Avi Rubin: All your devices can be hacked

44,157 views ใƒป 2015-07-15

TED


ืื ื ืœื—ืฅ ืคืขืžื™ื™ื ืขืœ ื”ื›ืชื•ื‘ื™ื•ืช ื‘ืื ื’ืœื™ืช ืœืžื˜ื” ื›ื“ื™ ืœื”ืคืขื™ืœ ืืช ื”ืกืจื˜ื•ืŸ.

00:00
Translator: Joseph Geni Reviewer: Morton Bast
0
0
7000
ืžืชืจื’ื: Ido Dekkers ืžื‘ืงืจ: Sonia Barchilon
00:12
I'm a computer science professor,
1
12588
3031
ืื ื™ ืžืจืฆื” ืœืžื“ืขื™ ื”ืžื—ืฉื‘,
00:15
and my area of expertise is
2
15619
2313
ื•ืชื—ื•ื ื”ืžื•ืžื—ื™ื•ืช ืฉืœื™
00:17
computer and information security.
3
17932
2199
ื”ื•ื ืžื—ืฉื‘ื™ื ื•ืื‘ื˜ื—ืช ืžื™ื“ืข.
00:20
When I was in graduate school,
4
20131
2320
ื›ืฉื”ื™ื™ืชื™ ื‘ืœื™ืžื•ื“ื™ ื”ืชื•ืืจ ื”ืฉื ื™,
00:22
I had the opportunity to overhear my grandmother
5
22451
2601
ื”ื–ื“ืžืŸ ืœื™ ืœืฉืžื•ืข ืืช ืกื‘ืชื ืฉืœื™
00:25
describing to one of her fellow senior citizens
6
25052
4134
ืžืชืืจืช ื‘ืื•ื–ื ื™ ืื—ื“ ืžื™ื“ื™ื“ื™ื”, ืื“ื ืžื‘ื•ื’ืจ,
00:29
what I did for a living.
7
29186
2369
ืžืžื” ืื ื™ ืžืชืคืจื ืก.
00:31
Apparently, I was in charge of making sure that
8
31555
3562
ื”ืกืชื‘ืจ ืœื™ ืฉืื ื™ ืื—ืจืื™ ืœื”ื‘ื˜ื™ื—
00:35
no one stole the computers from the university. (Laughter)
9
35117
3900
ืฉืื™ืฉ ืœื ื™ื’ื ื•ื‘ ืžื—ืฉื‘ื™ื ืžื”ืื•ื ื™ื‘ืจืกื™ื˜ื”. [ืฆื—ื•ืง]
00:39
And, you know, that's a perfectly reasonable thing
10
39017
2744
ื•ืžื‘ื—ื™ื ืชื” ื–ื” ื”ื’ื™ื•ื ื™ ืžืื•ื“,
00:41
for her to think, because I told her I was working
11
41761
1920
ื›ื™ ืืžืจืชื™ ืœื” ืฉืื ื™ ืขื•ื‘ื“
00:43
in computer security,
12
43681
1507
ื‘ืื‘ื˜ื—ืช ืžื—ืฉื‘ื™ื,
00:45
and it was interesting to get her perspective.
13
45188
3597
ื•ื”ื™ื” ืžืขื ื™ื™ืŸ ืœืฉืžื•ืข ืืช ื ืงื•ื“ืช ื”ื”ืฉืงืคื” ืฉืœื”.
00:48
But that's not the most ridiculous thing I've ever heard
14
48785
2617
ืื‘ืœ ืœื ื–ื” ื”ื“ื‘ืจ ื”ื›ื™ ืžื’ื•ื—ืš ืฉืฉืžืขืชื™
00:51
anyone say about my work.
15
51402
2017
ืฉืื•ืžืจื™ื ืขืœ ืขื‘ื•ื“ืชื™.
00:53
The most ridiculous thing I ever heard is,
16
53419
2284
ื”ื“ื‘ืจ ื”ื›ื™ ืžื’ื•ื—ืš ืฉืฉืžืขืชื™,
00:55
I was at a dinner party, and a woman heard
17
55703
3134
ื”ื™ื” ื‘ืืจื•ื—ืช ืขืจื‘, ื•ืื™ืฉื” ืื—ืช ืฉืฉืžืขื”
00:58
that I work in computer security,
18
58837
1783
ืฉืื ื™ ืขื•ื‘ื“ ื‘ืื‘ื˜ื—ืช ืžื—ืฉื‘ื™ื,
01:00
and she asked me if -- she said her computer had been
19
60620
3517
ืฉืืœื” ืื•ืชื™ ืื - ื”ื™ื ืืžืจื” ืฉื”ืžื—ืฉื‘ ืฉืœื”
01:04
infected by a virus, and she was very concerned that she
20
64137
3436
ื—ื˜ืฃ ื•ื™ืจื•ืก, ื•ื”ื™ื ื”ื™ืชื” ืžืื•ื“ ืžื•ื“ืื’ืช
01:07
might get sick from it, that she could get this virus. (Laughter)
21
67573
3951
ืฉื’ื ื”ื™ื ืชื—ืœื”, ื•ืชื™ื“ื‘ืง ื‘ื•ื•ื™ืจื•ืก ื”ื–ื”. [ืฆื—ื•ืง]
01:11
And I'm not a doctor, but I reassured her
22
71524
2943
ืื ื™ ืืžื ื ืœื ืจื•ืคื, ืื‘ืœ ื”ื‘ื˜ื—ืชื™ ืœื”
01:14
that it was very, very unlikely that this would happen,
23
74467
3144
ืฉืžืื•ื“ ืžืื•ื“ ืœื ืกื‘ื™ืจ ืฉื–ื” ื™ืงืจื”,
01:17
but if she felt more comfortable, she could be free to use
24
77611
2801
ืื‘ืœ ืื ื–ื” ื™ืจื’ื™ืข ืื•ืชื”, ื”ื™ื ื™ื›ื•ืœื” ืœื”ืฉืชืžืฉ
01:20
latex gloves when she was on the computer,
25
80412
1848
ื‘ื›ืคืคื•ืช ื’ื•ืžื™ ื›ืฉื”ื™ื ืขื•ื‘ื“ืช ื‘ืžื—ืฉื‘,
01:22
and there would be no harm whatsoever in that.
26
82260
3392
ื•ืœื ื™ื™ื’ืจื ืœื” ืฉื•ื ื ื–ืง.
01:25
I'm going to get back to this notion of being able to get
27
85652
2507
ืขื•ื“ ืื—ื–ื•ืจ ืœืจืขื™ื•ืŸ ื”ื–ื”, ืฉืœ ื”ืืคืฉืจื•ืช ืœื—ื˜ื•ืฃ ื•ื™ืจื•ืก
01:28
a virus from your computer, in a serious way.
28
88159
3508
ืžื”ืžื—ืฉื‘, ื‘ื ื™ืžื” ืจืฆื™ื ื™ืช ื™ื•ืชืจ.
01:31
What I'm going to talk to you about today
29
91667
1640
ื”ื™ื•ื ืื“ื‘ืจ ืื™ืชื›ื
01:33
are some hacks, some real world cyberattacks that people
30
93307
4846
ืขืœ ื›ืžื” ืคืจื™ืฆื•ืช-ืžื—ืฉื‘, ื›ืžื” ืžืชืงืคื•ืช ืื™ื ื˜ืจื ื˜ ืืžื™ืชื™ื•ืช
01:38
in my community, the academic research community,
31
98153
2554
ืฉืื ืฉื™ื ืžื”ืงื”ื™ืœื” ืฉืœื™, ืงื”ื™ืœืช ื”ืžื—ืงืจ ื”ืืงื“ืžื™,
01:40
have performed, which I don't think
32
100707
2794
ื‘ื™ืฆืขื•, ื•ืœื ื ืจืื” ืœื™
01:43
most people know about,
33
103501
1208
ืฉืจื•ื‘ ื”ืื ืฉื™ื ื™ื•ื“ืขื™ื ืขืœื™ื”ืŸ,
01:44
and I think they're very interesting and scary,
34
104709
3028
ื•ืœื“ืขืชื™ ื”ืŸ ืžืื•ื“ ืžืขื ื™ื™ื ื•ืช ื•ืžืคื—ื™ื“ื•ืช,
01:47
and this talk is kind of a greatest hits
35
107737
2441
ื•ื”ื”ืจืฆืื” ื”ื–ื• ื”ื™ื ืžืฉื”ื• ื›ืžื• ืžื™ื˜ื‘ ื”ืœื”ื™ื˜ื™ื
01:50
of the academic security community's hacks.
36
110178
2991
ืฉืœ ืคืจื™ืฆื•ืช ื”ืžื—ืฉื‘ ืฉืœ ืงื”ื™ืœืช ืื‘ื˜ื—ืช ื”ืžื™ื“ืข ื”ืืงื“ืžื™ืช.
01:53
None of the work is my work. It's all work
37
113169
1987
ืื™ืŸ ื›ืืŸ ืขื‘ื•ื“ื•ืช ืฉืœื™. ืืœื ื›ื•ืœื ืขื‘ื•ื“ื•ืช
01:55
that my colleagues have done, and I actually asked them
38
115156
2174
ืฉืขืžื™ืชื™ื™ ื‘ื™ืฆืขื•, ื•ืื ื™ ื‘ื™ืงืฉืชื™ ืžื”ื ืฉืงื•ืคื™ื•ืช
01:57
for their slides and incorporated them into this talk.
39
117330
2557
ื•ืฉื™ืœื‘ืชื™ ืื•ืชืŸ ื‘ื”ืจืฆืื” ื”ื–ื•.
01:59
So the first one I'm going to talk about
40
119887
1742
ื”ืขื‘ื•ื“ื” ื”ืจืืฉื•ื ื” ืฉืขืœื™ื” ืื“ื‘ืจ
02:01
are implanted medical devices.
41
121629
2674
ืขื•ืกืงืช ื‘ื”ืฉืชืœืช ืื‘ื™ื–ืจื™ื ืจืคื•ืื™ื™ื.
02:04
Now medical devices have come a long way technologically.
42
124303
3040
ืื‘ื™ื–ืจื™ื ืจืคื•ืื™ื™ื ืขืฉื• ื›ื‘ืจืช ื“ืจืš ืžื‘ื—ื™ื ื” ื˜ื›ื ื•ืœื•ื’ื™ืช.
02:07
You can see in 1926 the first pacemaker was invented.
43
127343
3856
ืืคืฉืจ ืœืจืื•ืช ืฉื‘-1926 ื”ื•ืžืฆื ืงื•ืฆื‘ ื”ืœื‘ ื”ืจืืฉื•ืŸ.
02:11
1960, the first internal pacemaker was implanted,
44
131199
3552
1960 - ื”ื•ืฉืชืœ ืงื•ืฆื‘ ื”ืœื‘ ื”ืคื ื™ื-ื’ื•ืคื™ ื”ืจืืฉื•ืŸ,
02:14
hopefully a little smaller than that one that you see there,
45
134751
2552
ื ืงื•ื•ื” ืฉื”ื•ื ื”ื™ื” ืงื˜ืŸ ื™ื•ืชืจ ืžื–ื” ืฉืจื•ืื™ื ื›ืืŸ,
02:17
and the technology has continued to move forward.
46
137303
2968
ื•ื”ื˜ื›ื ื•ืœื•ื’ื™ื” ื”ืžืฉื™ื›ื” ืœื”ืชืงื“ื.
02:20
In 2006, we hit an important milestone from the perspective
47
140271
4633
ื‘-2006 ื”ื’ืขื ื• ืœืื‘ืŸ-ื“ืจืš ื—ืฉื•ื‘ื”
02:24
of computer security.
48
144904
3167
ืžื‘ื—ื™ื ืช ืื‘ื˜ื—ืช ื”ืžื—ืฉื•ื‘.
02:28
And why do I say that?
49
148071
1341
ื•ืžื“ื•ืข ืื ื™ ืื•ืžืจ ื–ืืช?
