --> 00:01:37,170
the classic IT department.
35
00:01:37,170 --> 00:01:39,510
When we think about supply chains,
36
00:01:39,510 --> 00:01:42,450
we normally talk about
three fundamental flows:
37
00:01:42,450 --> 00:01:45,180
the flow of goods, the
flow of information,
38
00:01:45,180 --> 00:01:48,060
and the flow of money, the financial flow.
39
00:01:48,060 --> 00:01:51,420
A cyberattack can
certainly disrupt all three
40
00:01:51,420 --> 00:01:52,353
at the same time.
41
00:01:53,460 --> 00:01:56,820
And on top of that, you have the potential
42
00:01:56,820 --> 00:02:00,390
reputational damage and
impacts on customers,
43
00:02:00,390 --> 00:02:04,860
on suppliers, employees, you
know, your entire network.
44
00:02:04,860 --> 00:02:08,490
Shortly after, Miguel and I
started interviewing companies
45
00:02:08,490 --> 00:02:11,010
that had actually
experienced cyber incidents,
46
00:02:11,010 --> 00:02:13,710
and something became very clear:
47
00:02:13,710 --> 00:02:17,910
the classic problem,
organizations working in silos.
48
00:02:17,910 --> 00:02:20,460
So, on the one side,
you had like, you know,
49
00:02:20,460 --> 00:02:21,930
the supply chain people.
50
00:02:21,930 --> 00:02:25,680
On the other side, we had the
cybersecurity and IT teams.
51
00:02:25,680 --> 00:02:29,040
Both dealing with, you
know, the same disruption,
52
00:02:29,040 --> 00:02:32,640
but from completely
different perspectives.
53
00:02:32,640 --> 00:02:37,590
That is when, like,
Miguel and I saw the need
54
00:02:37,590 --> 00:02:41,340
to create the Supply
Chain Cybersecurity Lab.
55
00:02:41,340 --> 00:02:44,280
Our goal is to understand
how cyberattacks translate
56
00:02:44,280 --> 00:02:47,010
into physical operational disruption
57
00:02:47,010 --> 00:02:49,890
and then help organizations
become better prepared
58
00:02:49,890 --> 00:02:54,890
to prevent, mitigate, and
respond to those disruptions.
59
00:02:55,320 --> 00:02:56,580
- For listeners unfamiliar
60
00:02:56,580 --> 00:02:58,590
with supply chain cybersecurity,
61
00:02:58,590 --> 00:03:02,280
why should non-IT leaders
care about this issue?
62
00:03:02,280 --> 00:03:05,910
What makes cyberattacks relevant
across supply chains today?
63
00:03:05,910 --> 00:03:08,700
- I started working actually on cyber
64
00:03:08,700 --> 00:03:10,470
and kind of like the connection
65
00:03:10,470 --> 00:03:13,470
with supply chains about three years ago
66
00:03:13,470 --> 00:03:16,260
because we had a pretty big
project with the U.S. government
67
00:03:16,260 --> 00:03:18,090
to basically find
different vulnerabilities
68
00:03:18,090 --> 00:03:22,020
in the warehousing space when
it came to technology risks.
69
00:03:22,020 --> 00:03:24,210
Of course, cyber was the biggest one
70
00:03:24,210 --> 00:03:26,550
that came out of that study,
71
00:03:26,550 --> 00:03:30,600
and because of that, we
realized, "Wow, cybercrime,
72
00:03:30,600 --> 00:03:33,990
mainly after COVID, was basically
going through the roof."
73
00:03:33,990 --> 00:03:37,740
Just in 2025, to give you an idea,
74
00:03:37,740 --> 00:03:40,560
the impact of cybercrime
in the world was estimated
75
00:03:40,560 --> 00:03:45,150
between $2 trillion and $10
trillion, with a T, right?
76
00:03:45,150 --> 00:03:47,910
Just the average of that
would be $6 trillion.
77
00:03:47,910 --> 00:03:50,130
That's the third-largest country by GDP,
78
00:03:50,130 --> 00:03:51,600
only after the U.S. and China.
79
00:03:51,600 --> 00:03:56,600
So, the scale, you know, of
cybercrime worldwide was huge,
80
Load full transcript
00:03:56,640 --> 00:03:58,290
and the targets were changing.
81
00:03:58,290 --> 00:04:00,270
I think most of the people
in the audience will think
82
00:04:00,270 --> 00:04:02,240
about cybercrime and maybe, you know,
83
00:04:02,240 --> 00:04:05,177
they are gonna envision a
guy in a room, you know,
84
00:04:05,177 --> 00:04:10,177
like a teenage, like, kid,
like back in the, you know,
85
00:04:10,470 --> 00:04:12,660
90s or so when the internet was started.
86
00:04:12,660 --> 00:04:14,221
That's not the reality anymore, right?
87
00:04:14,221 --> 00:04:15,054
- Yeah.
88
00:04:15,054 --> 00:04:16,080
- We have a bunch of organized crime.
89
00:04:16,080 --> 00:04:20,340
We have a lot of state-sponsored,
like, cybercrime nowadays.
90
00:04:20,340 --> 00:04:21,900
Also geopolitics is not helping
91
00:04:21,900 --> 00:04:24,060
in that sense, unfortunately.
92
00:04:24,060 --> 00:04:24,893
But the reality
93
00:04:24,893 --> 00:04:27,120
is that because of that,
we have seen a change
94
00:04:27,120 --> 00:04:30,450
in how the attackers are
actually performing the attacks.
95
00:04:30,450 --> 00:04:33,960
So we see more targets
that are just aiming
96
00:04:33,960 --> 00:04:36,420
at disrupting operations, and that means
97
00:04:36,420 --> 00:04:38,910
a lot of the times stopping
physical supply chains,
98
00:04:38,910 --> 00:04:41,730
like disrupting a manufacturing
plant from producing,
99
00:04:41,730 --> 00:04:44,370
a warehouse from sending
product, and so on.
100
00:04:44,370 --> 00:04:48,660
So that is why nowadays
anybody in the operations world
101
00:04:48,660 --> 00:04:50,280
and in supply chain in particular needs
102
00:04:50,280 --> 00:04:52,200
to think about this in
a different way, right?
103
00:04:52,200 --> 00:04:54,660
Because these attacks
are already impacting us,
104
00:04:54,660 --> 00:04:56,100
and it's huge.
105
00:04:56,100 --> 00:04:58,860
- So there's a very high
level of sophistication,
106
00:04:58,860 --> 00:04:59,880
- Yeah.
- That we're looking at.
107
00:04:59,880 --> 00:05:01,980
- So super high level of sophistication
108
00:05:01,980 --> 00:05:05,070
when it comes to the new
attacks and, at the same time,
109
00:05:05,070 --> 00:05:07,230
unfortunately, a very low level
110
00:05:07,230 --> 00:05:08,670
of sophistication a lot of the times
111
00:05:08,670 --> 00:05:12,450
on the operational technology
layers of our supply chains,
112
00:05:12,450 --> 00:05:15,840
and that's 'cause we rely
still a lot in legacy systems.
113
00:05:15,840 --> 00:05:18,120
A lot of them, if you
think about warehouses,
114
00:05:18,120 --> 00:05:22,020
manufacturing plants,
even vessels or ports,
115
00:05:22,020 --> 00:05:25,200
we've had pretty, like, I would say,
116
00:05:25,200 --> 00:05:28,050
good automation systems, like
physical automation systems,
117
00:05:28,050 --> 00:05:30,900
for like 50 years or more.
118
00:05:30,900 --> 00:05:34,350
Like after the Industrial
Revolution and after World War II,
119
00:05:34,350
--> 00:05:36,180
for instance, manufacturing
was the first sector
120
00:05:36,180 --> 00:05:37,860
that moved quite fast in that sense,
121
00:05:37,860 --> 00:05:40,770
right, in terms of automating
a lot of the process.
122
00:05:40,770 --> 00:05:43,020
But those robots were not, you know,
123
00:05:43,020 --> 00:05:44,460
supposed to be on the internet.
124
00:05:44,460 --> 00:05:47,160
Nowadays, they are, but you cannot
125
00:05:47,160 --> 00:05:48,360
patch them a lot of times.
126
00:05:48,360 --> 00:05:51,240
So a lot of the traditional operations
127
00:05:51,240 --> 00:05:54,420
are now super vulnerable
to modern attacks.
128
00:05:54,420 --> 00:05:55,380
- Yeah, I agree with that.
129
00:05:55,380 --> 00:05:57,930
So, like, essentially, people
130
00:05:57,930 --> 00:06:00,540
who deeply understand cybersecurity
131
00:06:00,540 --> 00:06:04,380
do not necessarily understand
how a warehouse operates,
132
00:06:04,380 --> 00:06:06,780
how transportation networks work,
133
00:06:06,780 --> 00:06:09,150
what happens on a production line,
134
00:06:09,150 --> 00:06:11,370
you know, or what, like, the operational
135
00:06:11,370 --> 00:06:14,670
consequences are when a
critical system goes down.
136
00:06:14,670 --> 00:06:15,660
And on the other side,
137
00:06:15,660 --> 00:06:18,150
supply chain professionals
are not always trained
138
00:06:18,150 --> 00:06:20,013
to understand cyber risk.
139
00:06:20,850 --> 00:06:25,727
So we need people who can
speak both languages, right?
140
00:06:26,850 --> 00:06:29,100
We are, I would say, trying
141
00:06:29,100 --> 00:06:31,680
to help companies continue operating even
142
00:06:31,680 --> 00:06:33,210
when technology fails.
143
00:06:33,210 --> 00:06:34,770
- And it may seem obvious,
144
00:06:34,770 --> 00:06:36,720
but in terms of the motivation
145
00:06:36,720 --> 00:06:38,940
that attackers have
for doing cyberattacks,
146
00:06:38,940 --> 00:06:42,510
primarily money, primarily
wreaking havoc, a mix of both?
147
00:06:42,510 --> 00:06:44,850
- So we have seen a change in that.
148
00:06:44,850 --> 00:06:49,200
Like, for many years,
money was the main one.
149
00:06:49,200 --> 00:06:50,610
Ransom is still the main type
150
00:06:50,610 --> 00:06:53,250
of cyberattack that basically encrypts all
151
00:06:53,250 --> 00:06:57,720
of your data and locks
your systems, in a way.
152
00:06:57,720 --> 00:07:00,150
And even some of the recent
ones, they're still ransom,
153
00:07:00,150 --> 00:07:02,580
but they are not actually
asking for any money back.
154
00:07:02,580 --> 00:07:03,690
Instead, they are just trying
155
00:07:03,690 --> 00:07:05,280
to disrupt the operations anyway.
156
00:07:05,280 --> 00:07:07,380
And that has to do with the fact
157
00:07:07,380 --> 00:07:10,710
that we are seeing more
state-sponsored cybercrime
158
00:07:10,710 --> 00:07:12,270
in the last couple of years.
159
00:07:12,270 --> 00:07:15,484
So that is unfortunately
also a big change in, yeah,
160
00:07:15,484 --> 00:07:17,460
in the motivation of malicious actors.
161
00:07:17,460 --> 00:07:19,710
- I wanna speak to a specific example
162
00:07:19,710 --> 00:07:21,720
that you can walk us through.
163
00:07:21,720 --> 00:07:23,310
Can you walk us through
a recent cyberattack
164
00:07:23,310 --> 00:07:24,900
and its operational impact,
165
00:07:24,900 --> 00:07:26,700
just to give us a frame of reference?
166
00:07:26,700 --> 00:07:27,780
- Yeah, sure.
167
00:07:27,780 --> 00:07:31,620
There are many, unfortunately,
but I think 2025,
168
00:07:31,620 --> 00:07:34,470
the Jaguar Land Rover attack
was one of the, I would say,
169
00:07:34,470 --> 00:07:35,610
biggest ones in history,
170
00:07:35,610 --> 00:07:39,720
but also more, let's
say, educational, right?
