WEBVTT

1
00:00:01.320 --> 00:00:04.360
<v Speaker 1>Welcome back, everybody. Today, we're gonna be taking a different look.

2
00:00:04.599 --> 00:00:08.480
<v Speaker 1>We're gonna be looking at the park bench problem. Housepies

3
00:00:08.519 --> 00:00:11.240
<v Speaker 1>really meet their source. You see you in a hundred times.

4
00:00:11.240 --> 00:00:13.080
<v Speaker 1>A man walks into a park sits on a bench.

5
00:00:13.119 --> 00:00:15.439
<v Speaker 1>A second person appears a few minutes later and sits

6
00:00:15.439 --> 00:00:18.519
<v Speaker 1>beside him. Neither looks at the other. They stare straight

7
00:00:18.559 --> 00:00:23.039
<v Speaker 1>ahead while quietly exchanging classified information. Because my favorite part,

8
00:00:23.440 --> 00:00:25.800
<v Speaker 1>one of them casually slides a envelope towards the other.

9
00:00:26.800 --> 00:00:29.719
<v Speaker 1>The circ second person takes it, stands up, and walks away,

10
00:00:29.800 --> 00:00:33.759
<v Speaker 1>and apparently nobody notices. Watching home man I saw Sall

11
00:00:33.840 --> 00:00:37.280
<v Speaker 1>Bearnson engage in variations of this classic spy movie behavior.

12
00:00:38.960 --> 00:00:42.000
<v Speaker 1>You might be asking yourself, well, this is actually doesn't

13
00:00:42.039 --> 00:00:45.679
<v Speaker 1>this actually look more suspicious if two people legitimately knew

14
00:00:45.719 --> 00:00:48.280
<v Speaker 1>each other? Wouldn't it look more natural if they were talking,

15
00:00:48.359 --> 00:00:52.079
<v Speaker 1>or drinking coffee or laughing? The answer introduces us to

16
00:00:52.119 --> 00:00:56.000
<v Speaker 1>a fascinating problem in real world espionage, what I would

17
00:00:56.039 --> 00:00:59.439
<v Speaker 1>call the park bench problem. Real intel officers aren't simply

18
00:00:59.479 --> 00:01:02.359
<v Speaker 1>trying to looking suspicious. They also try to avoid creating

19
00:01:02.359 --> 00:01:06.359
<v Speaker 1>evidence of a relationship. Those aren't necessarily the same thing.

20
00:01:06.519 --> 00:01:10.159
<v Speaker 1>So back through Saul Berenston, imagine he was actually talking

21
00:01:10.200 --> 00:01:12.799
<v Speaker 1>to a Russian intelligence officer, and imagine what that would

22
00:01:12.799 --> 00:01:17.079
<v Speaker 1>look like. Secretly meeting an American intelligence officer. A random

23
00:01:17.079 --> 00:01:20.120
<v Speaker 1>person walking through the park probably doesn't matter. But if

24
00:01:20.599 --> 00:01:24.799
<v Speaker 1>Russian or American counterintelligence is already following you, now everything changes.

25
00:01:25.280 --> 00:01:28.599
<v Speaker 1>They're documenting where you go, when you arrive, who appears long,

26
00:01:28.760 --> 00:01:30.959
<v Speaker 1>how long they remain, where they go afterward, and where

27
00:01:31.000 --> 00:01:34.760
<v Speaker 1>the same person appears again next week. Suddenly having coffee

28
00:01:34.799 --> 00:01:38.799
<v Speaker 1>and laughing together isn't necessarily better either. You've just demonstrated

29
00:01:38.840 --> 00:01:41.760
<v Speaker 1>that you know each other, and that brings us to

30
00:01:41.840 --> 00:01:47.000
<v Speaker 1>our first intelligence concept, the pattern of association. Intelligence services

31
00:01:47.000 --> 00:01:50.280
<v Speaker 1>don't always need to hear what two people are saying. Sometimes,

32
00:01:50.319 --> 00:01:53.760
<v Speaker 1>simply establishing the person A repeatedly encounter's person B is

33
00:01:53.799 --> 00:01:58.000
<v Speaker 1>extremely valuable. If a CAA officer station under diplomatic cover

34
00:01:58.120 --> 00:02:02.239
<v Speaker 1>repeatedly appears near an employee of the Russian government, counterintelligence

35
00:02:02.280 --> 00:02:06.359
<v Speaker 1>doesn't necessarily need a recording of their conversation. The relationship