02:29
Because that's when implanted devices inside of people
50
149412
2890
ื›ื™ ืื– ื”ืื‘ื™ื–ืจื™ื ืฉื”ื•ืฉืชืœื• ื‘ื‘ื ื™-ืื“ื
02:32
started to have networking capabilities.
51
152302
2745
ื”ื—ืœื• ืœื”ื™ื•ืช ื‘ืขืœื™ ื™ื›ื•ืœื•ืช ืฉืœ ืจืฉืช.
02:35
One thing that brings us close to home is we look
52
155047
1880
ืื—ื“ ื”ื“ื‘ืจื™ื ืฉืžืงืจื‘ื™ื ืื•ืชื ื• ืœื ื•ืฉื ืฉืœื ื•
02:36
at Dick Cheney's device, he had a device that
53
156927
2705
ื”ื•ื ื”ื”ืชืงืŸ ืฉืœ ื“ื™ืง ืฆ'ื™ื™ื ื™. ื”ื™ื” ืœื• ื”ืชืงืŸ
02:39
pumped blood from an aorta to another part of the heart,
54
159632
3869
ืฉืฉืื‘ ื“ื ืžืื‘ื™-ื”ืขื•ืจืงื™ื ืœื—ืœืง ืื—ืจ ื‘ืœื‘,
02:43
and as you can see at the bottom there,
55
163501
1183
ื•ื›ืคื™ ืฉืืชื ืจื•ืื™ื ื›ืืŸ ืœืžื˜ื”,
02:44
it was controlled by a computer controller,
56
164684
3009
ื”ื•ื ื ืฉืœื˜ ืขืœ-ื™ื“ื™ ื‘ืงืจ ืžืžื•ื—ืฉื‘,
02:47
and if you ever thought that software liability
57
167693
2517
ื•ืžื™ ืฉื—ื•ืฉื‘ ืฉืืžื™ื ื•ืช ืฉืœ ืชื•ื›ื ื”
02:50
was very important, get one of these inside of you.
58
170210
3589
ื”ื™ื ื—ืฉื•ื‘ื” ืžืื•ื“, ืฉื™ื›ื ื™ืก ื“ื‘ืจ ื›ื–ื” ืœื’ื•ืฃ ืฉืœื•.
02:53
Now what a research team did was they got their hands
59
173799
3695
ืื– ืฆื•ื•ืช ืžื—ืงืจ ื”ื ื™ื— ืืช ื™ื“ื™ื•
02:57
on what's called an ICD.
60
177494
1420
ืขืœ ืžื” ืฉืžื›ื•ื ื” ืฉืชืœ ื“ืคื™ื‘ืจื™ืœื˜ื•ืจ.
02:58
This is a defibrillator, and this is a device
61
178914
2070
ื”ื“ืคื™ื‘ืจื™ืœื˜ื•ืจ ื”ื•ืคืš ืืช ืคื™ืจืคื•ืจื™ ืœื‘, ื•ื–ื”ื• ื”ืชืงืŸ
03:00
that goes into a person to control their heart rhythm,
62
180984
4336
ืฉืžื›ื ื™ืกื™ื ืœื’ื•ืฃ ื›ื“ื™ ืœืฉืœื•ื˜ ื‘ืงืฆื‘ ื”ืœื‘,
03:05
and these have saved many lives.
63
185320
2338
ื•ื”ืžื›ืฉื™ืจื™ื ื”ืืœื” ื”ืฆื™ืœื• ื”ืžื•ืŸ ืื ืฉื™ื.
03:07
Well, in order to not have to open up the person
64
187658
2472
ืื‘ืœ ื›ื“ื™ ืœืžื ื•ืข ืืช ื”ืฆื•ืจืš ืœืคืชื•ื— ืืช ื”ืžื˜ื•ืคืœ
03:10
every time you want to reprogram their device
65
190130
2194
ื‘ื›ืœ ืคืขื ืฉืจื•ืฆื™ื ืœืชื›ื ืช ืžื—ื“ืฉ ืืช ื”ื”ืชืงืŸ
03:12
or do some diagnostics on it, they made the thing be able
66
192324
2455
ืื• ืœื‘ื“ื•ืง ืืช ื”ืžื›ืฉื™ืจ, ื‘ื ื• ืื•ืชื• ื›ืš ืฉื™ื•ื›ืœ
03:14
to communicate wirelessly, and what this research team did
67
194779
3102
ืœืขื‘ื•ื“ ื‘ืชืงืฉื•ืจืช ืืœื—ื•ื˜ื™ืช, ื•ืฆื•ื•ืช ื”ืžื—ืงืจ
03:17
is they reverse engineered the wireless protocol,
68
197881
2610
ืคื™ืจืง ืœื’ื•ืจืžื™ื ื•ื‘ืขื–ืจืช ื”ื ื“ืกื” ื”ืคื•ื›ื” ืฉืœ ื”ืคืจื•ื˜ื•ืงื•ืœ ื”ืืœื—ื•ื˜ื™,
03:20
and they built the device you see pictured here,
69
200491
1872
ื”ื ื‘ื ื• ืืช ื”ื”ืชืงืŸ ืฉืžืฆื•ืœื ื›ืืŸ,
03:22
with a little antenna, that could talk the protocol
70
202363
2760
ืขื ื”ืื ื˜ื ื” ื”ืงื˜ื ื”, ืฉื™ื›ื•ืœื” ืœืงืฉืจ ื‘ื™ืŸ ื”ืคืจื•ื˜ื•ืงื•ืœ
03:25
to the device, and thus control it.
71
205123
4475
ืœื”ืชืงืŸ, ื•ื›ืš ืœืฉืœื•ื˜ ื‘ื•.
03:29
In order to make their experience real -- they were unable
72
209598
2689
ื›ื“ื™ ืœื‘ื“ื•ืง ืืช ื”ื ื™ืกื•ื™ ื‘ืžืฆื™ืื•ืช - ื”ื ืœื ื”ืฆืœื™ื—ื•
03:32
to find any volunteers, and so they went
73
212287
2472
ืœืžืฆื•ื ืžืชื ื“ื‘ื™ื, ื•ืœื›ืŸ ื”ื ื”ืœื›ื•
03:34
and they got some ground beef and some bacon
74
214759
2144
ื•ืงื ื• ืงืฆืช ื‘ืฉืจ ื˜ื—ื•ืŸ ื•ื‘ื™ื™ืงื•ืŸ
03:36
and they wrapped it all up to about the size
75
216903
1788
ื•ืืจื–ื• ืื•ืชื ื‘ื™ื—ื“ ื‘ืขืจืš ืœื’ื•ื“ืœ
03:38
of a human being's area where the device would go,
76
218691
2798
ืฉืœ ื”ืื–ื•ืจ ืฉื‘ื• ื™ื•ื›ื ืก ื”ื”ืชืงืŸ ื‘ื’ื•ืฃ ื”ืื“ื,
03:41
and they stuck the device inside it
77
221489
1454
ื•ืชืงืขื• ื‘ืชื•ื›ื• ืืช ื”ื”ืชืงืŸ
03:42
to perform their experiment somewhat realistically.
78
222943
3132
ื›ื“ื™ ืœื‘ืฆืข ืืช ื”ื ื™ืกื•ื™ ื‘ืฆื•ืจื” ืžืฆื™ืื•ืชื™ืช-ืžืฉื”ื•.
03:46
They launched many, many successful attacks.
79
226075
3020
ื”ื ื‘ื™ืฆืขื• ื”ืžื•ืŸ ื”ืžื•ืŸ ืžืชืงืคื•ืช ืžื•ืฆืœื—ื•ืช.
03:49
One that I'll highlight here is changing the patient's name.
80
229095
3056
ื•ืื—ืช ืžื”ืŸ, ืฉื”ื“ื’ืฉืชื™ ื›ืืŸ, ืžืฉื ื” ืืช ืฉื ื”ื—ื•ืœื”.
03:52
I don't know why you would want to do that,
81
232151
993
ืื™ืŸ ืœื™ ืžื•ืฉื’ ืœืžื” ื–ื” ื˜ื•ื‘,
03:53
but I sure wouldn't want that done to me.
82
233144
2104
ืื‘ืœ ื‘ื˜ื— ืฉืœื ื”ื™ื™ืชื™ ืจื•ืฆื” ืฉื™ืขืฉื• ืœื™ ืืช ื–ื”.
03:55
And they were able to change therapies,
83
235248
2331
ื•ื”ื ื”ืฆืœื™ื—ื• ืœืฉื ื•ืช ืืช ื”ื˜ื™ืคื•ืœ,
03:57
including disabling the device -- and this is with a real,
84
237579
2495
ื›ื•ืœืœ ื”ืฉื‘ืชืช ื”ื”ืชืงืŸ - ื•ืžื“ื•ื‘ืจ ื›ืืŸ ื‘ื”ืชืงืŸ ืืžื™ืชื™,
04:00
commercial, off-the-shelf device --
85
240074
1896
ืžืกื—ืจื™, ื”ืชืงืŸ ืฉื ื™ืชืŸ ื‘ืคื™ืงื•ื— ื‘ืœื‘ื“ -
04:01
simply by performing reverse engineering and sending
86
241970
2046
ืคืฉื•ื˜ ืขืœ ื™ื“ื™ ื”ื ื“ืกื” ื”ืคื•ื›ื” ืฉืœ ื”ืžืขืจื›ืช
04:04
wireless signals to it.
87
244016
2989
ื•ืฉื™ื’ื•ืจ ืื•ืชื•ืช ืืœื—ื•ื˜ื™ื™ื.
04:07
There was a piece on NPR that some of these ICDs
88
247005
3580
ื‘ืจื“ื™ื• ื”ืฆื™ื‘ื•ืจื™ ืฉื•ื“ืจื” ื›ืชื‘ื” ื•ื”ืกืชื‘ืจ ืฉื‘ื›ืžื” ืžืฉืชืœื™ ื”ื™ืคื•ืš ื”ืคืจืคื•ืจื™ื
04:10
could actually have their performance disrupted
89
250585
2422
ื”ื‘ื™ืฆื•ืขื™ื ืขืœื•ืœื™ื ืœื”ืฉืชื‘ืฉ
04:13
simply by holding a pair of headphones onto them.
90
253007
3651
ืืคื™ืœื• ืื ืžืงืจื‘ื™ื ืืœื™ื”ื ื–ื•ื’ ืื•ื–ื ื™ื•ืช ืจื’ื™ืœื•ืช.
04:16
Now, wireless and the Internet
91
256658
1409
ื”ืื™ื ื˜ืจื ื˜ ื•ื”ืจืฉืชื•ืช ื”ืืœื—ื•ื˜ื™ื•ืช
04:18
can improve health care greatly.
92
258067
1652
ื™ื›ื•ืœื•ืช ืœืชืจื•ื ื”ืžื•ืŸ ืœื‘ืจื™ืื•ืช.
04:19
There's several examples up on the screen
93
259719
2087
ื”ื ื” ื›ืžื” ื“ื•ื’ืžืื•ืช ืขืœ ื”ืžืกืš
04:21
of situations where doctors are looking to implant devices
94
261806
3107
ืฉืœ ืจื•ืคืื™ื ืฉืจื•ืฆื™ื ืœื‘ืฆืข ื”ืฉืชืœื•ืช ืฉืœ ื”ืชืงื ื™ื
04:24
inside of people, and all of these devices now,
95
264913
2865
ื‘ืชื•ืš ื‘ื ื™-ืื“ื, ื•ื›ื™ื•ื ื›ืœ ื”ื”ืชืงื ื™ื ื”ืืœื”
04:27
it's standard that they communicate wirelessly,
96
267778
3125
ืžืกื•ื’ืœื™ื ืœืชืงืฉื•ืจืช ืืœื—ื•ื˜ื™ืช ื›ื—ืœืง ืžื”ืชืงืŸ ืฉืœื”ื,
04:30
and I think this is great,
97
270903
1412
ื•ืื ื™ ื—ื•ืฉื‘ ืฉื–ื” ื ื”ื“ืจ,
04:32
but without a full understanding of trustworthy computing,
98
272315
3105
ืื‘ืœ ืื ืœื ืžื‘ื™ื ื™ื ืœื’ืžืจื™ ืžื”ื• ืžื—ืฉื•ื‘ ืžื”ื™ืžืŸ,
04:35
and without understanding what attackers can do
99
275420
2407
ื•ืื ืœื ืžื‘ื™ื ื™ื ืžื” ืคื•ืจืฆื™ ืžื—ืฉื‘ ืžืกื•ื’ืœื™ื ืœืขื•ืœืœ
04:37
and the security risks from the beginning,
100
277827
2147
ื•ืžื”ื ืกื™ื›ื•ื ื™ ื”ืื‘ื˜ื—ื” ืžืœื›ืชื—ื™ืœื”,
04:39
there's a lot of danger in this.