171
00:07:39,720 --> 00:07:41,100
This was an attack that basically
172
00:07:41,100 --> 00:07:43,800
targeted the car manufacturer
Jaguar Land Rover
173
00:07:43,800 --> 00:07:47,070
in the U.K. They had to
shut down operations,
174
00:07:47,070 --> 00:07:48,570
like, in all manufacturing plants
175
00:07:48,570 --> 00:07:51,900
in the country for five to six weeks.
176
00:07:51,900 --> 00:07:54,360
But the thing was that
not only was, you know,
177
00:07:54,360 --> 00:07:57,120
Jaguar Land Rover affected,
but when you have to shut down,
178
00:07:57,120 --> 00:07:59,550
you know, for six weeks production,
179
00:07:59,550 --> 00:08:02,520
like, your tier one, tier two, tier three,
180
00:08:02,520 --> 00:08:06,120
tier four suppliers are
impacted very, very rapidly.
181
00:08:06,120 --> 00:08:07,440
And the reason for that is 'cause
182
00:08:07,440 --> 00:08:09,480
also a lot of those suppliers,
usually they are, you know,
183
00:08:09,480 --> 00:08:12,630
small and medium-sized companies
that have basically no,
184
00:08:12,630 --> 00:08:15,850
you know, cash runway to keep
operating without getting paid
185
00:08:16,710 --> 00:08:18,120
for a long period of time, right?
186
00:08:18,120 --> 00:08:19,650
So the total impact actually
187
00:08:19,650 --> 00:08:21,990
for the U.K. economy of that single attack
188
00:08:21,990 --> 00:08:24,970
was between $2 to $3
billion on a single attack
189
00:08:24,970 --> 00:08:27,510
to a single company, right?
190
00:08:27,510 --> 00:08:28,530
- Billion with a B.
191
00:08:28,530 --> 00:08:29,850
- Yeah, billion with a B.
192
00:08:29,850 --> 00:08:33,120
For Jaguar and Land Rover specifically,
193
00:08:33,120 --> 00:08:35,640
just the cost of cyber
recovery, which is basically
194
00:08:35,640 --> 00:08:39,810
when you have to, you know,
call consultants, lawyers,
195
00:08:39,810 --> 00:08:43,290
and usually, like, a recovery
battalion, cybersecurity team
196
00:08:43,290 --> 00:08:44,370
that just, you know, goes in
197
00:08:44,370 --> 00:08:46,590
when the worst (laughs) happens.
198
00:08:46,590 --> 00:08:49,050
They spent more than $200
million just on that,
199
00:08:49,050 --> 00:08:52,110
and that's not accounting
for all the
revenue losses
200
00:08:52,110 --> 00:08:53,907
that were estimated between, like, $50
201
00:08:53,907 --> 00:08:57,450
to $100 million per
week of lost production.
202
00:08:57,450 --> 00:08:58,920
So it was huge.
203
00:08:58,920 --> 00:09:00,900
They had to reset everything,
204
00:09:00,900 --> 00:09:02,760
basically isolate all the network
205
00:09:02,760 --> 00:09:05,370
and basically start almost from scratch.
206
00:09:05,370 --> 00:09:06,630
But yeah, it took them months.
207
00:09:06,630 --> 00:09:09,150
Until mid-November, they were not back up.
208
00:09:09,150 --> 00:09:11,040
I wanna mention now 'cause one
209
00:09:11,040 --> 00:09:13,890
of the reasons why we
launched the lab is 'cause
210
00:09:13,890 --> 00:09:16,140
actually there is not a lot
of visibility when it comes
211
00:09:16,140 --> 00:09:18,540
to all of these attacks
impacting the supply chain.
212
00:09:18,540 --> 00:09:20,707
And one of the first
things we did actually was,
213
00:09:20,707 --> 00:09:23,250
"Okay, how do we bring
this to life?" Right?
214
00:09:23,250 --> 00:09:26,730
So we created a tool,
and it's called TRACE.
215
00:09:26,730 --> 00:09:28,860
Camilo, did you wanna
maybe explain a little bit
216
00:09:28,860 --> 00:09:30,330
about what TRACE does?
217
00:09:30,330 --> 00:09:31,163
- Absolutely.
218
00:09:31,163 --> 00:09:33,660
So it is very important to learn
219
00:09:33,660 --> 00:09:35,430
from those attacks, you know,
220
00:09:35,430 --> 00:09:40,320
like, the case of Jaguar Land
Rover or Asahi or many others.
221
00:09:40,320 --> 00:09:44,070
But usually, what about the lessons
222
00:09:44,070 --> 00:09:46,920
that other companies can
learn from those incidents?
223
00:09:46,920 --> 00:09:48,930
And not, like, isolating them,
224
00:09:48,930 --> 00:09:52,500
but you know, like, to
look at them globally.
225
00:09:52,500 --> 00:09:54,890
So as Miguel mentioned,
there was no, at the time,
226
00:09:54,890 --> 00:09:59,100
or to the best of knowledge, a
platform with structured data
227
00:09:59,100 --> 00:10:02,880
about the incidents that
affect the supply chains.
228
00:10:02,880 --> 00:10:07,290
And then, you know, we
decided to develop TRACE,
229
00:10:07,290 --> 00:10:11,880
which is the World Supply Chain
Cybersecurity Observatory.
230
00:10:11,880 --> 00:10:16,260
It's free-to-access platform
where we kind of collect
231
00:10:16,260 --> 00:10:19,080
and organize publicly
available information
232
00:10:19,080 --> 00:10:21,810
about cyber incidents that, in particular,
233
00:10:21,810 --> 00:10:24,450
have affected supply
chains around the world.
234
00:10:24,450 --> 00:10:27,720
And the idea behind it is very simple.
235
00:10:27,720 --> 00:10:29,670
We wanna learn from the past
236
00:10:29,670 --> 00:10:32,493
so that we can better be
prepared for the future.
237
00:10:33,390 --> 00:10:36,480
For each incident, we try to capture
238
00:10:36,480 --> 00:10:40,020
and understand not only what
kind of cyberattack it was,
239
00:10:40,020 --> 00:10:43,590
but also what happened to the operation.
240
00:10:43,590 --> 00:10:46,920
So which industry was affected,
241
00:10:46,920 --> 00:10:49,200
which part of the organization was hit,
242
00:10:49,200 --> 00:10:51,390
was manufacturing interrupted,
243
00:10:51,390 --> 00:10:53,580
how long did the disruption last,
244
00:10:53,580 --> 00:10:56,340
what systems were involved,
so on and so forth.
245
00:10:56,340 --> 00:10:59,250
And if the information is available,
246
00:10:59,250 --> 00:11:02,687
also, what was the financial impact of it?
247
00:11:02,687 --> 00:11:03,690
- One of the issues
248
00:11:03,690 --> 00:11:05,430
that we have is that
a lot of these attacks
249
00:11:05,430 --> 00:11:06,870
are not publicly available,
250
00:11:06,870 --> 00:11:07,703
- Mm-hmm.
251
00:11:07,703 --> 00:11:08,536
- Or publicly disclosed, right?
252
00:11:08,536 --> 00:11:09,369
- That's what I was gonna
ask you about 'cause
253
00:11:09,369 --> 00:11:10,270
no company, of course, wants the PR of,
254
00:11:10,270 --> 00:11:11,383
- Yeah, yeah.
255
00:11:11,383 --> 00:11:12,216
- You know, "We have a cyberattack."
256
00:11:12,216 --> 00:11:13,590
- That's still a limitation.
257
00:11:13,590 --> 00:11:15,577
And that's one of the
reasons also why we wanted
258
00:11:15,577 --> 00:11:16,877
to create the new initiative, right?
259
00:11:16,877 --> 00:11:19,680
'Cause we wanna learn from
a lot of these companies
260
00:11:19,680 --> 00:11:21,600
that have suffered, you
know, these cyberattacks.
261
00:11:21,600 --> 00:11:26,400
And I think it's funny that you just said,
262
00:11:26,400 --> 00:11:28,893
for instance, like, you know, the PR side.
263
00:11:29,910 --> 00:11:31,830
I would say maybe 10 or 15 years ago,
264
00:11:31,830 --> 00:11:33,840
that was a good reason for it.
265
00:11:33,840 --> 00:11:36,464
Nowadays, everybody, you
know, gets hacked. (laughs)
266
00:11:36,464 --> 00:11:37,297
- Yeah.
267
00:11:37,297 --> 00:11:39,104
- So yeah, you might have a, you know,
268
00:11:39,104 --> 00:11:42,810
short loss in, you know, your
stock price for a week or two,
269
00:11:42,810 --> 00:11:44,790
but also there is a lot of research now
270
00:11:44,790 --> 00:11:47,310
that shows that that usually
recovers pretty fast,
271
00:11:47,310 --> 00:11:49,530
'cause as long as it's
not one of these massive,
272
00:11:49,530 --> 00:11:51,750
very harmful attacks, the impact
273
00:11:51,750 --> 00:11:54,180
on the operations might not
be, you know, that high.
274
00:11:54,180 --> 00:11:57,120
And those, we know them, 'cause
those, they are disclosed-
275
00:11:57,120 --> 00:11:58,140
- Sure.
- No matter what, right?
276
00:11:58,140 --> 00:12:00,870
So, but yeah, there's still
a lot of maybe attacks
277
00:12:00,870 --> 00:12:03,270
that have, you know, like, a medium
278
00:12:03,270 --> 00:12:06,750
to small impact that
probably we're not aware of.
279
00:12:06,750 --> 00:12:09,480
- Yeah, and you know,
there's a lot
changing,
280
00:12:09,480 --> 00:12:11,490
as you all mentioned, in cyberattacks.
281
00:12:11,490 --> 00:12:13,530
It almost feels like you have to be ahead
282
00:12:13,530 --> 00:12:15,780
of the curve or at least on the pulse,
283
00:12:15,780 --> 00:12:18,090
because if you're behind,
then you're vulnerable.
284
00:12:18,090 --> 00:12:19,320
Can you all speak to, you know,
285
00:12:19,320 --> 00:12:22,893
what's changing in cyberattacks
on supply chains right now?
286
00:12:24,120 --> 00:12:26,790
- I think one of the biggest things
287
00:12:26,790 --> 00:12:29,670
changing is the attack surface.
288
00:12:29,670 --> 00:12:34,670
So supply chains today
are much more connected
289
00:12:35,108 --> 00:12:38,010
than they were even a few years ago.
290
00:12:38,010 --> 00:12:42,990
Companies rely on cloud
platforms, software providers,
291
00:12:42,990 --> 00:12:46,830
technology vendors, connected
equipment, suppliers,
292
00:12:46,830 --> 00:12:50,640
logistics partners, so on and
so forth, and of course, AI.
293
00:12:50,640 --> 00:12:54,420
Every new connection creates value,
294
00:12:54,420 --> 00:12:56,970
but that can also create
another potential point
295
00:12:56,970 --> 00:12:57,873
of vulnerability.
296
00:12:58,710 --> 00:13:02,730
So AI, in fact, is a very good example.
297
00:13:02,730 --> 00:13:06,630
Right now there is almost
a race amongst companies
298
00:13:06,630 --> 00:13:08,070
to implement AI.
299
00:13:08,070 --> 00:13:12,060
There is this perception
that if we are not doing it,
300
00:13:12,060 --> 00:13:14,490
our competitors probably are, right?
301
00:13:14,490 --> 00:13:16,260
And they're getting ahead of us,
302
00:13:16,260 --> 00:13:18,120
so you know, they're scared, and they want
303
00:13:18,120 --> 00:13:20,640
to implement and, you
know, launch and develop.