36
00:02:06.400 --> 00:02:10.759
<v Speaker 1>itself becomes interesting. So intel officers face a paradox. You

37
00:02:10.800 --> 00:02:13.680
<v Speaker 1>need to communicate with your source, but communicating creates evidence

38
00:02:13.680 --> 00:02:18.599
<v Speaker 1>that the relationship exists. Another trade care AFT concept does

39
00:02:18.639 --> 00:02:21.319
<v Speaker 1>cover for action, which means having an innocent explanation for

40
00:02:21.400 --> 00:02:25.560
<v Speaker 1>what you're doing. Imagine again, the intelligence officer repeatedly entering

41
00:02:25.560 --> 00:02:28.159
<v Speaker 1>the same cafe not suspicious if there's no reason for

42
00:02:28.240 --> 00:02:31.719
<v Speaker 1>him to be there. But suppose the cafe is across

43
00:02:31.759 --> 00:02:34.039
<v Speaker 1>the street from his office. If he buys coffee every morning,

44
00:02:34.039 --> 00:02:37.840
<v Speaker 1>Now his presence has a perfectly reasonable explanation. The best

45
00:02:37.840 --> 00:02:40.840
<v Speaker 1>clandestine behavior often isn't something that looks secret, it's something

46
00:02:40.840 --> 00:02:45.479
<v Speaker 1>that looks boring. However, another problem. Before meeting a sensitive

47
00:02:45.479 --> 00:02:47.840
<v Speaker 1>source and an intel officer may need to determine whether

48
00:02:47.879 --> 00:02:52.039
<v Speaker 1>someone is following him. This introduces the surveillance detection route,

49
00:02:52.319 --> 00:02:56.680
<v Speaker 1>sometimes abbreviated the SDR. The basic concept to straightforward. Before

50
00:02:56.719 --> 00:03:00.439
<v Speaker 1>conducting an important clandestine activity, the officer must move through

51
00:03:00.599 --> 00:03:03.919
<v Speaker 1>environment in ways that provide opportunities to determine whether surveillance

52
00:03:03.960 --> 00:03:07.639
<v Speaker 1>is present. That doesn't mean dramatically looking over your shoulder

53
00:03:07.639 --> 00:03:10.680
<v Speaker 1>every fifteen seconds. In fact, doing that could advertise that

54
00:03:10.759 --> 00:03:14.800
<v Speaker 1>your surveillance conscious. Instead, the objective is to identify whether

55
00:03:14.840 --> 00:03:19.520
<v Speaker 1>the same people, vehicles, or surveillance elements continue appearing during

56
00:03:19.560 --> 00:03:24.919
<v Speaker 1>otherwise explainable movements and sophisticated surveillance teams make this extremely difficult.

57
00:03:25.319 --> 00:03:28.800
<v Speaker 1>Professional surveillance isn't necessarily one guy in sunglasses following you

58
00:03:28.840 --> 00:03:33.280
<v Speaker 1>for six blocks. Teams can rotate, personnel, vehicles can change

59
00:03:33.280 --> 00:03:36.400
<v Speaker 1>Observers can position themselves ahead of the target, and that's

60
00:03:36.439 --> 00:03:39.800
<v Speaker 1>why spy movies may sometimes underestimate the difficulty of determining

61
00:03:39.840 --> 00:03:42.599
<v Speaker 1>where they're actually clean. Now, suppose you don't want to

62
00:03:42.680 --> 00:03:45.120
<v Speaker 1>risk a meeting at all either. One of espionage's oldest

63
00:03:45.159 --> 00:03:48.599
<v Speaker 1>solutions the dead drop, which allows information and material to

64
00:03:48.599 --> 00:03:51.719
<v Speaker 1>be transferred without the two participants meeting, just leaving something

65
00:03:51.759 --> 00:03:56.360
<v Speaker 1>got a predetermined location and later retrieving it huge advantage

66
00:03:56.360 --> 00:04:00.759
<v Speaker 1>for the intelligence. Cara intelligence identifies the sun savenants can

67
00:04:00.800 --> 00:04:05.360
<v Speaker 1>potentially reveal who services it. Another technique is the brush past.

68
00:04:05.400 --> 00:04:07.639
<v Speaker 1>The idea is a very brief encounter during which material

69
00:04:07.719 --> 00:04:11.199
<v Speaker 1>changes hands while both individuals continue moving through an ordinary environment.

70
00:04:11.840 --> 00:04:14.240
<v Speaker 1>The attraction is obvious instead of sitting together for twenty

71
00:04:14.319 --> 00:04:19.040
<v Speaker 1>minutes to contact last seconds. But again, professional surveillance changes everything.