101
279974
2390
ื™ืฉ ื‘ื›ืš ืกื›ื ื” ืจื‘ื”.
04:42
Okay, let me shift gears and show you another target.
102
282364
1477
ื›ืขืช ืืขื‘ื™ืจ ื”ื™ืœื•ืš ื•ืืจืื” ืœื›ื ื™ืขื“ ื ื•ืกืฃ.
04:43
I'm going to show you a few different targets like this,
103
283841
2088
ืืจืื” ืœื›ื ืžืกืคืจ ื™ืขื“ื™ื ืฉื•ื ื™ื ื›ืžื• ื–ื”,
04:45
and that's my talk. So we'll look at automobiles.
104
285929
2917
ื•ื–ื” ื ื•ืฉื ื”ื”ืจืฆืื” ืฉืœื™. ื ื™ืงื— ืœื“ื•ื’ืžื” ื›ืœื™ ืจื›ื‘.
04:48
This is a car, and it has a lot of components,
105
288846
2896
ื–ืืช ืžื›ื•ื ื™ืช, ื™ืฉ ืœื” ื”ืžื•ืŸ ืžืจื›ื™ื‘ื™ื,
04:51
a lot of electronics in it today.
106
291742
1620
ื•ื”ื™ื•ื ื’ื ื”ืžื•ืŸ ืืœืงื˜ืจื•ื ื™ืงื”.
04:53
In fact, it's got many, many different computers inside of it,
107
293362
4377
ืœืžืขืฉื”, ื”ื™ื ืžื›ื™ืœื” ื”ืจื‘ื” ืžื—ืฉื‘ื™ื ืฉื•ื ื™ื,
04:57
more Pentiums than my lab did when I was in college,
108
297739
3155
ื™ื•ืชืจ ืžื—ืฉื‘ื™ "ืคื ื˜ื™ื•ื" ืžื›ืคื™ ืฉื”ื™ื• ื‘ืžืขื‘ื“ื” ืฉืœื™ ื‘ืงื•ืœื’',
05:00
and they're connected by a wired network.
109
300894
3639
ื•ื”ื ืžื—ื•ื‘ืจื™ื ื‘ืจืฉืช ืงื•ื•ื™ืช.
05:04
There's also a wireless network in the car,
110
304533
3431
ื‘ืžื›ื•ื ื™ืช ื™ืฉ ื’ื ืจืฉืช ืืœื—ื•ื˜ื™ืช,
05:07
which can be reached from many different ways.
111
307964
3233
ืฉื ื™ืชืŸ ืœื’ืฉืช ืืœื™ื” ื‘ื›ืžื” ื“ืจื›ื™ื.
05:11
So there's Bluetooth, there's the FM and XM radio,
112
311197
3701
ื™ืฉ "ื‘ืœื•ื˜ื•ืช'", ื™ืฉ ืจื“ื™ื• FM ื•-XM,
05:14
there's actually wi-fi, there's sensors in the wheels
113
314898
2820
ื™ืฉ "ื•ื•ื™ื™-ืคื™ื™", ื™ืฉ ื—ื™ื™ืฉื ื™ื ื‘ื’ืœื’ืœื™ื
05:17
that wirelessly communicate the tire pressure
114
317718
2153
ืฉืžื•ืกืจื™ื ื‘ืื•ืคืŸ ืืœื—ื•ื˜ื™ ืืช ืœื—ืฅ ื”ืฆืžื™ื’ื™ื
05:19
to a controller on board.
115
319871
1806
ืœื‘ืงืจ ืฉื ืžืฆื ื‘ืžื›ื•ื ื™ืช.
05:21
The modern car is a sophisticated multi-computer device.
116
321677
4918
ื”ืžื›ื•ื ื™ืช ื”ืžื•ื“ืจื ื™ืช ื”ื™ื ืžืชืงืŸ ืžืชื•ื—ื›ื ื•ืžืจื•ื‘ื”-ืžื—ืฉื‘ื™ื.
05:26
And what happens if somebody wanted to attack this?
117
326595
3322
ื•ืžื” ื™ืงืจื” ืื ืžื™ืฉื”ื• ื™ืจืฆื” ืœืชืงื•ืฃ ืื•ืชื”?
05:29
Well, that's what the researchers
118
329917
1317
ื–ื” ืžื” ืฉืขืฉื• ื”ื—ื•ืงืจื™ื
05:31
that I'm going to talk about today did.
119
331234
1871
ืฉื”ื™ื•ื ืื“ื‘ืจ ืขืœื™ื”ื.
05:33
They basically stuck an attacker on the wired network
120
333105
2977
ื”ื ื‘ืขืฆื ื“ื—ืคื• ื”ืชืงืŸ ืชืงื™ืคื” ืœืจืฉืช ื”ืงื•ื•ื™ืช
05:36
and on the wireless network.
121
336082
2322
ื•ืœืจืฉืช ื”ืืœื—ื•ื˜ื™ืช.
05:38
Now, they have two areas they can attack.
122
338404
2699
ืื– ื™ืฉ ืœื”ื ืฉื ื™ ืื–ื•ืจื™ื ืฉืื•ืชื ื”ื ื™ื›ื•ืœื™ื ืœืชืงื•ืฃ.
05:41
One is short-range wireless, where you can actually
123
341103
2038
ืื—ื“ ื”ื•ื ื”ืจืฉืช ื”ืืœื—ื•ื˜ื™ืช ืงืฆืจืช-ื”ื˜ื•ื•ื—, ืฉื‘ื” ืืคืฉืจ
05:43
communicate with the device from nearby,
124
343141
1781
ืœืชืงืฉืจ ืขื ื”ื”ืชืงืŸ ืžืžืจื—ืง ืงืฆืจ,
05:44
either through Bluetooth or wi-fi,
125
344922
2137
ื‘ืขื–ืจืช ื‘ืœื•ื˜ื•ืช' ืื• ื•ื•ื™ื™-ืคื™ื™,
05:47
and the other is long-range, where you can communicate
126
347059
2174
ื•ื”ืฉื ื™ ื”ื•ื ืืจื•ืš-ื˜ื•ื•ื—, ืฉื‘ื• ื ื™ืชืŸ ืœืชืงืฉืจ
05:49
with the car through the cellular network,
127
349233
1782
ืขื ื”ืžื›ื•ื ื™ืช ื‘ืจืฉืช ื”ืกืœื•ืœืจื™ืช,
05:51
or through one of the radio stations.
128
351015
1960
ืื• ื“ืจืš ืื—ืช ืžืชื—ื ื•ืช ื”ืจื“ื™ื•.
05:52
Think about it. When a car receives a radio signal,
129
352975
3049
ื—ื™ืฉื‘ื• ืขืœ ื›ืš. ื›ืฉืžื›ื•ื ื™ืช ืžืงื‘ืœืช ืื•ืช ืจื“ื™ื•,
05:56
it's processed by software.
130
356024
2201
ื”ื•ื ืžืขื•ื‘ื“ ืขืœ ื™ื“ื™ ืชื•ื›ื ื”.
05:58
That software has to receive and decode the radio signal,
131
358225
3061
ื”ืชื•ื›ื ื” ืฆืจื™ื›ื” ืœืงื‘ืœ ื•ืœืคืขื ื— ืืช ืื•ืช ื”ืจื“ื™ื•,
06:01
and then figure out what to do with it,
132
361286
1119
ื•ืื– ืœื—ืฉื•ื‘ ืžื” ืœืขืฉื•ืช ืื™ืชื•,
06:02
even if it's just music that it needs to play on the radio,
133
362405
3024
ื’ื ืื ืžื“ื•ื‘ืจ ืจืง ื‘ืžื•ืกื™ืงื” ืฉื”ื™ื ืฆืจื™ื›ื” ืœื”ืฉืžื™ืข ื‘ืžื›ืฉื™ืจ ื”ืจื“ื™ื•,
06:05
and that software that does that decoding,
134
365429
2268
ื•ื”ืชื•ื›ื ื” ืฉืžื‘ืฆืขืช ืืช ื”ืคื™ืขื ื•ื— ื”ื–ื”,
06:07
if it has any bugs in it, could create a vulnerability
135
367697
3093
ืื ื™ืฉ ืœื” ื‘ืื’ื™ื, ืขืœื•ืœื” ืœื”ื•ื•ืช ื ืงื•ื“ืช-ืชื•ืจืคื”
06:10
for somebody to hack the car.
136
370790
3035
ืฉื™ื ืฆืœ ืžื™ืฉื”ื• ืฉื™ืจืฆื” ืœืคืจื•ืฅ ืืช ืชื•ื›ื ื•ืช ื”ืจื›ื‘.
06:13
The way that the researchers did this work is,
137
373825
2952
ื”ื“ืจืš ืฉื‘ื” ื”ื—ื•ืงืจื™ื ื‘ื™ืฆืขื• ืืช ื”ืขื‘ื•ื“ื” ื”ื–ืืช,
06:16
they read the software in the computer chips
138
376777
4223
ื”ื ืงืจืื• ืืช ื”ืชื•ื›ื ื” ืฉื‘ืฉื‘ื‘ื™ ื”ืžื—ืฉื‘
06:21
that were in the car, and then they used sophisticated
139
381000
3193
ืฉืœ ื”ืžื›ื•ื ื™ืช, ื•ืื– ื”ืฉืชืžืฉื• ื‘ื›ืœื™ื ืžืชื•ื—ื›ืžื™ื
06:24
reverse engineering tools
140
384193
1414
ืฉืœ ื”ื ื“ืกื” ื”ืคื•ื›ื”
06:25
to figure out what that software did,
141
385607
2055
ื›ื“ื™ ืœื’ืœื•ืช ืžื” ื”ืชื•ื›ื ื” ืขื•ืฉื”.
06:27
and then they found vulnerabilities in that software,
142
387662
3041
ื•ืื– ื”ื ืžืฆืื• ื ืงื•ื“ื•ืช-ืชื•ืจืคื” ื‘ืื•ืชื” ืชื•ื›ื ื”,
06:30
and then they built exploits to exploit those.
143
390703
3346
ื•ื‘ื ื• ืงื•ื“ ืชื•ื›ื ื” ื–ื“ื•ื ื™ ื›ื“ื™ ืœื ืฆืœ ืืช ืื•ืชืŸ ื ืงื•ื“ื•ืช ืชื•ืจืคื”.
06:34
They actually carried out their attack in real life.
144
394049
2382
ื”ื ืžืžืฉ ื‘ื™ืฆืขื• ืืช ื”ื”ืชืงืคื” ื”ื–ื• ื‘ืคื•ืขืœ.
06:36
They bought two cars, and I guess
145
396431
1350
ื”ื ืงื ื• ืฉืชื™ ืžื›ื•ื ื™ื•ืช, ื•ื ืจืื” ืœื™
06:37
they have better budgets than I do.
146
397781
2918
ืฉื™ืฉ ืœื”ื ื™ื•ืชืจ ืชืงืฆื™ื‘ ืžืืฉืจ ืœื™.