304
00:13:20,640 --> 00:13:23,910
So some companies are
encouraging employees
305
00:13:23,910 --> 00:13:27,450
to experiment with AI and
become more productive,
306
00:13:27,450 --> 00:13:29,190
and that is good, you know.
307
00:13:29,190 --> 00:13:33,030
That many, like, innovations
may come out of it.
308
00:13:33,030 --> 00:13:35,220
But the challenge is that the adoption
309
00:13:35,220 --> 00:13:38,760
can sometimes move faster than governance.
310
00:13:38,760 --> 00:13:42,270
So a company might officially implement,
311
00:13:42,270 --> 00:13:45,570
I don't know, Microsoft
Copilot, for example,
312
00:13:45,570 --> 00:13:49,200
while employees are also
using their personal ChatGPT,
313
00:13:49,200 --> 00:13:52,500
Claude, and other AI accounts.
314
00:13:52,500 --> 00:13:56,280
They may be using them
for personal devices,
315
00:13:56,280 --> 00:13:59,460
uploading documents, analyzing
data, you know, connecting,
316
00:13:59,460 --> 00:14:03,210
you know, different applications,
or even developing things.
317
00:14:03,210 --> 00:14:05,700
And suddenly the
organizations may not even
318
00:14:05,700 --> 00:14:09,750
have complete visibility into
which AI tools are being used,
319
00:14:09,750 --> 00:14:11,760
what information is being shared,
320
00:14:11,760 --> 00:14:14,340
or where that information
is going, you know?
321
00:14:14,340 --> 00:14:17,370
And again, AI is just one part of it.
322
00:14:17,370 --> 00:14:19,443
The other one, OT.
323
00:14:20,340 --> 00:14:23,430
- Yeah, I mean, OT, as I mentioned before,
324
00:14:23,430 --> 00:14:25,840
the operation technology
side is a big part
325
00:14:25,840 --> 00:14:28,500
of it 'cause we have
seen a huge proliferation
326
00:14:28,500 --> 00:14:31,080
of technology providers in the space.
327
00:14:31,080 --> 00:14:33,240
But the reality is not
only on the OT side,
328
00:14:33,240 --> 00:14:37,170
it's also on the maybe IT
side of supply chains, right?
329
00:14:37,170 --> 00:14:39,150
If we think about, like, maybe 10,
330
00:14:39,150 --> 00:14:42,900
15 years ago, and e-commerce
had a lot to do with it,
331
00:14:42,900 --> 00:14:46,470
the amount of technology
suppliers was pretty much,
332
00:14:46,470 --> 00:14:48,600
like, I would say, a small number.
333
00:14:48,600 --> 00:14:51,450
The big suppliers of supply
chain technology were a few.
334
00:14:51,450 --> 00:14:52,500
You could list them across
335
00:14:52,500 --> 00:14:54,210
in a warehouse management systems,
336
00:14:54,210 --> 00:14:56,160
transportation management systems,
337
00:14:56,160 --> 00:14:59,430
of course, you know,
the big ERPs and so on.
338
00:14:59,430 --> 00:15:02,250
Now there are so many
providers of everything.
339
00:15:02,250 --> 00:15:03,690
Even, you know, if you think
340
00:15:03,690 --> 00:15:07,320
about robotics technologies
that, like, in warehousing,
341
00:15:07,320 --> 00:15:10,820
which is my background, like, again, 10,
342
00:15:10,820 --> 00:15:15,060
15 years ago we had just a
few that were kind of reliable
343
00:15:15,060 --> 00:15:19,830
and were, I would say, you
know, top tier in the field.
344
00:15:19,830 --> 00:15:21,780
Nowadays you have hundreds of them.
345
00:15:21,780 --> 00:15:26,190
So that proliferation
of, you know, suppliers,
346
00:15:26,190 --> 00:15:29,460
usually small, like a lot of
them are even startups, right?
347
00:15:29,460 --> 00:15:30,930
That means that most
348
00:15:30,930 --> 00:15:34,080
of those companies do not
have the cybersecurity,
349
00:15:34,080 --> 00:15:37,680
you know, capabilities that a
lot of the big players have.
350
00:15:37,680 --> 00:15:42,270
That is a huge, you know,
open door for cybercrime.
351
00:15:42,270 --> 00:15:44,400
And at the same time, we're implementing,
352
00:15:44,400 --> 00:15:46,140
again, more technologies.
353
00:15:46,140 --> 00:15:47,940
That means, as Camilo mentioned,
354
00:15:47,940 --> 00:15:51,450
that we have basically a
bigger attack surface, right?
355
00:15:51,450 --> 00:15:53,580
As you have more
technologies interconnected,
356
00:15:53,580 --> 00:15:55,560
your attack surface just increases.
357
00:15:55,560 -->
00:15:56,820
Attack surface meaning where
358
00:15:56,820 --> 00:15:58,380
the attacks can come from, right?
359
00:15:58,380 --> 00:15:59,780
I remember now that one
360
00:15:59,780 --> 00:16:02,700
of the main logistics providers worldwide,
361
00:16:02,700 --> 00:16:05,430
having more than 250
different technology providers
362
00:16:05,430 --> 00:16:07,020
for the warehouses worldwide,
363
00:16:07,020 --> 00:16:09,960
even though each of the
warehouses usually uses only
364
00:16:09,960 --> 00:16:12,180
between 10 to 20 different technologies.
365
00:16:12,180 --> 00:16:15,150
But still, because of course
every facility is different,
366
00:16:15,150 --> 00:16:17,490
you end up with 250 different.
367
00:16:17,490 --> 00:16:19,020
Are you gonna be able
to really control all
368
00:16:19,020 --> 00:16:20,580
of them to, you know, audit them
369
00:16:20,580 --> 00:16:22,110
from a cybersecurity perspective?
370
00:16:22,110 --> 00:16:23,670
Of course not. It's super hard.
371
00:16:23,670 --> 00:16:26,367
And also for attackers,
they know that, right?
372
00:16:26,367 --> 00:16:29,310
And these tech providers
are the weakest links.
373
00:16:29,310 --> 00:16:33,390
So that is one of the, I would
say, changing, like, fastest,
374
00:16:33,390 --> 00:16:36,040
you know, changing things
that we're seeing now
375
00:16:36,960 --> 00:16:38,880
in supply chains with modern technology,
376
00:16:38,880 --> 00:16:41,580
that technology providers
are becoming the new target
377
00:16:41,580 --> 00:16:45,300
for cybercrime 'cause once
an attacker hits one of them,
378
00:16:45,300 --> 00:16:46,890
basically the impact then cascades
379
00:16:46,890 --> 00:16:49,230
across the different organizations too.
380
00:16:49,230 --> 00:16:50,820
And there are many examples of those.
381
00:16:50,820 --> 00:16:54,820
I would say maybe one that I
suffered directly last year
382
00:16:55,888 --> 00:16:58,380
was a pretty good example
'cause it happened in Europe,
383
00:16:58,380 --> 00:17:01,170
and it happened to many
different airports at once.
384
00:17:01,170 --> 00:17:02,413
- Yes.
385
00:17:02,413 --> 00:17:04,620
- 'Cause the idea of how can you actually
386
00:17:04,620 --> 00:17:07,680
shut down more than one
airport at once, well,
387
00:17:07,680 --> 00:17:09,960
I mean, probably we couldn't
do it, like, 10 years ago.
388
00:17:09,960 --> 00:17:13,530
But now these guys, they
basically target a small supplier
389
00:17:13,530 --> 00:17:16,620
that nobody was paying
attention to probably,
390
00:17:16,620 --> 00:17:20,430
that was handling the check-in
and boarding processes.
391
00:17:20,430 --> 00:17:22,020
So a small software supplier,
392
00:17:22,020 --> 00:17:26,700
and basically they had to
freeze all the check-in kiosks.
393
00:17:26,700 --> 00:17:29,160
They halted all the, you
know, automated baggage
394
00:17:29,160 --> 00:17:32,130
tag printers across more
than four airports in Europe.
395
00:17:32,130 --> 00:17:33,630
I think it was Brussels,
396
00:17:33,630 --> 00:17:35,430
Berlin, Heathrow in London,
397
00:17:35,430 --> 00:17:38,070
and I think Dublin in Ireland, so.
398
00:17:38,070 --> 00:17:39,270
- And was that one just for havoc?
399
00:17:39,270 --> 00:17:40,560
There was no ransom.
400
00:17:40,560 --> 00:17:42,780
- So that one was also ransom, but again-
401
00:17:42,780 --> 00:17:43,613
- It was.
402
00:17:43,613 --> 00:17:45,450
- They didn't ask for,
like, a payment again.
403
00:17:45,450 --> 00:17:47,727
So yeah, the goal was
disrupting operations, yeah.
404
00:17:47,727 --> 00:17:49,830
- And in manufacturing,
405
00:17:49,830 --> 00:17:51,000
something similar is happening.
406
00:17:51,000 --> 00:17:55,383
So many companies still operate
equipment and legacy systems
407
00:17:55,383 --> 00:18:00,360
that were designed 10,
20, or even 30 years ago.
408
00:18:00,360 --> 00:18:02,370
And originally most
409
00:18:02,370 --> 00:18:05,250
of those systems were relatively isolated.
410
00:18:05,250 --> 00:18:07,080
They were not designed to be connected
411
00:18:07,080 --> 00:18:09,600
to the internet, to the cloud.
412
00:18:09,600 --> 00:18:13,830
So today those systems
are increasingly connected
413
00:18:13,830 --> 00:18:17,100
to IT systems, cloud environments,
414
00:18:17,100 --> 00:18:22,100
vendors, as you were describing,
and remote access tools.
415
00:18:22,590 --> 00:18:27,570
So you have old technology
operating in a new,
416
00:18:27,570 --> 00:18:30,030
highly connected environment.
417
00:18:30,030 --> 00:18:31,707
And as you said, Miguel,
418
00:18:31,707 --> 00:18:33,480
(laughs) attackers are using that.
419
00:18:33,480 --> 00:18:35,580
- Yeah, yeah, yeah, and
you cannot patch them,
420
00:18:35,580 --> 00:18:36,810
most of the time, so.
421
00:18:36,810 --> 00:18:38,550
- Our vision, just to give,
422
00:18:38,550 --> 00:18:41,280
like, an overview of our research approach
423
00:18:41,280 --> 00:18:43,710
and how we are translating
that into certain,
424
00:18:43,710 --> 00:18:47,760
like, tools, is, like, we
have our research agenda.
425
00:18:47,760 --> 00:18:51,660
We wanna learn from the past,
so for that we built TRACE.
426
00:18:51,660 --> 00:18:52,800
You know, look at the past,
427
00:18:52,800 --> 00:18:55,020
learn from those incidents, et cetera.
428
00:18:55,020 --> 00:18:56,640
We are building another one,
429
00:18:56,640 --> 00:19:00,030
which is RADAR, to anticipate the future.
430
00:19:00,030 --> 00:19:04,080
Essentially it's to
identify emerging risks,
431
00:19:04,080 --> 00:19:06,840
threats that could affect supply chains
432
00:19:06,840 --> 00:19:09,270
from the standpoint of cybersecurity.
433
00:19:09,270 --> 00:19:14,270
And by connecting the inputs
or the outcomes from those two,
434
00:19:14,490 --> 00:19:17,700
we are building SCOUT
to shape the present.
435
00:19:17,700 --> 00:19:21,180
So SCOUT stands for Supply Chain Operation
436
00:19:21,180 --> 00:19:24,120
and Vulnerability Threat Assessment.
437
00:19:24,120 --> 00:19:28,320
And companies have, you know,
hundreds or even thousands
438
00:19:28,320 --> 00:19:33,320
of unknown vulnerabilities
across the organization,
439
00:19:33,390 --> 00:19:35,250
and also upstream.