72
00:04:19.079 --> 00:04:21.959
<v Speaker 1>If one participant is already being closely watched, even a brief,

73
00:04:22.040 --> 00:04:26.519
<v Speaker 1>unexplained interaction may become significant, which brings us back to

74
00:04:26.600 --> 00:04:31.160
<v Speaker 1>Paul to Saul sitting on the bench. Is it completely ridiculous, No,

75
00:04:31.160 --> 00:04:33.800
<v Speaker 1>there's some logic behind it. If two people sit near

76
00:04:33.839 --> 00:04:37.360
<v Speaker 1>one another without obviously acknowledging each other, a casual observer

77
00:04:37.480 --> 00:04:41.560
<v Speaker 1>might not ever realize they're communicating. The distinction is causal

78
00:04:41.720 --> 00:04:45.319
<v Speaker 1>observation and professional surveillance are completely different things. The stranger

79
00:04:45.360 --> 00:04:47.759
<v Speaker 1>walking his dog may notice nothing, but a trained surveillance

80
00:04:47.800 --> 00:04:53.079
<v Speaker 1>team already interested in Saul may notice everything. From a

81
00:04:53.120 --> 00:04:57.920
<v Speaker 1>filmmaking perspective, it's brilliant. The audience immediately understands secret information

82
00:04:57.959 --> 00:05:01.519
<v Speaker 1>has changed hands. From a counter intelligence perspective, you just

83
00:05:01.560 --> 00:05:06.720
<v Speaker 1>create a potentially observable physical transaction. And modern technology makes

84
00:05:06.720 --> 00:05:10.560
<v Speaker 1>the problem even greater. With high resolution cameras, face recognition,

85
00:05:11.120 --> 00:05:16.120
<v Speaker 1>license plate information also and communication metadata can establish patterns,

86
00:05:17.160 --> 00:05:20.079
<v Speaker 1>so a real time detection and retrospective detection. This is

87
00:05:20.120 --> 00:05:24.199
<v Speaker 1>an important distinction. Maybe nobody realizes today that your park

88
00:05:24.240 --> 00:05:27.959
<v Speaker 1>bench encounter mattered, but six months later investigators discovered that

89
00:05:28.000 --> 00:05:30.720
<v Speaker 1>the person's sitting byside you was a Russian intel officer.

90
00:05:31.319 --> 00:05:34.079
<v Speaker 1>Now they can go backward. Where else were you both present?

91
00:05:34.120 --> 00:05:37.360
<v Speaker 1>Did your phones appear in similar locations? Did either person's

92
00:05:37.360 --> 00:05:41.879
<v Speaker 1>behavior change after the encounter. Did sensitive information disappear? Suddenly

93
00:05:41.920 --> 00:05:44.680
<v Speaker 1>that harmless park bench becomes evidence. And that's the fundamental

94
00:05:44.720 --> 00:05:48.639
<v Speaker 1>problem of clan destined communication. Every contact creates potential information,

95
00:05:49.639 --> 00:05:53.879
<v Speaker 1>and meeting creates a location. Phone call creates metadata physically change,

96
00:05:53.920 --> 00:05:57.759
<v Speaker 1>it creates an observable event. It's the meeting that produces

97
00:05:57.759 --> 00:06:00.959
<v Speaker 1>It's the meeting that produces no meaningful intelligen for the adversary.

98
00:06:01.439 --> 00:06:04.439
<v Speaker 1>Sometimes that means creating an innocent explanation for being together.

99
00:06:04.959 --> 00:06:09.399
<v Speaker 1>Sometimes it means minimizing contact. Sometimes it means communicating in directly.

100
00:06:10.959 --> 00:06:14.000
<v Speaker 1>That's why the classic Hollywood park bench isn't completely fictional,

101
00:06:14.040 --> 00:06:17.800
<v Speaker 1>It's simply and complete. Real espionage isn't about wearing sunglasses

102
00:06:17.800 --> 00:06:20.360
<v Speaker 1>and staring straight ahead and whispering the package has arrived.

103
00:06:20.720 --> 00:06:24.079
<v Speaker 1>It's about understanding something much more difficult. Your adversary doesn't

104
00:06:24.079 --> 00:06:26.680
<v Speaker 1>necessarily need to hear what you said that. You only

105
00:06:26.720 --> 00:06:29.000
<v Speaker 1>need to discover who you were sitting next to.