06:40
The first threat model was to see what someone could do
147
400699
2590
ืžื•ื“ืœ ื”ืื™ื•ื ื”ืจืืฉื•ืŸ ื”ื™ื” ืœืžืฆื•ื ืžื” ืืคืฉืจ ืœืขืฉื•ืช
06:43
if an attacker actually got access
148
403289
2144
ืื ืคื•ืจืฅ ืื›ืŸ ืžืฆืœื™ื— ืœื”ืฉื™ื’ ื’ื™ืฉื”
06:45
to the internal network on the car.
149
405433
2053
ืœืจืฉืช ื”ืคื ื™ืžื™ืช ืฉืœ ื”ืจื›ื‘.
06:47
Okay, so think of that as, someone gets to go to your car,
150
407486
2603
ื—ื™ืฉื•ื‘ ืขืœ ื–ื” ื›ืš: ืžื™ืฉื”ื• ืžืฆืœื™ื— ืœื”ื’ื™ืข ืœืจื›ื‘ ืฉืœื›ื,
06:50
they get to mess around with it, and then they leave,
151
410089
2904
ืžืชืขืกืง ืื™ืชื•, ื•ืžืกืชืœืง,
06:52
and now, what kind of trouble are you in?
152
412993
2368
ื•ืขื›ืฉื™ื•, ื‘ืื™ื–ื• ืฆืจื” ืืชื?
06:55
The other threat model is that they contact you
153
415361
2792
ืžื•ื“ืœ ื”ืื™ื•ื ื”ืฉื ื™ ื”ื™ื” ืœื™ืฆื•ืจ ืชืงืฉื•ืจืช
06:58
in real time over one of the wireless networks
154
418153
2457
ื‘ื–ืžืŸ ืืžื™ืชื™ ื“ืจืš ืื—ืช ื”ืจืฉืชื•ืช ื”ืืœื—ื•ื˜ื™ื•ืช
07:00
like the cellular, or something like that,
155
420610
2055
ื›ืžื• ื”ืจืฉืช ื”ืกืœื•ืœืจื™ืช, ืื• ืžืฉื”ื• ื›ื–ื”
07:02
never having actually gotten physical access to your car.
156
422665
4000
ืœืœื ืฉื•ื ื’ื™ืฉื” ืคื™ื–ื™ืช ืืœ ื”ืžื›ื•ื ื™ืช.
07:06
This is what their setup looks like for the first model,
157
426665
2824
ื›ืš ื ืจืื” ื”ืžืขืจืš ืฉืœื”ื ืขื‘ื•ืจ ื”ืžื•ื“ืœ ื”ืจืืฉื•ืŸ,
07:09
where you get to have access to the car.
158
429489
1683
ืฉื‘ื• ื™ืฉ ืฆื•ืจืš ื‘ื’ื™ืฉื” ืœืžื›ื•ื ื™ืช.
07:11
They put a laptop, and they connected to the diagnostic unit
159
431172
3387
ื”ื ื”ื›ื ื™ืกื• ืžื—ืฉื‘ ื ื™ื™ื“, ื•ื”ื ื—ื™ื‘ืจื• ืœื™ื—ื™ื“ืช ื”ื‘ืงืจื”
07:14
on the in-car network, and they did all kinds of silly things,
160
434559
2939
ืฉื‘ืจืฉืช ื”ืคื ื™ืžื™ืช ืฉืœ ื”ืจื›ื‘, ื•ืขืฉื• ื›ืœ ืžื™ื ื™ ืฉื˜ื•ื™ื•ืช,
07:17
like here's a picture of the speedometer
161
437498
2783
ื›ืžื•, ื”ื ื” ืชืžื•ื ื” ืฉืœ ืžื“ ื”ืžื”ื™ืจื•ืช
07:20
showing 140 miles an hour when the car's in park.
162
440281
2816
ืฉืžืจืื”220 ืงืž"ืฉ ื›ืฉื”ืžื›ื•ื ื™ืช ื‘ื”ื™ืœื•ืš ื—ื ื™ื”.
07:23
Once you have control of the car's computers,
163
443097
2373
ื‘ืจื’ืข ืฉืžืฉื™ื’ื™ื ืฉืœื™ื˜ื” ื‘ืžื—ืฉื‘ื™ ื”ืจื›ื‘,
07:25
you can do anything.
164
445470
919
ืืคืฉืจ ืœืขืฉื•ืช ื”ื›ืœ.
07:26
Now you might say, "Okay, that's silly."
165
446389
1616
ืื•ืœื™ ืชื’ื™ื“ื•, "ื–ื” ืžื˜ื•ืคืฉ."
07:28
Well, what if you make the car always say
166
448005
1659
ืื‘ืœ ืžื” ืื ื ื’ืจื•ื ืœืจื›ื‘ ืฉืœื›ื ืœื”ืจืื•ืช ืชืžื™ื“
07:29
it's going 20 miles an hour slower than it's actually going?
167
449664
2741
ืฉื”ื•ื ื ื•ืกืข 30 ืงืž"ืฉ ืœืื˜ ื™ื•ืชืจ ืžืžื”ื™ืจื•ืชื• ื‘ืคื•ืขืœ?
07:32
You might produce a lot of speeding tickets.
168
452405
2542
ืืชื ืขืœื•ืœื™ื ืœื—ื˜ื•ืฃ ื”ืžื•ืŸ ื“ื•ื—ื•ืช.
07:34
Then they went out to an abandoned airstrip with two cars,
169
454947
3856
ืื—ืจ ื›ืš ื”ื ื™ืฆืื• ืขื ืฉืชื™ ืžื›ื•ื ื™ื•ืช ืœืฉื“ื” ืชืขื•ืคื” ื ื˜ื•ืฉ,
07:38
the target victim car and the chase car,
170
458803
2745
ืจื›ื‘ ื”ืžื˜ืจื”, ื”ืงื•ืจื‘ืŸ, ื•ืจื›ื‘ ื”ืžืจื“ืฃ,
07:41
and they launched a bunch of other attacks.
171
461548
2746
ื•ื”ืคืขื™ืœื• ื›ืžื” ื”ืชืงืคื•ืช ื ื•ืกืคื•ืช.
07:44
One of the things they were able to do from the chase car
172
464294
2766
ืื—ื“ ื”ื“ื‘ืจื™ื ืฉื”ื ื”ืฆืœื™ื—ื• ืœื‘ืฆืข ืžืžื›ื•ื ื™ืช ื”ืžืจื“ืฃ
07:47
is apply the brakes on the other car,
173
467060
1974
ื”ื™ื” ืœืœื—ื•ืฅ ืขืœ ื”ื‘ืœืžื™ื ื‘ืจื›ื‘ ื”ืฉื ื™,
07:49
simply by hacking the computer.
174
469034
1560
ืคืฉื•ื˜ ื‘ื›ืš ืฉืคืจืฆื• ืœืžื—ืฉื‘.
07:50
They were able to disable the brakes.
175
470594
2431
ื”ื ื”ืฆืœื™ื—ื• ื’ื ืœื ื˜ืจืœ ืืช ื”ื‘ืœืžื™ื.
07:53
They also were able to install malware that wouldn't kick in
176
473025
3178
ื”ื ื™ื›ืœื• ืœืฉืชื•ืœ ืชื•ื›ื ื” ื–ื“ื•ื ื™ืช ืฉืœื ืชื™ื›ื ืก ืœืคืขื•ืœื”
07:56
and wouldn't trigger until the car was doing something like
177
476203
2425
ื•ืœื ืชื•ืคืขืœ ืขื“ ืฉื”ืจื›ื‘ ื™ืขืฉื” ืžืฉื”ื• ื›ืžื•
07:58
going over 20 miles an hour, or something like that.
178
478628
3746
ืœื ืกื•ืข ื‘ืžื”ื™ืจื•ืช ืฉืœ ืžืขืœ 30 ืงืž"ืฉ, ืื• ืžืฉื”ื• ื›ื–ื”.
08:02
The results are astonishing, and when they gave this talk,
179
482374
2758
ื”ืชื•ืฆืื•ืช ื”ื™ื• ืžื“ื”ื™ืžื•ืช, ื•ื›ืฉื”ื ื”ืจืฆื• ืขืœ ื–ื”,
08:05
even though they gave this talk at a conference
180
485132
1716
ื’ื ื›ืฉื”ื ื”ืจืฆื• ืขืœ ื–ื” ื‘ื›ื ืก
08:06
to a bunch of computer security researchers,
181
486848
1726
ืฉืœ ื—ื‘ื•ืจืช ื—ื•ืงืจื™ ืื‘ื˜ื—ืช ืžื—ืฉื‘ื™ื,
08:08
everybody was gasping.
182
488574
1700
ื”ื ืฉื™ืžื” ืฉืœ ื›ื•ืœื ื ืขืชืงื”.
08:10
They were able to take over a bunch of critical computers
183
490274
3699
ื”ื ื”ืฆืœื™ื—ื• ืœื”ืฉืชืœื˜ ืขืœ ืžืกืคืจ ืžื—ืฉื‘ื™ื ืงืจื™ื˜ื™ื™ื
08:13
inside the car: the brakes computer, the lighting computer,
184
493973
3761
ื‘ืชื•ืš ืžื›ื•ื ื™ืช: ืžื—ืฉื‘ ื”ื‘ืœืžื™ื, ืžื—ืฉื‘ ื”ืชืื•ืจื”,
08:17
the engine, the dash, the radio, etc.,
185
497734
2827
ื”ืžื ื•ืข, ืœื•ื— ื”ืฉืขื•ื ื™ื, ื”ืจื“ื™ื• ื•ื›ื•',
08:20
and they were able to perform these on real commercial
186
500561
2293
ื•ื”ื ื”ืฆืœื™ื—ื• ืœืขืฉื•ืช ืืช ื–ื” ื‘ืžื›ื•ื ื™ื•ืช ืจื’ื™ืœื•ืช ืฉื ืžืฆืื•ืช ื‘ืฉื•ืง,
08:22
cars that they purchased using the radio network.
187
502854
3027
ื‘ืืžืฆืขื•ืช ืจืฉืช ื”ืจื“ื™ื•.
08:25
They were able to compromise every single one of the
188
505881
3003
ื”ื ื”ืฆืœื™ื—ื• ืœืคืจื•ืฅ ื›ืœ ืื—ืช
08:28
pieces of software that controlled every single one
189
508884
2466
ืžื”ืชื•ื›ื ื•ืช ืฉืฉื•ืœื˜ื•ืช ื‘ื›ืœ ืื—ืช
08:31
of the wireless capabilities of the car.
190
511350
3015
ืžื”ื™ื›ื•ืœื•ืช ื”ืืœื—ื•ื˜ื™ื•ืช ืฉืœ ื”ืจื›ื‘.
08:34
All of these were implemented successfully.
191
514365
2513
ื›ืœ ื”ื“ื‘ืจื™ื ื”ืืœื” ื™ื•ืฉืžื• ื‘ื”ืฆืœื—ื”.
08:36
How would you steal a car in this model?
192
516878
2352
ืื™ืš ื’ื•ื ื‘ื™ื ืžื›ื•ื ื™ืช ืœืคื ื™ ื”ืžื•ื“ืœ ื”ื–ื”?
08:39
Well, you compromise the car by a buffer overflow
193
519230
3680
ืคื•ืจืฆื™ื ืืช ืชื•ื›ื ืช ื”ืจื›ื‘ ื‘ืขื–ืจืช ื’ืœื™ืฉืช ื—ื•ืฆืฅ
08:42
of vulnerability in the software, something like that.
194
522910
2527
ื‘ื ืงื•ื“ืช-ืชื•ืจืคื” ืฉืœ ื”ืชื•ื›ื ื”, ืžืฉื”ื• ื›ื–ื”.
08:45
You use the GPS in the car to locate it.
195
525437
2203
ืžืฉืชืžืฉื™ื ื‘ืžืขืจื›ืช ื”ืื™ื›ื•ืŸ ื”ืœื•ื•ื™ื™ื ื™ ืฉืœ ื”ืจื›ื‘ ื›ื“ื™ ืœืืชืจ ืื•ืชื•.