440
00:19:35,250 --> 00:19:37,110
So the idea with SCOUT,
441
00:19:37,110 --> 00:19:40,740
imagine a platform where your
employees actually interact
442
00:19:40,740 --> 00:19:45,180
with it, learn by interacting,
which is another key piece,
443
00:19:45,180 --> 00:19:49,230
and also you can get a quick
assessment of your company
444
00:19:49,230 --> 00:19:53,310
across your different
functional areas and levels.
445
00:19:53,310 --> 00:19:58,260
And with that then, like,
determine a roadmap,
446
00:19:58,260 --> 00:20:01,590
prioritize what to do next, you know,
447
00:20:01,590 --> 00:20:04,590
and also being, like,
dynamic because again,
448
00:20:04,590 --> 00:20:06,540
like, technology is nowadays changing
449
00:20:06,540 --> 00:20:10,920
so fast that the traditional
classic assessment, you know,
450
00:20:10,920 --> 00:20:15,150
checklist that they do once a
year, you know, it's outdated.
451
00:20:15,150 --> 00:20:16,470
Even if they finish today,
452
00:20:16,470 --> 00:20:18,780
maybe tomorrow it's
already outdated, you know?
453
00:20:18,780 --> 00:20:20,310
So that is a SCOUT.
454
00:20:20,310 --> 00:20:23,190
It's a tool that we're
creating for organizations.
455
00:20:23,190 --> 00:20:25,500
Maybe large companies have their own
456
00:20:25,500 --> 00:20:28,770
vulnerability management
platforms or systems,
457
00:20:28,770 --> 00:20:32,820
but what about the tier
one, two, three suppliers?
458
00:20:32,820 --> 00:20:35,130
Maybe they have some certain visibility
459
00:20:35,130 --> 00:20:38,820
and control on their tier one
suppliers, but, like, in terms
460
00:20:38,820 --> 00:20:41,340
of compliance, but what about the others?
461
00:20:41,340 --> 00:20:43,612
- You know, we've talked
about how it needs
462
00:20:43,612 --> 00:20:45,690
to be beyond IT teams
within organizations looking
463
00:20:45,690 --> 00:20:47,100
at cybersecurity.
464
00:20:47,100 --> 00:20:51,240
What kinds of roles or teams
are you all envisioning
465
00:20:51,240 --> 00:20:53,670
as using something like SCOUT?
466
00:20:53,670 --> 00:20:55,518
Are we looking at operations teams?
467
00:20:55,518 --> 00:20:58,012
- Everyone. (laughs)
468
00:20:58,012 --> 00:20:59,130
Yeah.
469
00:20:59,130 --> 00:20:59,963
Everyone across the organization.
470
00:20:59,963 --> 00:21:02,400
In terms of cyber, in terms of security,
471
00:21:02,400 --> 00:21:04,323
it's a matter of everyone.
472
00:21:05,310 --> 00:21:08,310
As Miguel was mentioning, you
have the case of the airports,
473
00:21:08,310 --> 00:21:10,800
you know, and maybe, you know,
474
00:21:10,800 --> 00:21:14,340
you kind of create certain barriers
475
00:21:14,340 --> 00:21:17,700
or kind of protect key
things such as, you know,
476
00:21:17,700 --> 00:21:19,260
the radar or, you
know,
477
00:21:19,260 --> 00:21:22,470
like navigation systems
of the planes, whatever.
478
00:21:22,470 --> 00:21:24,600
They never thought that the way
479
00:21:24,600 --> 00:21:27,720
to create a disruption was, you know,
480
00:21:27,720 --> 00:21:30,540
that you cannot check in luggage, right?
481
00:21:30,540 --> 00:21:31,380
- As simple as that, yeah.
482
00:21:31,380 --> 00:21:32,910
- As simple as that.
483
00:21:32,910 --> 00:21:34,200
- So we've talked about some things
484
00:21:34,200 --> 00:21:36,470
that are changing quickly
in cybersecurity and some
485
00:21:36,470 --> 00:21:40,350
of the tools that you all
are offering through the lab.
486
00:21:40,350 --> 00:21:43,140
Can you give us a kind of
overview of the current challenges
487
00:21:43,140 --> 00:21:46,890
that companies face in supply
chain cybersecurity right now?
488
00:21:46,890 --> 00:21:48,870
- We look at, for instance now,
489
00:21:48,870 --> 00:21:50,190
supply chain cybersecurity
490
00:21:50,190 --> 00:21:53,940
in two kind of like axes,
and that's something
491
00:21:53,940 --> 00:21:58,590
that we started doing 'cause
I like the theoretical part
492
00:21:58,590 --> 00:21:59,880
of research, right?
493
00:21:59,880 --> 00:22:02,610
So I like to read from
other different fields
494
00:22:02,610 --> 00:22:03,960
and bring ideas together.
495
00:22:03,960 --> 00:22:05,397
So when I was thinking about cyber
496
00:22:05,397 --> 00:22:07,440
and the scale of these disruptions,
497
00:22:07,440 --> 00:22:10,530
I went back and read about,
for instance, like, you know,
498
00:22:10,530 --> 00:22:14,670
nuclear reactors basically
having issues and creating some
499
00:22:14,670 --> 00:22:18,630
of the, you know, largest
disasters that we've had on Earth,
500
00:22:18,630 --> 00:22:22,830
but also spaceships going
burst, which is also something
501
00:22:22,830 --> 00:22:24,210
that is very, very hard to, you know,
502
00:22:24,210 --> 00:22:26,427
to control and to predict, right?
503
00:22:26,427 --> 00:22:27,630
And when you launch something to space,
504
00:22:27,630 --> 00:22:31,380
there is so much uncertainty
in what can go wrong, right?
505
00:22:31,380 --> 00:22:33,840
So what we realized is that when you look
506
00:22:33,840 --> 00:22:36,390
at supply chain cybersecurity nowadays,
507
00:22:36,390 --> 00:22:39,210
it's moving across two axes
that are related to that.
508
00:22:39,210 --> 00:22:41,400
One is the level of what we call coupling,
509
00:22:41,400 --> 00:22:44,970
and coupling meaning the
interdependence of systems, right?
510
00:22:44,970 --> 00:22:47,490
When you think about, again,
511
00:22:47,490 --> 00:22:50,493
a spaceship that basically
goes, you know, to space,
512
00:22:51,390 --> 00:22:55,860
like when something goes wrong
in one part of that ship,
513
00:22:55,860 --> 00:22:59,160
it's so easy for that
issue to basically then,
514
00:22:59,160 --> 00:23:00,983
you know, travel to any
other part of the ship.
515
00:23:00,983 --> 00:23:04,350
Like if something starts
heating up, the next, you know,
516
00:23:04,350 --> 00:23:07,020
metal part is also gonna be
heating up, things like that.
517
00:23:07,020 --> 00:23:10,200
Nowadays in supply chains and
cybersecurity specifically
518
00:23:10,200 --> 00:23:11,850
because of the amount of interconnectivity
519
00:23:11,850 --> 00:23:13,590
that we have in IT systems,
520
00:23:13,590 --> 00:23:17,670
but also those IT systems
connected to the, you know, OT
521
00:23:17,670 --> 00:23:20,670
or physical equipment environment,
as Camilo was mentioning,
522
00:23:20,670 --> 00:23:23,550
makes it very, very easy for
those kind of like cyberattacks
523
00:23:23,550 --> 00:23:24,900
to cascade very fast, right?
524
00:23:24,900 --> 00:23:27,060
Somebody targeting one system
525
00:23:27,060 --> 00:23:29,100
can shut down a company right away.
526
00:23:29,100 --> 00:23:32,370
So that is, like, one of the
biggest challenges in the sense
527
00:23:32,370 --> 00:23:34,500
that the level of coupling
makes it very, very hard
528
00:23:34,500 --> 00:23:37,230
to basically stop, even if, of course,
529
00:23:37,230 --> 00:23:38,970
there is a lot of, you
know, network segmentation
530
00:23:38,970 --> 00:23:41,610
that can be done, and
cybersecurity teams have been doing
531
00:23:41,610 --> 00:23:44,730
that for years, but it's
much harder than before.
532
00:23:44,730 --> 00:23:46,440
And the other one I
would say is complexity.
533
00:23:46,440 --> 00:23:49,950
That's usually when you look
at catastrophic disasters,
534
00:23:49,950 --> 00:23:52,260
being able to explain
what went wrong, right?
535
00:23:52,260 --> 00:23:54,360
And how the system is gonna behave.
536
00:23:54,360 --> 00:23:56,010
For instance, in space,
537
00:23:56,010 --> 00:23:59,160
there's so many things that
we don't even know, right?
538
00:23:59,160 --> 00:24:00,723
How reactions are gonna
happen, things like that.
539
00:24:00,723 --> 00:24:03,120
It's almost impossible
to predict sometimes.
540
00:24:03,120 --> 00:24:04,770
With cyber nowadays, for instance,
541
00:24:04,770 --> 00:24:07,050
because we have a lot of
AI that is being deployed
542
00:24:07,050 --> 00:24:10,890
in modern supply chain
technologies, all across the board,
543
00:24:10,890 --> 00:24:13,020
that is gonna become also an issue, right?
544
00:24:13,020 --> 00:24:14,730
We don't know a lot of the times
545
00:24:14,730 --> 00:24:16,470
how these systems are gonna behave, right?
546
00:24:16,470 --> 00:24:18,840
I mean, that's part of
artificial intelligence
547
00:24:18,840 --> 00:24:20,460
that is nondeterministic.
548
00:24:20,460 --> 00:24:23,692
So that's gonna become, I
think, in the next, you know,
549
00:24:23,692 --> 00:24:25,010
five to 10 years, another big challenge
550
00:24:25,010 --> 00:24:28,110
in terms of how do we
deal with that uncertainty
551
00:24:28,110 --> 00:24:30,660
in terms of you have a bunch
of different, you know,
552
00:24:30,660 --> 00:24:34,587
AI tools
deployed, maybe AI
agents in your supply chain,
553
00:24:34,587 --> 00:24:37,740
and all of them start maybe misbehaving,
554
00:24:37,740 --> 00:24:39,957
but you don't understand it, right?
555
00:24:39,957 --> 00:24:42,630
And that can scale and
cascade pretty fast.
556
00:24:42,630 --> 00:24:44,490
- Yeah, and also I would say
557
00:24:44,490 --> 00:24:47,910
that another important
challenge is talent, right?
558
00:24:47,910 --> 00:24:50,700
So, there is a significant shortage
559
00:24:50,700 --> 00:24:53,460
of cybersecurity professionals worldwide.
560
00:24:53,460 --> 00:24:57,180
Some organizations estimate
that it's around 3 million,
561
00:24:57,180 --> 00:25:02,180
like, people, others around 5
million people, and, you know,
562
00:25:02,610 --> 00:25:05,910
it's a matter of, in this
case, not just about cyber,
563
00:25:05,910 --> 00:25:09,123
but having, you know, the
right combination of skills.
564
00:25:10,200 --> 00:25:11,550
There is a question, like, we want
565
00:25:11,550 --> 00:25:16,080
to also address with the lab,
is how to upskill, you know,
566
00:25:16,080 --> 00:25:21,000
supply chainers that have
no clue about these threats.
567
00:25:21,000 --> 00:25:22,830
- Yeah, and we already have an idea
568
00:25:22,830 --> 00:25:25,650
of a new position that
we are gonna be probably,
569
00:25:25,650 --> 00:25:26,610
you know, naming.
570
00:25:26,610 --> 00:25:27,443
What is it?
571
00:25:27,443 --> 00:25:30,460
- Sorry.
- Supply chain CISO. (laughs)
572
00:25:34,200 --> 00:25:37,980
Probably we are gonna have
dedicated CISOs per function
573
00:25:37,980 --> 00:25:39,330
in the organization, right?