08:47
You remotely unlock the doors through the computer
196
527640
2195
ืคื•ืชื—ื™ื ืืช ื”ื“ืœืชื•ืช ื‘ืฉืœื™ื˜ื” ืžืจื—ื•ืง ื“ืจืš ื”ืžื—ืฉื‘
08:49
that controls that, start the engine, bypass anti-theft,
197
529835
3138
ืฉืฉื•ืœื˜ ื‘ื”ืŸ, ืžืคืขื™ืœื™ื ืืช ื”ืžื ื•ืข, ืขื•ืงืคื™ื ืืช ื”ื”ื’ื ื” ื ื’ื“ ืคืจื™ืฆื”,
08:52
and you've got yourself a car.
198
532973
1668
ื•ื™ืฉ ืœื›ื ืžื›ื•ื ื™ืช.
08:54
Surveillance was really interesting.
199
534641
2487
ื”ืžืขืงื‘ ื”ื™ื” ืžืžืฉ ืžืขื ื™ื™ืŸ.
08:57
The authors of the study have a video where they show
200
537128
3209
ืœื›ื•ืชื‘ื™ ื”ืžื—ืงืจ ื™ืฉ ืกืจื˜ื•ืŸ ืฉื‘ื• ื”ื ืžืฆื™ื’ื™ื ืืช ืขืฆืžื
09:00
themselves taking over a car and then turning on
201
540337
2549
ืžืฉืชืœื˜ื™ื ืขืœ ืžื›ื•ื ื™ืช,
09:02
the microphone in the car, and listening in on the car
202
542886
2761
ืžืคืขื™ืœื™ื ื‘ื” ืืช ื”ืžื™ืงืจื•ืคื•ืŸ ื•ืžืฆื•ืชืชื™ื ืœืžื” ืฉืงื•ืจื” ื‘ื”
09:05
while tracking it via GPS on a map,
203
545647
3351
ืชื•ืš ืžืขืงื‘ ืื—ืจื™ื” ืขืœ ื™ื“ื™ ืื™ื›ื•ืŸ ืœื•ื•ื™ื™ื ื™ ืขืœ ื’ื‘ื™ ืžืคื”.
09:08
and so that's something that the drivers of the car
204
548998
1713
ื•ื–ื” ื“ื‘ืจ ืฉื ื”ื’ ื”ืžื›ื•ื ื™ืช
09:10
would never know was happening.
205
550711
2168
ืœืขื•ืœื ืœื ื™ื“ืข ืฉืงื•ืจื”.
09:12
Am I scaring you yet?
206
552879
2134
ื›ื‘ืจ ื”ืฆืœื—ืชื™ ืœื”ืคื—ื™ื“ ืืชื›ื?
09:15
I've got a few more of these interesting ones.
207
555013
1943
ื™ืฉ ืœื™ ืขื•ื“ ื›ืžื” ื“ื•ื’ืžืื•ืช ืžืขื ื™ื™ื ื•ืช ื›ืืœื”.
09:16
These are ones where I went to a conference,
208
556956
1833
ืืœื” ื“ื•ื’ืžืื•ืช ืฉืฉืžืขืชื™ ืคืขื ื‘ื›ื ืก
09:18
and my mind was just blown, and I said,
209
558789
1933
ื•ืคืฉื•ื˜ ื ื“ื”ืžืชื™ ื•ืืžืจืชื™,
09:20
"I have to share this with other people."
210
560722
1826
"ืื ื™ ื—ื™ื™ื‘ ืœืฉืชืฃ ื‘ื›ืš ืขื•ื“ ืื ืฉื™ื."
09:22
This was Fabian Monrose's lab
211
562548
1623
ื–ื” ื”ื™ื” ื‘ืžืขื‘ื“ื” ืฉืœ ืคื‘ื™ืืŸ ืžื•ื ืจื•ื–
09:24
at the University of North Carolina, and what they did was
212
564171
3456
ื‘ืื•ื ื™ื‘ืจืกื™ื˜ืช ืฆืคื•ืŸ-ืงืจื•ืœื™ื ื”, ื•ืžื” ืฉื”ื ืขืฉื•
09:27
something intuitive once you see it,
213
567627
2075
ื”ื•ื ืžื•ื‘ืŸ ืžืืœื™ื• ื›ืฉืจื•ืื™ื ืื•ืชื•,
09:29
but kind of surprising.
214
569702
1714
ืื‘ืœ ื“ื™ ืžืคืชื™ืข.
09:31
They videotaped people on a bus,
215
571416
2259
ื”ื ืฆื™ืœืžื• ื‘ื•ื•ื™ื“ื™ืื• ืื ืฉื™ื ื‘ืื•ื˜ื•ื‘ื•ืก,
09:33
and then they post-processed the video.
216
573675
2840
ื•ืื– ืขืฉื• ืขื™ื‘ื•ื“ ืฉืœ ื”ืกืจื˜ื•ืŸ.
09:36
What you see here in number one is a
217
576515
2463
ืžื” ืฉืืชื ืจื•ืื™ื ื‘ืžืกืคืจ 1
09:38
reflection in somebody's glasses of the smartphone
218
578978
4383
ื”ื•ื ื”ืฉืชืงืคื•ืช ื”ืกืžืจื˜ืคื•ืŸ ืฉืœ ืžื™ืฉื”ื• ื‘ืžืฉืงืคื™ื•
09:43
that they're typing in.
219
583361
1425
ื‘ื–ืžืŸ ืฉื”ื•ื ืžืงืœื™ื“ ื‘ื•.
09:44
They wrote software to stabilize --
220
584786
1975
ื”ื ื›ืชื‘ื• ืชื•ื›ื ื” ืฉืžื™ื™ืฆื‘ืช -
09:46
even though they were on a bus
221
586761
1365
ื’ื ืื ื–ื” ื‘ืื•ื˜ื•ื‘ื•ืก
09:48
and maybe someone's holding their phone at an angle --
222
588126
3211
ื•ืžื™ืฉื”ื• ืžื—ื–ื™ืง ืืช ื”ื˜ืœืคื•ืŸ ืฉืœื• ื‘ื–ื•ื•ื™ืช -
09:51
to stabilize the phone, process it, and
223
591337
2370
ื›ื“ื™ ืœื™ื™ืฆื‘ ืืช ื”ื˜ืœืคื•ืŸ, ืœืขื‘ื“ ืืช ื–ื”,
09:53
you may know on your smartphone, when you type
224
593707
1885
ื•ืื–, ืื•ืœื™ ืืชื ืžื›ื™ืจื™ื ืืช ื–ื”, ื›ืฉืžืงืœื™ื“ื™ื ื‘ืกืžืจื˜ืคื•ืŸ
09:55
a password, the keys pop out a little bit, and they were able
225
595592
2939
ืกื™ืกืžื”, ื”ืื•ืชื™ื•ืช ืงื•ืคืฆื•ืช ื”ื—ื•ืฆื”, ื•ื”ื ื”ืฆืœื™ื—ื•
09:58
to use that to reconstruct what the person was typing,
226
598531
2840
ืœืฉื—ื–ืจ ืืช ืžื” ืฉืื•ืชื• ืื“ื ื”ืงืœื™ื“,
10:01
and had a language model for detecting typing.
227
601371
4321
ื•ื”ื™ื” ืœื”ื ืžื•ื“ืœ ืฉืคื” ืœื–ื™ื”ื•ื™ ื”ืงืœื“ื”.
10:05
What was interesting is, by videotaping on a bus,
228
605692
2335
ืžื” ืฉื”ื™ื” ืžืขื ื™ื™ืŸ ื–ื”, ืขืœ ื™ื“ื™ ืฆื™ืœื•ื ื‘ืื•ื˜ื•ื‘ื•ืก,
10:08
they were able to produce exactly what people
229
608027
2129
ื”ื ื”ื™ื• ืžืกื•ื’ืœื™ื ืœื—ื–ื•ืจ ื‘ื“ื™ื•ืง ืขืœ ืžื” ืฉืื ืฉื™ื
10:10
on their smartphones were typing,
230
610156
2151
ื”ืงืœื™ื“ื• ื‘ืกืžืจื˜ืคื•ื ื™ื ืฉืœื”ื,
10:12
and then they had a surprising result, which is that
231
612307
2260
ื•ืื– ื”ื ื’ื™ืœื• ืžืฉื”ื• ืžืคืชื™ืข,
10:14
their software had not only done it for their target,
232
614567
2764
ืฉื”ืชื•ื›ื ื” ืฉืœื”ื ืœื ืขืฉืชื” ืืช ื–ื” ืจืง ืœืžื™ ืฉื”ืชื›ื•ื•ื ื•,
10:17
but other people who accidentally happened
233
617331
1403
ืืœื ื’ื ืœืื ืฉื™ื ืื—ืจื™ื ืฉื‘ืžืงืจื” ื”ื™ื•
10:18
to be in the picture, they were able to produce
234
618734
2086
ื‘ืชืžื•ื ื”, ื”ื ื”ื™ื• ืžืกื•ื’ืœื™ื ืœืฉื—ื–ืจ
10:20
what those people had been typing, and that was kind of
235
620820
2727
ืžื” ื”ืื ืฉื™ื ื”ืืœื” ื”ืงืœื™ื“ื•, ื•ื–ื” ื”ื™ื” ืกื•ื’ ืฉืœ
10:23
an accidental artifact of what their software was doing.
236
623547
3617
ืชื•ืคืขืช ืœื•ื•ืื™ ืžืงืจื™ืช ืฉืœ ืžื” ืฉื”ืชื•ื›ื ื” ืฉืœื”ื ืขื•ืฉื”.
10:27
I'll show you two more. One is P25 radios.
237
627164
4303
ืื ื™ ืืจืื” ืœื›ื ืขื•ื“ ืฉืชื™ื™ื. ืื—ืช ื”ื™ื ืžื›ืฉื™ืจื™ ืจื“ื™ื• P25.
10:31
P25 radios are used by law enforcement
238
631467
2800
ืžื›ืฉื™ืจื™ ืจื“ื™ื• P25 ื ืžืฆืื™ื ื‘ืฉื™ืžื•ืฉ ื‘ืžื ื’ื ื•ื ื™ ืื›ื™ืคืช ื”ื—ื•ืง
10:34
and all kinds of government agencies
239
634267
3407
ื•ื‘ื›ืœ ืžื™ื ื™ ืกื•ื›ื ื•ื™ื•ืช ืžืžืฉืœืชื™ื•ืช
10:37
and people in combat to communicate,
240
637674
1736
ื•ืœืชืงืฉื•ืจืช ืฉืœ ืื ืฉื™ื ื‘ืœื—ื™ืžื”,
10:39
and there's an encryption option on these phones.
241
639410
2833
ื•ื™ืฉ ืื•ืคืฆื™ืช ื”ืฆืคื ื” ืขืœ ื”ืžื›ืฉื™ืจื™ื ื”ืืœื”.
10:42
This is what the phone looks like. It's not really a phone.
242
642243
2728
ื›ื›ื” ื ืจืื” ื”ื˜ืœืคื•ืŸ. ื–ื” ืœื ื‘ืืžืช ื˜ืœืคื•ืŸ.
10:44
It's more of a two-way radio.
243
644971
1206
ื–ื” ื™ื•ืชืจ ืžื›ืฉื™ืจ ืงืฉืจ.
10:46
Motorola makes the most widely used one, and you can see
244
646177
3322
ืžื•ื˜ื•ืจื•ืœื” ืžื™ื™ืฆืจืช ืืช ื”ื˜ืœืคื•ืŸ ื”ื ืคื•ืฅ ื‘ื™ื•ืชืจ, ื•ืืชื ื™ื›ื•ืœื™ื ืœืจืื•ืช
10:49
that they're used by Secret Service, they're used in combat,
245
649499
2649
ืฉื”ื ื‘ืฉื™ืžื•ืฉ ื”ืฉืจื•ืช ื”ื—ืฉืื™, ื‘ืœื—ื™ืžื”,
10:52
it's a very, very common standard in the U.S. and elsewhere.
246
652148
3102
ื–ื” ืชืงืŸ ืžืื•ื“ ืžืื•ื“ ื ืคื•ืฅ ื‘ืืจื”"ื‘ ื•ื‘ืืจืฆื•ืช ืื—ืจื•ืช.
10:55
So one question the researchers asked themselves is,
247
655250
2305
ืื– ืฉืืœื” ืื—ืช ืฉื”ื—ื•ืงืจื™ื ืฉืืœื• ืืช ืขืฆืžื ื”ื™ื,
10:57
could you block this thing, right?
248
657555
2704
ื”ืื ืืคืฉืจ ืœื—ืกื•ื ืืช ื”ื“ื‘ืจื™ื ื”ืืœื”, ื›ืŸ?
11:00
Could you run a denial-of-service,
249
660259
1583
ื”ืื ืืคืฉืจ ืœื”ืคืขื™ืœ ื”ืชืงืคื” ืฉืœ ืžื ื™ืขืช ืฉืจื•ืช,
11:01
because these are first responders?
250
661842
1824
ืžืคื ื™ ืฉืืœื” ื”ืžื’ื™ื‘ื™ื ื”ืจืืฉื•ื ื™ื?
11:03
So, would a terrorist organization want to black out the
251
663666
1801
ืื– ืื ืืจื’ื•ืŸ ื˜ืจื•ืจ ื™ืจืฆื” ืœื ื˜ืจืœ ืืช
11:05
ability of police and fire to communicate at an emergency?
252
665467
4488
ื”ื™ื›ื•ืœืช ืฉืœ ื”ืžืฉื˜ืจื” ื•ืžื›ื‘ื™ ื”ืืฉ ืœืชืงืฉืจ ื‘ื–ืžืŸ ื—ื™ืจื•ื?
11:09
They found that there's this GirlTech device used for texting
253
669955
3072
ื”ื ืžืฆืื• ืฉื™ืฉ ืืช ืžื›ืฉื™ืจ ื”GirlTech ืฉืžืฉืžืฉ ืœืžืกืจื•ื ื™ื
11:13
that happens to operate at the same exact frequency
254
673027
2718
ืฉื‘ืžืงืจื” ืคื•ืขืœ ื‘ื“ื™ื•ืง ื‘ืื•ืชื• ืชื“ืจ
11:15
as the P25, and they built what they called
255
675745
2271
ื›ืžื• ืžื›ืฉื™ืจื™ ื”-P25, ื•ื”ื ื‘ื ื• ืืช ืžื” ืฉื”ื ื›ื™ื ื•
11:18
My First Jammer. (Laughter)
256
678016
4334
ื”ืžืฉื‘ืฉ ื”ืจืืฉื•ืŸ ืฉืœื™. (ืฆื—ื•ืง)
11:22
If you look closely at this device,
257
682350
2378
ืื ืชื‘ื—ื ื• ืืช ื”ืžื›ืฉื™ืจ ื”ื™ื˜ื‘,
11:24
it's got a switch for encryption or cleartext.
258
684728
3630
ื™ืฉ ืœื• ื›ืคืชื•ืจ ืœื˜ืงืกื˜ ืจื’ื™ืœ ืื• ืžื•ืฆืคืŸ.
11:28
Let me advance the slide, and now I'll go back.
259
688358
3050
ืชื ื• ืœื™ ืœื”ืขื‘ื™ืจ ืฉืงื•ืคื™ืช, ื•ืขื›ืฉื™ื• ืื ื™ ืื—ื–ื•ืจ.
11:31
You see the difference?
260
691408
2547
ืืชื ืจื•ืื™ื ืืช ื”ื”ื‘ื“ืœ?
11:33
This is plain text. This is encrypted.
261
693955
2557
ื–ื” ื˜ืงื˜ ืจื’ื™ืœ. ื–ื” ืžื•ืฆืคืŸ.
11:36
There's one little dot that shows up on the screen,
262
696512
2557
ื™ืฉ ื ืงื•ื“ื” ืงื˜ื ื” ืื—ืช ืฉืžื•ืคื™ืขื” ืขืœ ื”ืžืกืš,
11:39
and one little tiny turn of the switch.
263
699069
2085
ื•ืกื™ื‘ื•ื‘ ืงื˜ืงื˜ืŸ ืฉืœ ื›ืคืชื•ืจ.
11:41
And so the researchers asked themselves, "I wonder how
264
701154
1904
ื•ื”ื—ื•ืงืจื™ื ืฉืืœื• ืืช ืขืฆืžื, "ืžืขื ื™ื™ืŸ ื›ืžื”
11:43
many times very secure, important, sensitive conversations
265
703058
4257
ืคืขืžื™ื ืฉื™ื—ื•ืช ืžืื•ื“ ืžืื•ื‘ื˜ื—ื•ืช, ื—ืฉื•ื‘ื•ืช ื•ืจื’ื™ืฉื•ืช
11:47
are happening on these two-way radios where they forget
266
707315
1623
ืžื•ืคื™ืขื•ืช ื‘ืžื›ืฉื™ืจื™ ื”ืงืฉืจ ื”ืืœื” ื›ืฉื”ื ืฉื•ื›ื—ื™ื ืœื”ืฆืคื™ืŸ,
11:48
to encrypt and they don't notice that they didn't encrypt?"
267
708938
2910
ื•ืœื ืฉืžื™ื ืœื‘ ืฉื”ื ืฉื›ื—ื• ืœื”ืฆืคื™ืŸ?"
11:51
So they bought a scanner. These are perfectly legal
268
711848
3339
ืื– ื”ื ืงื ื• ืกื•ืจืง. ืืœื” ืกื•ืจืงื™ื ื—ื•ืงื™ื™ื ืœื’ืžืจื™
11:55
and they run at the frequency of the P25,
269
715187
3458
ื•ื”ื ืคื•ืขืœื™ื ื‘ืชื“ืจ ืฉืœ ื”-P25,
11:58
and what they did is they hopped around frequencies
270
718645
1767
ื•ืžื” ืฉื”ื ืขืฉื• ื–ื” ืฉื”ื ืขืœื• ืขืœ ืชื“ืจื™ื
12:00
and they wrote software to listen in.
271
720412
2510
ื•ื›ืชื‘ื• ืชื•ื›ื ื” ืœื”ืื–ื ื”.
12:02
If they found encrypted communication, they stayed
272
722922
2634
ืื ื”ื ืžืฆืื• ืชืงืฉื•ืจืช ืžื•ืฆืคื ืช,
12:05
on that channel and they wrote down, that's a channel
273
725556
1686
ื”ื ื ืฉืืจื• ืขืœ ื”ืขืจื•ืฅ ื”ื–ื” ื•ืฆื™ื™ื ื• ืฉื–ื” ืขืจื•ืฅ
12:07
that these people communicate in,
274
727242
1788
ืฉื”ืื ืฉื™ื ื”ืืœื” ืžืชืงืฉืจื™ื ื‘ื•,
12:09
these law enforcement agencies,
275
729030
1622
ืกื•ื›ื ื•ื™ื•ืช ืื›ื™ืคืช ื”ื—ื•ืง ื”ืืœื•,
12:10
and they went to 20 metropolitan areas and listened in
276
730652
3391
ื•ื”ื ื”ืœื›ื• ืœ-20 ืื–ื•ืจื™ื ืขื™ืจื•ื ื™ื™ื ื•ื”ืื–ื™ื ื•
12:14
on conversations that were happening at those frequencies.
277
734043
3475
ืœืฉื™ื—ื•ืช ืฉื”ื™ื• ืขืœ ื”ืชื“ืจื™ื ื”ืืœื”.
12:17
They found that in every metropolitan area,
278
737518
3239
ื”ื ืžืฆืื• ืฉื‘ื›ืœ ืื–ื•ืจ ืขื™ืจื•ื ื™,
12:20
they would capture over 20 minutes a day
279
740757
2154
ื”ื ื”ื™ื• ืœื•ื›ื“ื™ื ืžืขืœ 20 ื“ืงื•ืช ื‘ื™ื•ื
12:22
of cleartext communication.
280
742911
2375
ืฉืœ ืฉื™ื—ื•ืช ื‘ื˜ืงืกื˜ ื—ื•ืคืฉื™.
12:25
And what kind of things were people talking about?
281
745286
2000
ื•ืขืœ ืžื” ื”ืื ืฉื™ื ื”ืืœื” ื“ื™ื‘ืจื•?
12:27
Well, they found the names and information
282
747286
1484
ื•ื‘ื›ืŸ, ื”ื ื’ื™ืœื• ืฉืžื•ืช ื•ืžื™ื“ืข
12:28
about confidential informants. They found information
283
748770
2852
ืขืœ ืžื•ื“ื™ืขื™ื ื—ืฉืื™ื™ื. ื”ื ื’ื™ืœื• ืžื™ื“ืข
12:31
that was being recorded in wiretaps,
284
751622
2202
ืฉื”ื•ืงืœื˜ ื‘ื”ืื–ื ื•ืช ืกืชืจ,
12:33
a bunch of crimes that were being discussed,
285
753824
2710
ื›ืžื” ืคืฉืขื™ื ืฉื“ื™ื‘ืจื• ืขืœื™ื”ื,
12:36
sensitive information.
286
756534
1162
ืžื™ื“ืข ืจื’ื™ืฉ.
12:37
It was mostly law enforcement and criminal.
287
757696
3363
ื–ื” ื”ื™ื” ื‘ืขื™ืงืจ ืžื™ื“ืข ืฉืงืฉื•ืจ ืœืื›ื™ืคืช ื—ื•ืง ื•ืคืœื™ืœื™.
12:41
They went and reported this to the law enforcement
288
761059
1834
ื”ื ื”ืœื›ื• ื•ื“ื™ื•ื•ื—ื• ืขืœ ื–ื” ืœืจืฉื•ื™ื•ืช ื”ื—ื•ืง
12:42
agencies, after anonymizing it,
289
762893
2023
ืื—ืจื™ ืฉืžื—ืงื• ืืช ื”ืคืจื˜ื™ื ื”ืžื–ื”ื™ื,
12:44
and the vulnerability here is simply the user interface
290
764916
3000
ื•ื ืงื•ื“ืช ื”ืชื•ืจืคื” ืคื” ื”ื™ื ืคืฉื•ื˜ ืžืžืฉืง ื”ืžืฉืชืžืฉ
12:47
wasn't good enough. If you're talking
291
767916
1394
ืฉืœื ื”ื™ื” ื˜ื•ื‘ ืžืกืคื™ืง. ืื ืืชื ืžื“ื‘ืจื™ื
12:49
about something really secure and sensitive, it should
292
769310
2816
ืขืœ ืžืฉื”ื• ื‘ืืžืช ื—ืฉืื™ ื•ืจื’ื™ืฉ, ื–ื” ืฆืจื™ืš ืœื”ื™ื•ืช
12:52
be really clear to you that this conversation is encrypted.
293
772126
3293
ื‘ืจื•ืจ ืœื›ื ืœื—ืœื•ื˜ื™ืŸ ืฉื”ืฉื™ื—ื” ืžื•ืฆืคื ืช.
12:55
That one's pretty easy to fix.
294
775419
1886
ื–ื” ืžืฉื”ื• ืฉืžืื•ื“ ืคืฉื•ื˜ ืœืชืงืŸ.
12:57
The last one I thought was really, really cool,
295
777305
1669
ื”ืื—ืจื•ืŸ ืœื“ืขืชื™ ื”ื•ื ืžืžืฉ ืžืžืฉ ืžื’ื ื™ื‘,
12:58
and I just had to show it to you, it's probably not something
296
778974
2813
ื•ืคืฉื•ื˜ ื”ื™ื™ืชื™ ื—ื™ื™ื‘ ืœื”ืจืื•ืช ืœื›ื ืืช ื–ื”, ื–ื” ื›ื ืจืื” ืœื ืžืฉื”ื•
13:01
that you're going to lose sleep over
297
781787
1005
ืฉืœื ืชื™ืฉื ื• ื‘ื’ืœืœื•
13:02
like the cars or the defibrillators,
298
782792
1791
ื›ืžื• ื”ืžื›ื•ื ื™ื•ืช ืื• ื”ื“ื™ืคื™ื‘ื•ืœื˜ื•ืจื™ื,
13:04
but it's stealing keystrokes.