574
00:25:39,330 --> 00:25:40,290
Most of the companies,
575
00:25:40,290 --> 00:25:44,340
they do have a global chief
information security officer,
576
00:25:44,340 --> 00:25:46,050
but we know certain
industries, for instance,
577
00:25:46,050 --> 00:25:48,960
the financial institutions,
which are usually,
578
00:25:48,960 --> 00:25:50,670
like, ahead when it
comes to cyber (laughs)
579
00:25:50,670 --> 00:25:54,151
because of, you know, what
it entails to protect banks.
580
00:25:54,151 --> 00:25:54,984
- Yeah.
581
00:25:54,984 --> 00:25:57,360
- They do have different,
582
00:25:57,360 --> 00:26:00,510
like, CISO positions for,
like, specific verticals
583
00:26:00,510 --> 00:26:01,350
of the company, right?
584
00:26:01,350 --> 00:26:03,030
One might be for commercial banking,
585
00:26:03,030 --> 00:26:06,660
one might be for, like,
internal, you know, IT systems.
586
00:26:06,660 --> 00:26:09,780
So, to be honest,
considering the deployment
587
00:26:09,780 --> 00:26:11,760
of technology all across the world,
588
00:26:11,760 --> 00:26:13,980
and specifically in
supply chain operations,
589
00:26:13,980 --> 00:26:18,390
we're gonna need supply chain CISOs soon.
590
00:26:18,390 --> 00:26:20,640
So specific people that look at how
591
00:26:20,640 --> 00:26:23,110
cyberattacks can impact physical
operations in companies.
592
00:26:23,110 --> 00:26:24,540
- So we have a model in banking,
593
00:26:24,540 --> 00:26:26,487
and it should cascade throughout,
right, to protect that.
594
00:26:26,487 --> 00:26:29,490
- The benchmarking part is, and it's great
595
00:26:29,490 --> 00:26:30,990
that you say that, Mackenzie,
596
00:26:30,990 --> 00:26:33,060
'cause one of the biggest
things I think that we can do
597
00:26:33,060 --> 00:26:36,450
from the research perspective,
like, at MIT, is that.
598
00:26:36,450 --> 00:26:37,350
Is that kinda, like, you know,
599
00:26:37,350 --> 00:26:39,690
cross-pollination of knowledge
from different industries.
600
00:26:39,690 --> 00:26:40,830
So benchmarking, okay,
601
00:26:40,830 --> 00:26:44,250
what industries maybe are
ahead when it comes to cyber?
602
00:26:44,250 --> 00:26:46,950
Where industries are ahead
when it comes to supply chain?
603
00:26:46,950 --> 00:26:48,300
Bring all that knowledge together.
604
00:26:48,300 --> 00:26:49,133
- Certainly.
605
00:26:49,133 --> 00:26:50,760
In terms of supply chain,
606
00:26:50,760 --> 00:26:54,930
cybersecurity is not
just about technology.
607
00:26:54,930 --> 00:26:58,710
It's about people, processes, leadership.
608
00:26:58,710 --> 00:27:01,620
So those new positions, those new roles,
609
00:27:01,620 --> 00:27:05,310
you know, for these challenges
of the 21st century, right?
610
00:27:05,310 --> 00:27:07,140
So imagine, you know,
611
00:27:07,140 --> 00:27:12,140
that, like, how a company
manages a catastrophic event
612
00:27:13,680 --> 00:27:15,630
or this kind of crisis.
613
00:27:15,630 --> 00:27:17,580
Who makes the decisions, right?
614
00:27:17,580 --> 00:27:21,330
We were, you know,
discussing this with leaders,
615
00:27:21,330 --> 00:27:23,230
you know, from different companies
616
00:27:24,084 --> 00:27:27,720
in our roundtable that
we conducted here at MIT,
617
00:27:27,720 --> 00:27:30,390
and then the question is,
like, who communicates
618
00:27:30,390 --> 00:27:32,940
with suppliers and customers?
619
00:27:32,940 --> 00:27:35,910
Which systems should be isolated?
620
00:27:35,910 --> 00:27:40,910
Which operations can actually
continue operating normally,
621
00:27:41,220 --> 00:27:43,260
or manually, right?
622
00:27:43,260 --> 00:27:47,010
So the companies should train people
623
00:27:47,010 --> 00:27:50,520
on how to deal with these circumstances,
624
00:27:50,520 --> 00:27:52,860
you know, for the decision-making process.
625
00:27:52,860 --> 00:27:54,420
- Well, that's an interesting question,
626
00:27:54,420 --> 00:27:57,000
because we're having a push for companies
627
00:27:57,000 --> 00:27:59,910
to use more technology, be more digital,
628
00:27:59,910 --> 00:28:02,370
and yet if there is this
kind of catastrophe,
629
00:28:02,370 --> 00:28:05,010
of course, that knocks
those digital systems out,
630
00:28:05,010 --> 00:28:07,740
you have to have people who
know how to do things manually.
631
00:28:07,740 --> 00:28:12,580
How do you
all suggest
companies finding that balance?
632
00:28:15,300 --> 00:28:17,370
- I mean, given
633
00:28:17,370 --> 00:28:20,760
that they already experienced
these disruptions.
634
00:28:20,760 --> 00:28:23,640
So in our interviews, we have tried
635
00:28:23,640 --> 00:28:28,640
to understand what they do,
you know, when this happens.
636
00:28:30,090 --> 00:28:32,850
So as I described before,
we have the silos.
637
00:28:32,850 --> 00:28:35,970
So the IT team is trying
to recover the systems,
638
00:28:35,970 --> 00:28:39,270
while the supply chainers
are trying to, you know,
639
00:28:39,270 --> 00:28:40,950
continue operations.
640
00:28:40,950 --> 00:28:44,100
So this is, of course, contingent
to the type of company.
641
00:28:44,100 --> 00:28:46,410
And so, but some of them, like, "Well,
642
00:28:46,410 --> 00:28:50,130
we started creating our new systems,"
643
00:28:50,130 --> 00:28:54,690
like, communication
channels like WhatsApp.
644
00:28:54,690 --> 00:28:57,150
I don't know, getting information
645
00:28:57,150 --> 00:29:00,150
from some spreadsheets, phone calls.
646
00:29:00,150 --> 00:29:02,304
You know, like, going back to 1980.
647
00:29:02,304 --> 00:29:03,431
- Back to spreadsheets and phone calls.
648
00:29:03,431 --> 00:29:04,264
(Camilo laughs)
Yes.
649
00:29:04,264 --> 00:29:05,097
- Right?
650
00:29:05,097 --> 00:29:08,820
It's like turning the switch
off and operating manually.
651
00:29:08,820 --> 00:29:12,900
Because again, even if the
IT team says, like, "Oh,
652
00:29:12,900 --> 00:29:15,723
we got, like, the systems
restored," and whatever,
653
00:29:16,890 --> 00:29:21,120
operations is like, "Can
we trust that system?
654
00:29:21,120 --> 00:29:23,250
Is the malicious actor still there?"
655
00:29:23,250 --> 00:29:25,380
- Yeah, and there are a
lot of trade-offs there.
656
00:29:25,380 --> 00:29:27,300
'Cause for instance, okay,
657
00:29:27,300 --> 00:29:30,570
we have the IT system restored
or the main IT system,
658
00:29:30,570 --> 00:29:33,030
but what about all the
shadow AI, you know,
659
00:29:33,030 --> 00:29:36,780
tool that everybody's using
already but was not official?
660
00:29:36,780 --> 00:29:37,920
Is that actually restored?
661
00:29:37,920 --> 00:29:40,470
Or a lot of the times,
all the Excel macros
662
00:29:40,470 --> 00:29:42,840
that people were using for the daily work,
663
00:29:42,840 --> 00:29:44,520
and there is a real example
664
00:29:44,520 --> 00:29:47,100
of a big company that
suffered that, right?
665
00:29:47,100 --> 00:29:48,240
Not too long ago.
666
00:29:48,240 --> 00:29:50,340
Like, their systems went completely down.
667
00:29:50,340 --> 00:29:52,440
They have one of the best
cybersecurity teams worldwide,
668
00:29:52,440 --> 00:29:54,450
I would say, and they recovered
669
00:29:54,450 --> 00:29:58,200
the main systems pretty fast,
in less than a day. (laughs)
670
00:29:58,200 --> 00:29:59,370
what they didn't know is that people
671
00:29:59,370 --> 00:30:02,650
were actually using hundreds
of other systems in the back,
672
00:30:02,650 --> 00:30:04,920
like, and those, of course,
673
00:30:04,920 --> 00:30:07,020
couldn't be restored
officially 'cause nobody knew
674
00:30:07,020 --> 00:30:10,230
about them but the specific
teams that were using them.
675
00:30:10,230 --> 00:30:13,140
So really getting back to
the level of productivity
676
00:30:13,140 --> 00:30:16,470
that they had before restoring
all those shadow systems,
677
00:30:16,470 --> 00:30:18,000
that usually takes a lot longer.
678
00:30:18,000 --> 00:30:21,120
And the same can happen
thinking the other way around.
679
00:30:21,120 --> 00:30:23,280
So you might be able to restore,
680
00:30:23,280 --> 00:30:25,327
for instance, the operational
technology and say,
681
00:30:25,327 --> 00:30:26,760
"Okay, the equipment is ready."
682
00:30:26,760 --> 00:30:29,640
'Cause you need to make sure
that your systems are secure,
683
00:30:29,640 --> 00:30:32,280
that the attackers are gone
and not again in the system.
684
00:30:32,280 --> 00:30:34,200
So you need to do a lot of forensics
685
00:30:34,200 --> 00:30:36,483
for that to be, let's say, ready to go.
686
00:30:37,380 --> 00:30:39,990
But still, a lot of the data
687
00:30:39,990 --> 00:30:42,030
is probably not
synchronized anymore, right?
688
00:30:42,030 --> 00:30:43,320
After you went through an attack,
689
00:30:43,320 --> 00:30:46,200
you probably had to do a
bunch of, you know, cleaning.
690
00:30:46,200 --> 00:30:47,790
A lot of the data that was coming in,
691
00:30:47,790 --> 00:30:50,100
as you shut down your systems,
692
00:30:50,100 --> 00:30:53,490
it stopped, so probably you're
gonna have a gap in between.
693
00:30:53,490 --> 00:30:55,560
And a lot of the new technologies,
694
00:30:55,560 --> 00:30:57,570
they require, like, updated data
695
00:30:57,570 --> 00:31:01,110
and the pretty synchronized
data to keep moving, right?
696
00:31:01,110 --> 00:31:03,600
So even if the robot
697
00:31:03,600 --> 00:31:06,600
or the data specific
technology is ready to go,
698
00:31:06,600 --> 00:31:09,840
a lot of this data
synchronization takes a long time
699
00:31:09,840 --> 00:31:10,920
to really get back to it.
700
00:31:10,920 --> 00:31:13,950
So, and that, again, this example,
701
00:31:13,950 --> 00:31:17,250
the OT, the equipment, the
physical equipment was recovered
702
00:31:17,250 --> 00:31:18,480
in less than eight hours.
703
00:31:18,480 --> 00:31:21,270
Now, really to get back
to the productivity levels
704
00:31:21,270 --> 00:31:23,820
and get back to have the perfect data
705
00:31:23,820 --> 00:31:26,220
to keep operating took them days.
706
00:31:26,220 --> 00:31:27,777
- Wow.
- Yeah.
707
00:31:27,777 --> 00:31:29,880
- And for leaders
listening who are wanting
708
00:31:29,880 --> 00:31:33,000
to improve the cybersecurity
in their organizations,
709
00:31:33,000 --> 00:31:34,560
of course, everything is a trade-off,
710
00:31:34,560 --> 00:31:37,320
what people are, you
know,
choosing to focus on.