299
784583
3023
ืื‘ืœ ืžื“ื•ื‘ืจ ื‘ื’ื ื™ื‘ืช ื”ืงืœื“ื•ืช.
13:07
Now, we've all looked at smartphones upside down.
300
787606
2747
ืขื›ืฉื™ื•, ื›ื•ืœื ื• ื”ื‘ื˜ื ื• ืขืœ ืกืžืจื˜ืคื•ื ื™ื ืžื›ืœ ื›ื™ื•ื•ืŸ.
13:10
Every security expert wants to hack a smartphone,
301
790353
2190
ื›ืœ ืžื•ืžื—ื” ืื‘ื˜ื—ื” ืจื•ืฆื” ืœืคืจื•ืฅ ืœืกืžืจื˜ืคื•ืŸ,
13:12
and we tend to look at the USB port, the GPS for tracking,
302
792543
4612
ืื ื—ื ื• ื ื•ื˜ื™ื ืœื‘ื“ื•ืง ืืช ื›ื ื™ืกืช ื”-USB, ื”-GPS ืœืžืขืงื‘,
13:17
the camera, the microphone, but no one up till this point
303
797155
3208
ืืช ื”ืžืฆืœืžื”, ืืช ื”ืžื™ืงืจื•ืคื•ืŸ, ืื‘ืœ ืขื“ ืขื›ืฉื™ื•
13:20
had looked at the accelerometer.
304
800363
1580
ืืฃ ืื—ื“ ืœื ื‘ื“ืง ืืช ืžื“ ื”ืชืื•ืฆื”.
13:21
The accelerometer is the thing that determines
305
801943
1647
ืžื“ ื”ืชืื•ืฆื” ื”ื•ื ืžื” ืฉืงื•ื‘ืข
13:23
the vertical orientation of the smartphone.
306
803590
3494
ืืช ื”ืื•ืจื™ืื ื˜ืฆื™ื” ื”ืื ื›ื™ืช ืฉืœ ื”ืžื›ืฉื™ืจ.
13:27
And so they had a simple setup.
307
807084
1417
ืื– ื”ื™ื” ืœื”ื ื”ืชืงื ื” ืคืฉื•ื˜ื”.
13:28
They put a smartphone next to a keyboard,
308
808501
2758
ื”ื ืฉืžื• ืืช ื”ืžื›ืฉื™ืจ ืœื™ื“ ื”ืžืงืœื“ืช,
13:31
and they had people type, and then their goal was
309
811259
2712
ื•ื”ื ื ืชื ื• ืœืื ืฉื™ื ืœื”ืงืœื™ื“, ื•ื”ืžื˜ืจื” ืฉืœื”ื ื”ื™ืชื”
13:33
to use the vibrations that were created by typing
310
813971
2856
ืœื”ืฉืชืžืฉ ื‘ื•ื•ื™ื‘ืจืฆื™ื•ืช ืฉื ื•ืฆืจื• ืžื”ื”ืงืœื“ื”
13:36
to measure the change in the accelerometer reading
311
816827
4240
ื•ืœืžื“ื•ื“ ืืช ื”ืฉื™ื ื•ื™ ื‘ืงืจื™ืืช ืžื“ ื”ืชืื•ืฆื”
13:41
to determine what the person had been typing.
312
821067
3176
ื›ื“ื™ ืœืงื‘ื•ืข ืžื” ื”ืื“ื ื”ืงืœื™ื“.
13:44
Now, when they tried this on an iPhone 3GS,
313
824243
2576
ืขื›ืฉื™ื•, ื›ืฉื”ื ื ื™ืกื• ืืช ื–ื” ืขืœ ืื™ื™ืคื•ืŸ 3GS,
13:46
this is a graph of the perturbations that were created
314
826819
2769
ื–ื” ื”ื’ืจืฃ ืฉืœ ื”ืจืขื™ื“ื•ืช ืฉื ื•ืฆืจื• ืžื”ื”ืงืœื“ื”,
13:49
by the typing, and you can see that it's very difficult
315
829588
3241
ื•ืืชื ื™ื›ื•ืœื™ื ืœืจืื•ืช ืฉืžืื•ื“ ืงืฉื”
13:52
to tell when somebody was typing or what they were typing,
316
832829
3078
ืœื”ื’ื™ื“ ืžืชื™ ืžื™ืฉื”ื• ื”ืงืœื™ื“ ืื• ืžื” ื”ื•ื ื”ืงืœื™ื“,
13:55
but the iPhone 4 greatly improved the accelerometer,
317
835907
3090
ืื‘ืœ ื‘ืื™ื™ืคื•ืŸ 4 ืฉื™ืคืจื• ืžืฉืžืขื•ืชื™ืช ืืช ืžื“ ื”ืชืื•ืฆื”,
13:58
and so the same measurement
318
838997
3480
ื•ืื– ืื•ืชืŸ ืžื“ื™ื“ื•ืช
14:02
produced this graph.
319
842477
1832
ื™ืฆืจื• ืืช ื”ื’ืจืฃ ื”ื–ื”.
14:04
Now that gave you a lot of information while someone
320
844309
2486
ืขื›ืฉื™ื• ื–ื” ืžืกืคืง ื”ืจื‘ื” ื™ื•ืชืจ ืžื™ื“ืข ื›ืฉืžื™ืฉื”ื• ืžืงืœื™ื“
14:06
was typing, and what they did then is used advanced
321
846795
3241
ื•ืžื” ืฉื”ื ืขืฉื• ื–ื” ืœื”ืฉืชืžืฉ ื‘ื˜ื›ื ื™ืงื•ืช ืžืชืงื“ืžื•ืช
14:10
artificial intelligence techniques called machine learning
322
850036
3007
ืฉืœ ืื™ื ื˜ืœื™ื’ื ืฆื™ื” ืžืœืื›ื•ืชื™ืช ืฉื ืงืจืืช ืœื™ืžื•ื“ ืžื›ื•ื ื”
14:13
to have a training phase,
323
853043
1431
ืœืฉืœื‘ ื”ืื™ืžื•ืŸ,
14:14
and so they got most likely grad students
324
854474
2236
ืื– ื”ื ืœืงื—ื• ืžืŸ ื”ืกืชื ืกื˜ื•ื“ื ื˜ื™ื
14:16
to type in a whole lot of things, and to learn,
325
856710
3789
ืœื”ืงืœื™ื“ ื›ืœ ืžื™ื ื™ ื“ื‘ืจื™ื, ื•ืœืœืžื•ื“,
14:20
to have the system use the machine learning tools that
326
860499
2768
ื›ื“ื™ ืฉื”ืžืขืจื›ืช ืชืฉืชืžืฉ ื‘ื›ืœื™ ืœื™ืžื•ื“ ืžื›ื•ื ื”
14:23
were available to learn what it is that the people were typing
327
863267
2863
ืฉื”ื™ื• ื–ืžื™ื ื™ื ื›ื“ื™ ืœืœืžื•ื“ ืžื” ืื ืฉื™ื ืžืงืœื™ื“ื™ื
14:26
and to match that up
328
866130
2827
ื•ืœื”ืชืื™ื ืืช ื–ื”
14:28
with the measurements in the accelerometer.
329
868957
2477
ืœืžื“ื™ื“ื•ืช ืฉืœ ืžื“ ื”ืชืื•ืฆื”.
14:31
And then there's the attack phase, where you get
330
871434
1635
ื•ืื– ื™ืฉ ืืช ืฉืœื‘ ื”ืžืชืงืคื”, ื‘ื• ืžื‘ืงืฉื™ื ืžืžื™ืฉื”ื•
14:33
somebody to type something in, you don't know what it was,
331
873069
2811
ืœื”ืงืœื™ื“ ืžืฉื”ื•, ืืชื ืœื ื™ื•ื“ืขื™ื ืžื” ื–ื” ื”ื™ื”,
14:35
but you use your model that you created
332
875880
1297
ืื‘ืœ ืืชื ืžืฉืชืžืฉื™ื ื‘ืžื•ื“ืœ ืฉื™ืฆืจืชื ื‘ืื™ืžื•ืŸ
14:37
in the training phase to figure out what they were typing.
333
877177
3442
ื›ื“ื™ ืœื“ืขืช ืžื” ื”ื•ื ื”ืงืœื™ื“.
14:40
They had pretty good success. This is an article from the USA Today.
334
880619
3484
ื•ื”ื™ื• ืœื”ื ืชื•ืฆืื•ืช ื“ื™ ื˜ื•ื‘ื•ืช. ื–ื” ืžืืžืจ ืžื”ืขื™ืชื•ืŸ USA Today.
14:44
They typed in, "The Illinois Supreme Court has ruled
335
884103
2609
ื”ืงืœื™ื“ื•, "ื‘ื™ืช ื”ืžืฉืคื˜ ื”ืขืœื™ื•ืŸ ืฉืœ ืื™ืœื™ื ื•ื™ ืคืกืง
14:46
that Rahm Emanuel is eligible to run for Mayor of Chicago"
336
886712
2962
ืฉืจื ืขืžื ื•ืืœ ื–ื›ืื™ ืœืจื•ืฅ ืœืจืืฉื•ืช ื”ืขื™ืจ ืฉืœ ืฉื™ืงื’ื•"
14:49
โ€” see, I tied it in to the last talk โ€”
337
889674
1354
- ืจื•ืื™ื, ืงื™ืฉืจืชื™ ืœืฉื™ื—ื” ื”ืงื•ื“ืžืช -
14:51
"and ordered him to stay on the ballot."
338
891028
2118
"ื•ื”ื•ืจื• ืœื• ืœื”ื™ืฉืืจ ื‘ืžื™ืจื•ืฅ."
14:53
Now, the system is interesting, because it produced
339
893146
2771
ืขื›ืฉื™ื•, ื”ืžืขืจื›ืช ืžืขื ื™ื™ื ืช, ืžืคื ื™ ืฉื”ื™ื ื™ืฆืจื”
14:55
"Illinois Supreme" and then it wasn't sure.
340
895917
2886
"ื”ืขืœื™ื•ืŸ ืฉืœ ืื™ืœื™ื ื•ื™" ื•ืื– ื”ื™ื ืœื ื”ื™ืชื” ื‘ื˜ื•ื—ื”.
14:58
The model produced a bunch of options,
341
898803
1950
ื”ืžื•ื“ืœ ืกื™ืคืง ืžืกืคืจ ืืคืฉืจื•ื™ื•ืช,
15:00
and this is the beauty of some of the A.I. techniques,
342
900753
2709
ื•ื–ื” ื”ื™ื•ืคื™ ืฉืœ ื›ืžื” ืžืฉื™ื˜ื•ืช ื”ืื™ื ื˜ื™ืœื™ื’ื ืฆื™ื” ื”ืžืœืื›ื•ืชื™ืช,
15:03
is that computers are good at some things,
343
903462
2250
ื–ื” ืฉืžื—ืฉื‘ื™ื ื˜ื•ื‘ื™ื ื‘ื—ืœืง ืžื”ื“ื‘ืจื™ื,
15:05
humans are good at other things,
344
905712
1534
ืื ืฉื™ื ื˜ื•ื‘ื™ื ื‘ื“ื‘ืจื™ื ืื—ืจื™ื,
15:07
take the best of both and let the humans solve this one.
345
907246
1931
ืงื—ื• ืืช ื”ื˜ื•ื‘ ืžืฉื ื™ื”ื ื•ืชื ื• ืœืื ืฉื™ื ืœืคืชื•ืจ ืืช ื–ื”.
15:09
Don't waste computer cycles.
346
909177
1382
ืืœ ืชื‘ื–ื‘ื–ื• ืžืฉืื‘ื™ ืžื—ืฉื‘.
15:10
A human's not going to think it's the Supreme might.
347
910559
2136
ืื“ื ืœื ื™ื—ืฉื•ื‘ ืฉื–ื” "ื”ื›ื•ื— ื”ืขืœื™ื•ืŸ".
15:12
It's the Supreme Court, right?
348
912695
1740
ื–ื” ื‘ื™ืช ื”ืžืฉืคื˜ ื”ืขืœื™ื•ืŸ, ื›ืŸ?
15:14
And so, together we're able to reproduce typing
349
914435
2530
ื•ื›ืš, ื™ื—ื“ ืื ื—ื ื• ืžืกื•ื’ืœื™ื ืœื™ืฆื•ืจ ืžื—ื“ืฉ ื”ืงืœื“ื•ืช
15:16
simply by measuring the accelerometer.
350
916965
2949
ืจืง ืขืœ ื™ื“ื™ ืฉื™ืžื•ืฉ ื‘ืžื“ ืชืื•ืฆื”.
15:19
Why does this matter? Well, in the Android platform,
351
919914
3502
ืœืžื” ื–ื” ืžืฉื ื”? ื•ื‘ื›ืŸ, ื‘ืคืœื˜ืคื•ืจืžืช ืื ื“ืจื•ืื™ื“,
15:23
for example, the developers have a manifest
352
923416
4133
ืœื“ื•ื’ืžื”, ืœืžืคืชื—ื™ื ื™ืฉ ืžื ื™ืคืกื˜
15:27
where every device on there, the microphone, etc.,
353
927564
2584
ืฉื‘ื• ื›ืœ ืจื›ื™ื‘, ื”ืžื™ืงืจื•ืคื•ืŸ, ื•ื›ื•',
15:30
has to register if you're going to use it
354
930148
1956
ืฆืจื™ืš ืœื”ื™ืจืฉื ืื ืืชื ืขื•ืžื“ื™ื ืœื”ืฉืชืžืฉ ื‘ื•
15:32
so that hackers can't take over it,
355
932104
2316
ื›ืš ืฉื”ืืงืจื™ื ืœื ื™ื›ื•ืœื™ื ืœื”ืฉืชืœื˜ ืขืœื™ื•,
15:34
but nobody controls the accelerometer.
356
934420
3108
ืื‘ืœ ืืฃ ืื—ื“ ืœื ืžืคืงื— ืขืœ ืžื“ ื”ืชืื•ืฆื”.
15:37
So what's the point? You can leave your iPhone next to
357
937528
2216
ืื– ืžื” ื”ื ืงื•ื“ื”? ืืชื ื™ื›ื•ืœื™ื ืœื”ืฉืื™ืจ ืืช ื”ืื™ื™ืคื•ืŸ
15:39
someone's keyboard, and just leave the room,
358
939744
2106
ืœื™ื“ ืžืงืœื“ืช ืฉืœ ืžื™ืฉื”ื•, ื•ืคืฉื•ื˜ ืœืฆืืช ืžื”ื—ื“ืจ,
15:41
and then later recover what they did,
359
941850
1639
ื•ืื—ืจื™ ื–ื” ืœืฉื—ื–ืจ ืžื” ื”ื ื›ืชื‘ื•,
15:43
even without using the microphone.
360
943489
1711
ืืคื™ืœื• ื‘ืœื™ ืฉื™ืžื•ืฉ ื‘ืžื™ืงืจื•ืคื•ืŸ.
15:45
If someone is able to put malware on your iPhone,
361
945200
2174
ืื ืžื™ืฉื”ื• ืžืกื•ื’ืœ ืœืฉื™ื ืจื•ื’ืœื” ืขืœ ื”ืื™ื™ืคื•ืŸ ืฉืœื›ื,
15:47
they could then maybe get the typing that you do
362
947374
2848
ื”ื ื™ื•ื›ืœื• ืื– ืื•ืœื™ ืœืงื‘ืœ ืืช ืžื” ืฉืืชื ืžืงืœื™ื“ื™ื
15:50
whenever you put your iPhone next to your keyboard.
363
950222
2321
ื‘ื›ืœ ืคืขื ืฉื”ืื™ื™ืคื•ืŸ ืฉืœื›ื ืœื™ื“ ื”ืžืงืœื“ืช.
15:52
There's several other notable attacks that unfortunately
364
952543
2271
ื™ืฉ ืขื•ื“ ื›ืžื” ืžืชืงืคื•ืช ื‘ื•ืœื˜ื•ืช ืื—ืจื•ืช ืฉืœืฆืขืจื ื•
15:54
I don't have time to go into, but the one that I wanted
365
954814
2131
ืื™ืŸ ืœื™ ื–ืžืŸ ืœื”ื›ื ืก ืืœื™ื”ืŸ, ืื‘ืœ ื–ื• ืฉืจืฆื™ืชื™ ืœืฆื™ื™ืŸ
15:56
to point out was a group from the University of Michigan
366
956945
2277
ื”ื™ื ื–ื• ืฉืœ ืงื‘ื•ืฆื” ืžืื•ื ื™ื‘ืจืกื™ื˜ืช ืžื™ืฉื™ื’ืŸ
15:59
which was able to take voting machines,
367
959222
2441
ืฉื”ืฆืœื™ื—ื” ืœืงื—ืช ืžื›ื•ื ื•ืช ื”ืฆื‘ืขื”,
16:01
the Sequoia AVC Edge DREs that
368
961663
2498
ื”ืกืงื•ื™ื” AVC ืื“ื’' DRE
16:04
were going to be used in New Jersey in the election
369
964161
1555
ืฉืขืžื“ื• ืœื”ืฉืชืžืฉ ื‘ื”ืŸ ื‘ื‘ื—ื™ืจื•ืช ื‘ื ื™ื• ื’'ืจื–ื™
16:05
that were left in a hallway, and put Pac-Man on it.
370
965716
2161
ื•ื”ื•ืฉืืจื• ื‘ืื™ื–ื” ืžืกื“ืจื•ืŸ, ื•ื”ื›ื ื™ืกื• ืœืชื•ื›ืŸ ืคืงืžืŸ.
16:07
So they ran the Pac-Man game.
371
967877
3623
ืื– ื”ืŸ ื”ืจื™ืฆื• ืืช ืžืฉื—ืง ื”ืคืงืžืŸ.
16:11
What does this all mean?
372
971500
1747
ืžื” ื–ื” ืื•ืžืจ?
16:13
Well, I think that society tends to adopt technology
373
973247
3647
ื•ื‘ื›ืŸ, ืื ื™ ื—ื•ืฉื‘ ืฉื”ื—ื‘ืจื” ื ื•ื˜ื” ืœืืžืฅ ื˜ื›ื ื•ืœื•ื’ื™ื•ืช
16:16
really quickly. I love the next coolest gadget.
374
976894
2824
ืžืžืฉ ื‘ืžื”ื™ืจื•ืช. ืื ื™ ืื•ื”ื‘ ืืช ื”ื’ื“ื’'ื˜ ื”ืžื’ื ื™ื‘ ื”ื‘ื.
16:19
But it's very important, and these researchers are showing,
375
979718
2614
ืื‘ืœ ืžืžืฉ ื—ืฉื•ื‘, ื•ื”ื—ื•ืงืจื™ื ื”ืืœื” ืžืจืื™ื,
16:22
that the developers of these things
376
982332
1360
ืฉื”ืžืคืชื—ื™ื ืฉืœ ื”ื“ื‘ืจื™ื ื”ืืœื”
16:23
need to take security into account from the very beginning,
377
983692
2865
ืฆืจื™ื›ื™ื ืœืงื—ืช ื‘ื—ืฉื‘ื•ืŸ ืืช ื”ืื‘ื˜ื—ื” ืžื”ื”ืชื—ืœื”,
16:26
and need to realize that they may have a threat model,
378
986557
2785
ื•ืฆืจื™ื›ื™ื ืœื”ื‘ื™ืŸ ืฉืื•ืœื™ ื™ืฉ ืœื”ื ืžื•ื“ืœ ืกื™ื›ื•ื ื™ื,
16:29
but the attackers may not be nice enough
379
989342
2462
ืื‘ืœ ื”ืชื•ืงืคื™ื ืื•ืœื™ ืœื ื™ื”ื™ื• ื ื—ืžื“ื™ื ืžืกืคื™ืง
16:31
to limit themselves to that threat model,
380
991804
1777
ืœื”ื’ื‘ื™ืœ ืืช ืขืฆืžื ืœืื•ืชื• ืžื•ื“ืœ ื”ืกื™ื›ื•ืŸ,
16:33
and so you need to think outside of the box.
381
993581
2537
ื•ืœื›ืŸ ื—ื™ื™ื‘ื™ื ืœื—ืฉื•ื‘ ืžื—ื•ืฅ ืœืงื•ืคืกื”.
16:36
What we can do is be aware
382
996118
1578
ืžื” ืฉืื ื—ื ื• ื™ื›ื•ืœื™ื ืœืขืฉื•ืช ื›ื“ื™ ืœื”ื™ื•ืช ืžื•ื“ืขื™ื
16:37
that devices can be compromised,
383
997696
2479
ืฉืžื›ืฉื™ืจื™ื ื™ื›ื•ืœื™ื ืœื”ื™ื•ืช ื‘ืกื™ื›ื•ืŸ,
16:40
and anything that has software in it
384
1000175
1699
ื•ืœื›ืœ ืžื” ืฉืžื›ื™ืœ ืชื•ื›ื ื”
16:41
is going to be vulnerable. It's going to have bugs.
385
1001874
2649
ื™ื”ื™ื• ื ืงื•ื“ื•ืช ืชื•ืจืคื”. ื•ื™ื”ื™ื• ืœื• ื‘ืื’ื™ื.
16:44
Thank you very much. (Applause)
386
1004523
3497
ืชื•ื“ื” ืจื‘ื” ืœื›ื. (ืžื—ื™ืื•ืช ื›ืคื™ื™ื)
ืขืœ ืืชืจ ื–ื”

ืืชืจ ื–ื” ื™ืฆื™ื’ ื‘ืคื ื™ื›ื ืกืจื˜ื•ื ื™ YouTube ื”ืžื•ืขื™ืœื™ื ืœืœื™ืžื•ื“ ืื ื’ืœื™ืช. ืชื•ื›ืœื• ืœืจืื•ืช ืฉื™ืขื•ืจื™ ืื ื’ืœื™ืช ื”ืžื•ืขื‘ืจื™ื ืขืœ ื™ื“ื™ ืžื•ืจื™ื ืžื”ืฉื•ืจื” ื”ืจืืฉื•ื ื” ืžืจื—ื‘ื™ ื”ืขื•ืœื. ืœื—ืฅ ืคืขืžื™ื™ื ืขืœ ื”ื›ืชื•ื‘ื™ื•ืช ื‘ืื ื’ืœื™ืช ื”ืžื•ืฆื’ื•ืช ื‘ื›ืœ ื“ืฃ ื•ื™ื“ืื• ื›ื“ื™ ืœื”ืคืขื™ืœ ืืช ื”ืกืจื˜ื•ืŸ ืžืฉื. ื”ื›ืชื•ื‘ื™ื•ืช ื’ื•ืœืœื•ืช ื‘ืกื ื›ืจื•ืŸ ืขื ื”ืคืขืœืช ื”ื•ื•ื™ื“ืื•. ืื ื™ืฉ ืœืš ื”ืขืจื•ืช ืื• ื‘ืงืฉื•ืช, ืื ื ืฆื•ืจ ืื™ืชื ื• ืงืฉืจ ื‘ืืžืฆืขื•ืช ื˜ื•ืคืก ื™ืฆื™ืจืช ืงืฉืจ ื–ื”.

https://forms.gle/WvT1wiN1qDtmnspy7