711
00:31:37,320 --> 00:31:39,000
What trade-offs must leaders make
712
00:31:39,000 --> 00:31:41,880
when looking at
strengthening cybersecurity?
713
00:31:41,880 --> 00:31:44,430
- So I think we've
mentioned a few already.
714
00:31:44,430 --> 00:31:46,740
This one, for instance,
about the, you know,
715
00:31:46,740 --> 00:31:48,750
how fast you wanna recover
716
00:31:48,750 --> 00:31:51,573
and how many forensics you
wanna do, that's a big one.
717
00:31:52,470 --> 00:31:54,240
You can try to recover very fast
718
00:31:54,240 --> 00:31:57,300
and basically reset all
your systems without looking
719
00:31:57,300 --> 00:31:58,233
at what happened.
720
00:31:59,340 --> 00:32:01,680
And basically, the idea of doing
721
00:32:01,680 --> 00:32:06,037
that is to try to maximize
business continuity, right?
722
00:32:06,037 --> 00:32:08,400
"My plant needs to keep operating.
723
00:32:08,400 --> 00:32:10,380
I shut down everything, I reset it.
724
00:32:10,380 --> 00:32:12,240
In one hour, we are good to go."
725
00:32:12,240 --> 00:32:14,310
The issue with that is that
probably you had no time
726
00:32:14,310 --> 00:32:16,780
to do any forensics, so the
attackers might still be,
727
00:32:16,780 --> 00:32:17,760
- How much can you trust it, yeah?
728
00:32:17,760 --> 00:32:19,860
- Exactly. Like, are
your systems compromised?
729
00:32:19,860 --> 00:32:20,790
Probably you don't know.
730
00:32:20,790 --> 00:32:23,730
So the chances of having another attack,
731
00:32:23,730 --> 00:32:26,040
like, soon are much higher, of course.
732
00:32:26,040 --> 00:32:27,780
So that's one.
733
00:32:27,780 --> 00:32:31,350
We talked today a lot about the increasing
734
00:32:31,350 --> 00:32:32,973
in the amount of suppliers,
735
00:32:34,050 --> 00:32:37,920
mainly in the technology space
for supply chains, right?
736
00:32:37,920 --> 00:32:42,660
So usually in traditional risk
management in supply chain,
737
00:32:42,660 --> 00:32:44,610
we were always taught that, you know,
738
00:32:44,610 --> 00:32:47,580
you diversify suppliers
to basically de-risk.
739
00:32:47,580 --> 00:32:50,250
'Cause if something bad
happens to one of them,
740
00:32:50,250 --> 00:32:51,960
for whatever reason, you just,
741
00:32:51,960 --> 00:32:54,660
you know, swap it and you go to the next.
742
00:32:54,660 --> 00:32:57,630
In cyber, I mean, that still, you know,
743
00:32:57,630 --> 00:33:00,990
remains true, when it comes
to the physical products.
744
00:33:00,990 --> 00:33:04,230
But the issue is that as you
basically expand your number
745
00:33:04,230 --> 00:33:05,797
of tier one suppliers, you also expand,
746
00:33:05,797 --> 00:33:07,980
you know, your attack surface.
747
00:33:07,980 --> 00:33:10,020
And it's not only that direct supplier,
748
00:33:10,020 --> 00:33:11,580
but also all the tier two, three,
749
00:33:11,580 --> 00:33:14,190
four suppliers that
come with that supplier
750
00:33:14,190 --> 00:33:16,740
that are gonna be part of
your new attack surface.
751
00:33:16,740 --> 00:33:20,190
'Cause a cyberattack
basically cascades pretty fast
752
00:33:20,190 --> 00:33:21,660
through the supply chain.
753
00:33:21,660 --> 00:33:23,880
So that's a new important trade-off
754
00:33:23,880 --> 00:33:27,450
when it comes to how big is
gonna be your supplier base.
755
00:33:27,450 --> 00:33:29,850
'Cause you're de-risking from one side,
756
00:33:29,850 --> 00:33:32,850
but at the same time, you're
increasing your attack surface
757
00:33:32,850 --> 00:33:34,740
from a cybersecurity perspective.
758
00:33:34,740 --> 00:33:35,900
- And going back to the role
759
00:33:35,900 --> 00:33:37,800
of AI in cybersecurity,
760
00:33:37,800 --> 00:33:40,860
how is AI changing supply
chain cybersecurity,
761
00:33:40,860 --> 00:33:43,230
both as a threat and as a defense?
762
00:33:43,230 --> 00:33:47,850
- Well, companies can certainly
leverage AI for defense.
763
00:33:47,850 --> 00:33:51,960
So you know, they have these
mature cybersecurity teams
764
00:33:51,960 --> 00:33:55,650
that can, you know, have
access to certain models,
765
00:33:55,650 --> 00:34:00,650
improve or patch certain system
scripts, so on and so forth.
766
00:34:00,930 --> 00:34:01,980
But on the other hand,
767
00:34:01,980 --> 00:34:06,450
you have, like, these
malicious actors that can also,
768
00:34:06,450 --> 00:34:11,450
you know, use these new
models to potentially,
769
00:34:11,970 --> 00:34:13,530
you know, attack certain companies.
770
00:34:13,530 --> 00:34:16,440
Now, I believe that
something that's gonna happen
771
00:34:16,440 --> 00:34:21,060
is that these, let's say,
hackers are not gonna be able
772
00:34:21,060 --> 00:34:23,730
to really disrupt some large firms
773
00:34:23,730 --> 00:34:25,740
by attacking them directly.
774
00:34:25,740 --> 00:34:26,970
So they are going to go
775
00:34:26,970 --> 00:34:30,540
after small and medium-sized companies
776
00:34:30,540 --> 00:34:34,230
that do not have the capabilities,
777
00:34:34,230 --> 00:34:37,590
the resources, the
muscle to really prepare,
778
00:34:37,590 --> 00:34:39,600
you know, and defend from these attackers.
779
00:34:39,600 --> 00:34:42,150
They are not even aware of that.
780
00:34:42,150 --> 00:34:46,650
Small, medium-sized companies
that, with vibe coding,
781
00:34:46,650 --> 00:34:49,440
are developing certain tools, you know,
782
00:34:49,440 --> 00:34:52,860
that's great, you know, to
be, like, more efficient.
783
00:34:52,860 --> 00:34:55,980
But at the end of the day,
they do not know what's
784
00:34:55,980 --> 00:35:00,180
in the back end, and
hackers can exploit that.
785
00:35:00,180 --> 00:35:02,100
- And we have seen that
trend already, right?
786
00:35:02,100 --> 00:35:03,120
When we were talking
787
00:35:03,120 --> 00:35:05,820
about the technology
providers being the ones hit,
788
00:35:05,820 --> 00:35:09,240
'cause usually those tech
providers of the main firms,
789
00:35:09,240 --> 00:35:10,890
they're
actually smaller in size.
790
00:35:10,890 --> 00:35:14,220
They're SMEs, lower
cybersecurity capabilities.
791
00:35:14,220 --> 00:35:16,410
And of course, again,
yeah, with tech providers,
792
00:35:16,410 --> 00:35:18,870
it happens that the return
of investment of the attacker
793
00:35:18,870 --> 00:35:21,420
is the highest 'cause they can
have this cascading effect.
794
00:35:21,420 --> 00:35:23,730
Actually, in the
roundtable that we hosted,
795
00:35:23,730 --> 00:35:25,050
we did see a big difference, right?
796
00:35:25,050 --> 00:35:28,740
In certain companies being
very open about using AI
797
00:35:28,740 --> 00:35:32,250
and deploying it, others
being more reluctant for now,
798
00:35:32,250 --> 00:35:34,473
a little bit, yeah, hesitant.
799
00:35:35,632 --> 00:35:37,590
But still, you know,
those large companies,
800
00:35:37,590 --> 00:35:39,420
probably they are gonna have, you know,
801
00:35:39,420 --> 00:35:41,400
clear procedures of, okay,
802
00:35:41,400 --> 00:35:44,670
how do you take an AI tool
that someone is building
803
00:35:44,670 --> 00:35:47,520
to really be, like in the production stage
804
00:35:47,520 --> 00:35:49,350
and ready to deploy.
805
00:35:49,350 --> 00:35:51,030
But SMEs are gonna be deploying
806
00:35:51,030 --> 00:35:53,220
those, like, super fast 'cause of course,
807
00:35:53,220 --> 00:35:55,890
when you're a small company,
all that matters is business,
808
00:35:55,890 --> 00:35:58,980
and you don't have money usually
to pay for cybersecurity.
809
00:35:58,980 --> 00:36:00,330
And I think a good example
810
00:36:00,330 --> 00:36:02,820
of that actually is that, for instance,
811
00:36:02,820 --> 00:36:07,820
for the defense industrial
base of the U.S.,
812
00:36:08,250 --> 00:36:10,200
of course, you have the big primes, right?
813
00:36:10,200 --> 00:36:13,170
Lockheed, Northrop, and others.
814
00:36:13,170 --> 00:36:16,620
But there is a huge amount of small,
815
00:36:16,620 --> 00:36:19,980
medium-sized suppliers that
provide to those, right?
816
00:36:19,980 --> 00:36:21,990
Early this year, and throughout the year,
817
00:36:21,990 --> 00:36:23,280
they're having a lot of conversations
818
00:36:23,280 --> 00:36:25,440
in terms of increasing
the level of cybersecurity
819
00:36:25,440 --> 00:36:28,200
for a lot of those small
and medium-sized firms.
820
00:36:28,200 --> 00:36:30,840
And actually, the
government wanted to pass
821
00:36:30,840 --> 00:36:35,840
a new regulation called the
Cybersecurity Maturity Model,
822
00:36:35,850 --> 00:36:40,500
and they had to basically
stop it 'cause the level
823
00:36:40,500 --> 00:36:42,090
of cybersecurity requirements
824
00:36:42,090 --> 00:36:44,910
that it would put into
these small suppliers,
825
00:36:44,910 --> 00:36:47,040
they just couldn't handle
it from a cost perspective.
826
00:36:47,040 --> 00:36:48,907
It was way too expensive
for them to really,
827
00:36:48,907 --> 00:36:50,490
you know, get to the level
828
00:36:50,490 --> 00:36:53,370
of cybersecurity that
supposed to be required.
829
00:36:53,370 --> 00:36:55,440
I wanna go back now to the idea
830
00:36:55,440 --> 00:36:57,480
of why we wanna create a SCOUT, right?
831
00:36:57,480 --> 00:37:01,890
The main target of SCOUT, of
this new tool that is just,
832
00:37:01,890 --> 00:37:04,290
the beginning of that right now,
833
00:37:04,290 --> 00:37:05,670
'cause we know it's not gonna be easy
834
00:37:05,670 --> 00:37:07,620
to build and we are working already
835
00:37:07,620 --> 00:37:12,120
with companies to basically
do it together, is that it's
836
00:37:12,120 --> 00:37:15,450
gonna impact mainly small and
medium-sized firms, right?
837
00:37:15,450 --> 00:37:17,940
A tool that allows you to
map the vulnerabilities
838
00:37:17,940 --> 00:37:21,960
to basically be able to
see where you should act
839
00:37:21,960 --> 00:37:23,070
and how you should act
840
00:37:23,070 --> 00:37:25,560
to improve your cybersecurity capabilities
841
00:37:25,560 --> 00:37:28,650
in the supply chain, how you
need to mitigate certain risks,
842
00:37:28,650 --> 00:37:31,800
how you need to prepare for
recovery in case of an attack.
843
00:37:31,800 --> 00:37:33,540
- There's been a lot of conversation
844
00:37:33,540 --> 00:37:35,610
and emphasis on agentic AI.
845
00:37:35,610 --> 00:37:38,670
As AI agents take on more work,
846
00:37:38,670 --> 00:37:40,200
what new risks do they create,
847
00:37:40,200 --> 00:37:42,716
and what guardrails are needed?
848
00:37:42,716 --> 00:37:46,800
- 2026 has been the
year of agentic AI, and-
849
00:37:46,800 --> 00:37:48,780
- And seeing in the news recently,
850
00:37:48,780 --> 00:37:50,087
some agents misbehaving,
851
00:37:50,087 --> 00:37:50,920
- Yeah, are going rogue.
- Yeah.
852
00:37:50,920 --> 00:37:51,930
- Going rogue.
- They're going rogue.
853
00:37:51,930 --> 00:37:53,460
AI is nondeterministic.
854
00:37:53,460 --> 00:37:54,750
You can tell it whatever you want,
855
00:37:54,750 --> 00:37:56,790
but sometimes it's not
gonna do exactly what's
856
00:37:56,790 --> 00:37:58,260
in the prompt 'cause, of course,
857
00:37:58,260 --> 00:37:59,850
you can only prompt so much.
858
00:37:59,850 --> 00:38:01,500
I mean, (laughs) it's basically impossible
859
00:38:01,500 --> 00:38:05,880
to give infinite instructions,
right, and be super precise,
860
00:38:05,880 --> 00:38:08,190
and that's actually one
of the risks, right?
861
00:38:08,190 --> 00:38:12,270
AI still does not distinguish
very well between instructions
862
00:38:12,270 --> 00:38:15,840
and anything that comes
after that, like any input,
863
00:38:15,840 --> 00:38:17,730
and that's actually one
of the biggest risks
864
00:38:17,730 --> 00:38:19,980
with any kind of AI tool and,
of course, agents, right?
865
00:38:19,980 --> 00:38:21,240
Like prompt injection.
866
00:38:21,240 --> 00:38:23,070
- Sounds like at the toddler phase, right?
867
00:38:23,070 --> 00:38:24,450
- Yeah, yeah, of course.
868
00:38:24,450 -->
00:38:25,740
Like something that, you know,
869
00:38:25,740 --> 00:38:28,920
gets embedded into your AI model,
870
00:38:28,920 --> 00:38:31,470
from a malicious actor, which I think most
871
00:38:31,470 --> 00:38:34,680
of the cases that we've
seen so far are not related
872
00:38:34,680 --> 00:38:37,620
to that, are related to
just the AI misbehaving.
873
00:38:37,620 --> 00:38:39,450
But, of course, we are
definitely gonna see
874
00:38:39,450 --> 00:38:44,040
more and more, cyberattacks
targeting AI agents in the near
875
00:38:44,040 --> 00:38:46,650
future 'cause they're
gonna be everywhere, right?
876
00:38:46,650 --> 00:38:48,270
So we are gonna be seeing probably
877
00:38:48,270 --> 00:38:51,150
AI fighting AI more and more.
878
00:38:51,150 --> 00:38:53,910
But the other day also in conversations
879
00:38:53,910 --> 00:38:57,570
with some of the companies
that are our partners,
880
00:38:57,570 --> 00:39:00,390
some of them are deploying
AI as a decoy, right?
881
00:39:00,390 --> 00:39:03,180
They're actually basically duplicating
882
00:39:03,180 --> 00:39:06,180
all their main kinda like
infrastructure or some
883
00:39:06,180 --> 00:39:10,890
of it to basically show as a
decoy to any kind of attacker.
884
00:39:10,890 --> 00:39:13,320
So instead of the attack going straight
885
00:39:13,320 --> 00:39:17,370
to your organization,
like IT infrastructure,
886
00:39:17,370 --> 00:39:21,420
it goes to this AI,
kinda like fake company.
887
00:39:21,420 --> 00:39:22,590
- Give them an easy target.
888
00:39:22,590 --> 00:39:23,529
- Yeah, giving them an easy target.
889
00:39:23,529 --> 00:39:25,080
- And it crumbles.
- And because of most
890
00:39:25,080 --> 00:39:27,420
of the attacks are now AI-enabled,
891
00:39:27,420 --> 00:39:31,140
they're also spending a lot of
money to perform the attack.
892
00:39:31,140 --> 00:39:33,390
So (laughs) basically
what companies can do
893
00:39:33,390 --> 00:39:35,220
is with these AI fakes,
894
00:39:35,220 --> 00:39:37,230
they actually make the attacker spend
895
00:39:37,230 --> 00:39:39,930
a bunch of time just going
through the system, getting lost,
896
00:39:39,930 --> 00:39:41,670
at the same time that they're
spending a lot of money,
897
00:39:41,670 --> 00:39:43,080
right, because they're paying
898
00:39:43,080 --> 00:39:46,070
for tokens to really
perform those attacks.
899
00:39:46,070 --> 00:39:48,480
So yeah, (laughs) it's
becoming a pretty interesting,
900
00:39:48,480 --> 00:39:52,830
you know, battle between
AIs, nowadays, yeah.
901
00:39:52,830 --> 00:39:54,330
- For companies listening,
902
00:39:54,330 --> 00:39:56,640
what would you all
suggest that they can do
903
00:39:56,640 --> 00:39:58,650
to improve their resilience?
904
00:39:58,650 --> 00:40:00,780
'Cause it can feel like a tall order
905
00:40:00,780 --> 00:40:02,700
to improve your cybersecurity.
906
00:40:02,700 --> 00:40:04,000
What should they do first?
907
00:40:05,010 --> 00:40:07,020
- Three things.
908
00:40:07,020 --> 00:40:09,360
The first one, people.
909
00:40:09,360 --> 00:40:11,310
The second one, people.
910
00:40:11,310 --> 00:40:13,740
And the third one, guess what?
911
00:40:13,740 --> 00:40:15,270
People. (laughs)
912
00:40:15,270 --> 00:40:16,893
So yeah, technology matters,
913
00:40:18,236 --> 00:40:20,730
but people remain one
of the greatest sources
914
00:40:20,730 --> 00:40:24,180
of both vulnerability and resilience.
915
00:40:24,180 --> 00:40:26,850
As a company, you can invest
916
00:40:26,850 --> 00:40:30,660
in sophisticated cybersecurity tools,
917
00:40:30,660 --> 00:40:35,660
but if employees are
not aware of the risks,
918
00:40:35,760 --> 00:40:40,760
if leadership does not know
how to respond during a crisis,
919
00:40:40,830 --> 00:40:45,180
or if different functions didn't
know how to work together,
920
00:40:45,180 --> 00:40:49,920
the organization can still be
extremely vulnerable, right?
921
00:40:49,920 --> 00:40:53,430
So I would say cybersecurity
preparedness is ultimately
922
00:40:53,430 --> 00:40:58,230
about people, culture, and
leadership, not only technology.
923
00:40:58,230 --> 00:41:01,530
- Yeah, when we started working also
924
00:41:01,530 --> 00:41:03,960
in supply chain risk
management and then, of course,
925
00:41:03,960 --> 00:41:08,960
we got into cyber, we used
to look at three main ways
926
00:41:09,120 --> 00:41:11,790
of preparing for any kind
of disruption, right?
927
00:41:11,790 --> 00:41:14,130
In risk management,
you look at prevention,
928
00:41:14,130 --> 00:41:15,810
mitigation, and recovery.
929
00:41:15,810 --> 00:41:16,643
Of course, in cyber,
930
00:41:16,643 --> 00:41:18,900
the detection part is
a big part of it too.
931
00:41:18,900 --> 00:41:21,720
But I think, for most companies,
932
00:41:21,720 --> 00:41:24,660
to mapping those dependencies
across, you know,
933
00:41:24,660 --> 00:41:28,560
the IT and the physical
operation side is gonna be huge.
934
00:41:28,560 --> 00:41:30,780
Prevention, still pretty much,
935
00:41:30,780 --> 00:41:32,970
I would say, 90% of the
work is gonna be done
936
00:41:32,970 --> 00:41:35,850
by the cybersecurity teams and IT teams.
937
00:41:35,850 --> 00:41:39,060
But having the supply chain
people and operations people
938
00:41:39,060 --> 00:41:41,160
at least being aware of
it and getting involved
939
00:41:41,160 --> 00:41:43,710
in the conversations is
gonna be super important.
940
00:41:43,710 --> 00:41:46,170
And the same, that goes
into the mitigation
941
00:41:46,170 --> 00:41:48,270
and recovery part 'cause that's where most
942
00:41:48,270 --> 00:41:50,850
of the supply chain people,
operations people need
943
00:41:50,850 --> 00:41:53,940
to play a much bigger role
than they have in the past.
944
00:41:53,940 --> 00:41:56,310
So for instance, in cyber,
945
00:41:56,310 --> 00:41:59,160
it's very typical to do
tabletop exercises, right?
946
00:41:59,160 --> 00:42:03,356
To basically model how
an attack, you know,
947
00:42:03,356 --> 00:42:06,360
is happening, and once it
has hit the organization,
948
00:42:06,360 --> 00:42:07,830
how we are gonna respond to it.
949
00:42:07,830 --> 00:42:11,670
Usually, they're pretty
strategic, so most of the times,
950
00:42:11,670 --> 00:42:14,220
they remain at the high level
of the organization, right?
951
00:42:14,220 --> 00:42:16,800
C-suite, VP level, and so on.
952
00:42:16,800 --> 00:42:17,670
But it's critical now
953
00:42:17,670 --> 00:42:19,710
to start bringing
operations people into them.
954
00:42:19,710 --> 00:42:22,470
It's not just a data
breach, reputational issue.
955
00:42:22,470 --> 00:42:24,990
No, (laughs) this attack just
956
00:42:24,990 --> 00:42:27,690
stopped your manufacturing
plant from producing.
957
00:42:27,690 --> 00:42:30,330
So really having like the people
958
00:42:30,330 --> 00:42:34,200
from the field in those
exercises and going from, again,
959
00:42:34,200 --> 00:42:37,050
the prevention, mitigation,
and to the recovery part
960
00:42:37,050 --> 00:42:39,630
of it altogether is gonna be critical.
961
00:42:39,630 --> 00:42:43,500
It's super hard 'cause also when you think
962
00:42:43,500 --> 00:42:46,230
about the business
impact of a cyberattack,
963
00:42:46,230 --> 00:42:50,160
that might change like from
one day to another, right?
964
00:42:50,160 --> 00:42:51,720
Most of the attacks (laughs)
965
00:42:51,720 --> 00:42:54,780
don't happen during valleys
of demand, for instance.
966
00:42:54,780 --> 00:42:55,950
They happen during peak season,
967
00:42:55,950 --> 00:42:57,780
like, you know, Thanksgiving,
968
00:42:57,780 --> 00:43:01,860
Christmas is coming, all the
retailers are getting hit.
969
00:43:01,860 --> 00:43:03,240
'Cause, of course, for the attackers,
970
00:43:03,240 --> 00:43:06,150
I mean, they know like if
they're asking for a ransom,
971
00:43:06,150 --> 00:43:08,280
oh, they're gonna pay 'cause
they have all their, you know,
972
00:43:08,280 --> 00:43:11,793
Thanksgiving shipments ready
to go, and they need to, right?
973
00:43:12,960 --> 00:43:14,373
How did you actually model the risk
974
00:43:14,373 --> 00:43:16,620
in a different way when
it comes to that, right?
975
00:43:16,620 --> 00:43:19,650
Probably you're not gonna
have the same kind of recovery
976
00:43:19,650 --> 00:43:22,140
plan for a normal day than, you know,
977
00:43:22,140 --> 00:43:24,360
the week of Thanksgiving or
two weeks before Thanksgiving
978
00:43:24,360 --> 00:43:26,310
when you're basically
shipping out everything,
979
00:43:26,310 --> 00:43:28,410
Black Friday and so on, right?
980
00:43:28,410 --> 00:43:31,350
So those kind of like dynamic
business continuity plans,
981
00:43:31,350 --> 00:43:32,640
for instance, is something that, you know,
982
00:43:32,640 --> 00:43:34,440
the operations and supply chain teams need
983
00:43:34,440 --> 00:43:37,320
to be involved in the
conversations early on.
984
00:43:37,320 --> 00:43:38,820
- And I will add to that that one
985
00:43:38,820 --> 00:43:41,670
practical recommendation,
you know, for companies
986
00:43:41,670 --> 00:43:46,670
would be identify the operations
you cannot afford to lose.
987
00:43:48,960 --> 00:43:53,960
Understand what they depend
on, and prepare your people
988
00:43:54,420 --> 00:43:56,763
to operate when those dependencies fail.
989
00:43:57,900 --> 00:43:59,550
That is critical.
990
00:43:59,550 --> 00:44:02,280
- You all mentioned
it's hard to look ahead,
991
00:44:02,280 --> 00:44:04,110
because there's so much
changing in cybersecurity,
992
00:44:04,110 --> 00:44:08,280
but if you look at the next
five to 10 years, how do you all
993
00:44:08,280 --> 00:44:11,820
think supply chain cybersecurity
may evolve in that time?
994
00:44:11,820 --> 00:44:14,760
- Oh, well, we are
developing a tool for that.
995
00:44:14,760 --> 00:44:18,150
It's called MIT RADAR, and
it's essentially a tool
996
00:44:18,150 --> 00:44:22,830
that is helping us to identify
emerging cybersecurity risks
997
00:44:22,830 --> 00:44:26,850
before they become a
major supply chain threat.
998
00:44:26,850 --> 00:44:30,120
The idea is to respect
the different domains,
999
00:44:30,120 --> 00:44:34,560
you know, like physical operations
1000
00:44:34,560 --> 00:44:37,800
and IT and others, even like, you know,
1001
00:44:37,800 --> 00:44:41,280
signals like quantum
computing, how these, you know,
1002
00:44:41,280 --> 00:44:44,910
technologies are gonna play a
critical role in the future,
1003
00:44:44,910 --> 00:44:48,660
how companies should
prepare their systems,
1004
00:44:48,660 --> 00:44:50,940
organization to those trends.
1005
00:44:50,940 --> 00:44:52,440
It's gonna happen, right?
1006
00:44:52,440 --> 00:44:56,740
So again, like the idea is
to combine the learnings
1007
00:44:57,772 --> 00:45:01,050
from TRACE and start like, you know,
1008
00:45:01,050 --> 00:45:02,670
trying to anticipate the future
1009
00:45:02,670 --> 00:45:07,670
with RADAR to shape the
present, the decisions today.
1010
00:45:07,752 --> 00:45:09,123
- And you all have spoken
a lot about, you know,
1011
00:45:09,123 --> 00:45:11,250
what the lab is working
on in terms of tools,
1012
00:45:11,250 --> 00:45:14,490
but in terms of everything
else, what is the lab doing now?
1013
00:45:14,490 --> 00:45:15,360
What are the priorities,
1014
00:45:15,360 --> 00:45:17,040
maybe things that you haven't covered yet?
1015
00:45:17,040 --> 00:45:17,873
- Hmm.
1016
00:45:17,873 --> 00:45:20,520
- Yeah, to sum up the,
maybe the tools part,
1017
00:45:20,520 --> 00:45:22,890
'cause it's something
that we're very proud of.
1018
00:45:22,890 --> 00:45:24,360
TRACE, looking at the past
1019
00:45:24,360 --> 00:45:26,970
or mapping all the cyberattacks worldwide
1020
00:45:26,970 --> 00:45:29,310
that are having an impact
on physical operations.
1021
00:45:29,310 --> 00:45:30,720
RADAR, looking at the future,
1022
00:45:30,720 --> 00:45:32,490
so looking at all the
emerging technologies
1023
00:45:32,490
--> 00:45:36,270
that are coming to supply chain
and basically crossing all
1024
00:45:36,270 --> 00:45:38,850
that information with
current vulnerabilities
1025
00:45:38,850 --> 00:45:41,857
that we're seeing in the space
already to tell a company,
1026
00:45:41,857 --> 00:45:44,190
"Hey, if you're gonna be having humanoids
1027
00:45:44,190 --> 00:45:46,800
in your warehouse tomorrow,
how are you gonna have to think
1028
00:45:46,800 --> 00:45:48,480
about cyber from that perspective?"
1029
00:45:48,480 --> 00:45:49,830
And of course this is gonna be huge
1030
00:45:49,830 --> 00:45:52,200
'cause we are gonna have more
and more autonomous systems,
1031
00:45:52,200 --> 00:45:54,900
not only agentic AI on the digital world,
1032
00:45:54,900 --> 00:45:57,540
but also, you know, autonomous
trucks coming to the roads,
1033
00:45:57,540 --> 00:46:00,570
humanoids coming to manufacturing
plants, warehouses, so.
1034
00:46:00,570 --> 00:46:03,150
And then SCOUT, that
basically puts all together,
1035
00:46:03,150 --> 00:46:04,530
kind of like takes information
1036
00:46:04,530 --> 00:46:08,550
from those two and becomes a
vulnerability assessment tool,
1037
00:46:08,550 --> 00:46:12,000
dynamic tool, for a company to be able
1038
00:46:12,000 --> 00:46:15,360
to actually just improve their
cybersecurity capabilities
1039
00:46:15,360 --> 00:46:16,710
when it comes to supply chain.
1040
00:46:16,710 --> 00:46:17,760
Those are definitely, I would say,
1041
00:46:17,760 --> 00:46:20,580
the main pillars of what
we are working on already.
1042
00:46:20,580 --> 00:46:22,620
And of course, then we
are doing interviews.
1043
00:46:22,620 --> 00:46:24,690
We call that the impact study
1044
00:46:24,690 --> 00:46:27,000
with a lot of the companies
that have suffered some
1045
00:46:27,000 --> 00:46:29,220
of these large cyberattacks worldwide
1046
00:46:29,220 --> 00:46:31,500
and events like the roundtables.
1047
00:46:31,500 --> 00:46:34,440
So the idea is to bring experts together
1048
00:46:34,440 --> 00:46:36,000
and to learn from all of them, right?
1049
00:46:36,000 --> 00:46:37,530
To see what's going on in this space.
1050
00:46:37,530 --> 00:46:40,380
A lot of these cyberattacks
are not being disclosed openly,
1051
00:46:40,380 --> 00:46:43,860
and I think every time I
actually teach a class,
1052
00:46:43,860 --> 00:46:45,900
I always ask, right,
like, "Who has, you know,
1053
00:46:45,900 --> 00:46:47,730
been impacted by a cyberattack?"
1054
00:46:47,730 --> 00:46:49,710
I think four or five years ago
1055
00:46:49,710 --> 00:46:52,410
when I started having kind
of like these ideas in mind,
1056
00:46:52,410 --> 00:46:54,480
like there were basically a couple
1057
00:46:54,480 --> 00:46:56,250
people raising their hands. (laughs)
1058
00:46:56,250 --> 00:46:57,960
In the last class, I taught,
1059
00:46:57,960 --> 00:47:01,440
like 70% of the class were
already like, "Eh, me."
1060
00:47:01,440 --> 00:47:03,420
So that's definitely,
you know, a
big change.
1061
00:47:03,420 --> 00:47:06,615
So the third pillar at MIT,
and specifically at CTL,
1062
00:47:06,615 --> 00:47:07,980
right, is educational piece.
1063
00:47:07,980 --> 00:47:12,210
So I've been teaching already
for a while, on cyber,
1064
00:47:12,210 --> 00:47:16,080
but we are already thinking
ahead, and considering a course
1065
00:47:16,080 --> 00:47:18,840
on supply chain cybersecurity,
a specific course.
1066
00:47:18,840 --> 00:47:22,920
We don't have dates yet, but
we think 2027 is pretty likely
1067
00:47:22,920 --> 00:47:26,750
for us to have that ready, and
we're super excited for it.
1068
00:47:26,750 --> 00:47:27,720
- So those are all the questions
1069
00:47:27,720 --> 00:47:30,540
I have, but anything else
that you all wanna add?
1070
00:47:30,540 --> 00:47:32,670
So great to have you both on the podcast.
1071
00:47:32,670 --> 00:47:33,503
- Thank you.
1072
00:47:33,503 --> 00:47:35,490
Sure, for our listeners,
1073
00:47:35,490 --> 00:47:39,810
we are here to help you, so
please feel free to reach out,
1074
00:47:39,810 --> 00:47:43,260
you know, to explore
our website, our tools.
1075
00:47:43,260 --> 00:47:46,860
We work every day to help
you, to help companies.
1076
00:47:46,860 --> 00:47:48,360
It may sound a little bit romantic,
1077
00:47:48,360 --> 00:47:49,860
but at the end of the day,
1078
00:47:49,860 --> 00:47:51,690
we are doing this for a better world.
1079
00:47:51,690 --> 00:47:53,730
- From my side, big thank you too.
1080
00:47:53,730 --> 00:47:54,720
It was a pleasure.
1081
00:47:54,720 --> 00:47:55,560
Our goal, again,
1082
00:47:55,560 --> 00:47:58,200
is to protect physical supply
chains from cybercrime.
1083
00:47:58,200 --> 00:48:00,030
That's why we created this new lab.
1084
00:48:00,030 --> 00:48:03,480
Any company that wants to
talk to us, we wanna learn
1085
00:48:03,480 --> 00:48:05,190
about what they're going
through, and hopefully
1086
00:48:05,190 --> 00:48:07,530
then we can bring that
back to the industry
1087
00:48:07,530 --> 00:48:10,350
and make cybersecurity for supply chains,
1088
00:48:10,350 --> 00:48:13,560
you know, just better
for everybody out there.
1089
00:48:13,560 --> 00:48:14,940
- That wraps up this episode
1090
00:48:14,940 --> 00:48:16,590
of "Supply Chain Frontiers."
1091
00:48:16,590 --> 00:48:18,150
A big thank you to Miguel and Camilo
1092
00:48:18,150 --> 00:48:20,070
for sharing their expertise and insights
1093
00:48:20,070 --> 00:48:23,100
into the current state of
supply chain cybersecurity.
1094
00:48:23,100 --> 00:48:24,360
To learn more about the lab,
1095
00:48:24,360 --> 00:48:28,897
visit it on the research
page of ctl.mit.edu.
1096
00:48:28,897 --> 00:48:30,690
"Supply Chain Frontiers" is recorded
1097
00:48:30,690 --> 00:48:33,840
on the MIT campus in
Cambridge, Massachusetts.
1098
00:48:33,840 --> 00:48:36,780
Our sound editor is at
David Benjamin Sound,
1099
00:48:36,780 --> 00:48:39,030
and our audio engineer
today is Kurt Schneider
1100
00:48:39,030 --> 00:48:41,580
of MIT Audiovisual Services.
1101
00:48:41,580 --> 00:48:44,070
Our producer is myself, Mackenzie Berry.
1102
00:48:44,070 --> 00:48:45,750
Be sure to check out previous episodes
1103
00:48:45,750 --> 00:48:50,750
of "Supply Chain Frontiers"
at ctl.mit.edu/podcast
1104
00:48:51,120 --> 00:48:54,180
or search for us on your
preferred podcast platform.
1105
00:48:54,180 --> 00:48:55,140
I'm Mackenzie Berry.
1106
00:48:55,140 --> 00:48:57,240
Thanks for listening, and
we'll catch you next time
1107
00:48:57,240 --> 00:48:58,612
on "Supply Chain Frontiers."
1108
00:48:58,612 --> 00:49:01,195
(upbeat music)