1
00:00:02,840 --> 00:00:05,680
Speaker 1: Hi, it's West Kasova. We're taking a break this week

2
00:00:05,680 --> 00:00:08,799
Speaker 1: for the holidays, so here's an episode you might have

3
00:00:08,920 --> 00:00:16,759
Speaker 1: missed by now. We all know Google and other tech

4
00:00:16,800 --> 00:00:20,480
Speaker 1: companies are collecting huge amounts of information about our digital

5
00:00:20,640 --> 00:00:24,759
Speaker 1: and actual lives, what we search for, what we watch

6
00:00:24,840 --> 00:00:28,920
Speaker 1: and listen to, and where we are at any given time.

7
00:00:29,800 --> 00:00:32,640
Speaker 1: In large part that's so advertisers who pay for that

8
00:00:32,800 --> 00:00:36,680
Speaker 1: data can pester us to buy things. But Bloomberg's Davy

9
00:00:36,720 --> 00:00:40,480
Speaker 1: Alba and Julia Love report that this vastrophe of personal

10
00:00:40,479 --> 00:00:44,880
Speaker 1: information about millions of people is now increasingly also being

11
00:00:44,920 --> 00:00:49,839
Speaker 1: sought by police departments across the US. They're asking judges

12
00:00:49,880 --> 00:00:53,680
Speaker 1: for warrants to compel Google to turn over location and

13
00:00:53,840 --> 00:00:55,000
Speaker 1: other user data.

14
00:00:55,480 --> 00:01:00,240
Speaker 2: Google says that it receives more than sixty thousand and

15
00:01:00,320 --> 00:01:04,120
Speaker 2: search warrants, and this was true in the US last year.

16
00:01:04,560 --> 00:01:07,480
Speaker 2: That's more than double the number from twenty nineteen.

17
00:01:08,319 --> 00:01:12,680
Speaker 3: I think these requests put Google in a difficult spot

18
00:01:12,920 --> 00:01:16,959
Speaker 3: because these are court orders. They're not just optional.

19
00:01:17,760 --> 00:01:22,280
Speaker 1: And later, Mesa, Arizona detective Travis Stobb tells us how

20
00:01:22,319 --> 00:01:23,759
Speaker 1: police use this information.

21
00:01:24,680 --> 00:01:28,200
Speaker 4: It's another tool on our toolkit, and it's a welcome one.

22
00:01:28,319 --> 00:01:30,280
Speaker 5: It just gives us another avenue.

23
00:01:29,880 --> 00:01:33,040
Speaker 4: To turn to when we're coming up dry on more

24
00:01:33,080 --> 00:01:34,920
Speaker 4: traditional investigative techniques.

25
00:01:40,680 --> 00:01:44,319
Speaker 1: I'm Wescosova today on the Big Take Cops cast a

26
00:01:44,520 --> 00:01:58,560
Speaker 1: Digital Dragnet. I asked Davy why she and Julia decided

27
00:01:58,600 --> 00:02:01,920
Speaker 1: to look into how police departments are using all this

28
00:02:02,080 --> 00:02:04,720
Speaker 1: information that Google collects about us.

29
00:02:05,600 --> 00:02:10,600
Speaker 2: I think we've known that Google is collecting an enormous

30
00:02:10,600 --> 00:02:13,960
Speaker 2: amount of data on each and every one of us. Basically,

31
00:02:13,960 --> 00:02:17,240
Speaker 2: if you have a Google account, you are handing over

32
00:02:17,880 --> 00:02:21,840
Speaker 2: data on where you've been, what you've searched, and Google

33
00:02:22,520 --> 00:02:26,079
Speaker 2: slurps up all this data to use in a lot

34
00:02:26,080 --> 00:02:29,960
Speaker 2: of cases for advertising. So you have all of this

35
00:02:30,240 --> 00:02:34,560
Speaker 2: history that you're keeping with Google that's in your account,

36
00:02:35,240 --> 00:02:38,240
Speaker 2: and it just so happens that cops are also interested

37
00:02:38,280 --> 00:02:40,799
Speaker 2: in that when they are trying to solve a crime.

38
00:02:41,520 --> 00:02:47,919
Speaker 2: So we suspected that there was this match between all

39
00:02:48,000 --> 00:02:51,120
Speaker 2: the data that we were giving Google and users are

40
00:02:51,120 --> 00:02:54,880
Speaker 2: giving Google, and the way that the cops try to

41
00:02:54,919 --> 00:02:58,800
Speaker 2: mind that data to solve these crimes that have taken place,

42
00:02:59,560 --> 00:03:03,920
Speaker 2: And we wanted to know whether cops were using this

43
00:03:04,200 --> 00:03:09,680
Speaker 2: for very serious crimes or if they were also interested

44
00:03:09,760 --> 00:03:16,880
Speaker 2: in this data for petty ones theft, graffiti, vandalism, And

45
00:03:17,000 --> 00:03:21,000
Speaker 2: what we found was that this practice was incredibly widespread

46
00:03:21,320 --> 00:03:24,720
Speaker 2: all over the US. Tiny police departments all over the

47
00:03:24,760 --> 00:03:27,800
Speaker 2: country are using Google data to try to solve crimes,

48
00:03:28,480 --> 00:03:33,000
Speaker 2: and the crimes weren't always necessarily serious ones. They did

49
00:03:33,040 --> 00:03:36,440
Speaker 2: include some of these more minor offenses as well.

50
00:03:38,360 --> 00:03:39,840
Speaker 1: And Julia, how'd you find this out?

51
00:03:40,480 --> 00:03:45,520
Speaker 3: If cops want to dip into Doodle's data, they have

52
00:03:45,640 --> 00:03:49,560
Speaker 3: to obtain a search warrant from a judge. The company

53
00:03:49,720 --> 00:03:53,840
Speaker 3: will not simply open up its files for any law

54
00:03:53,920 --> 00:03:58,080
Speaker 3: enforcement officer who comes knocking. They have to have a

55
00:03:58,160 --> 00:04:01,800
Speaker 3: valid court order, and so in order to do this,

56
00:04:02,720 --> 00:04:06,200
Speaker 3: police officers will write up an affi David that it

57
00:04:06,200 --> 00:04:10,440
Speaker 3: explains the trime that occurred and why they feel that

58
00:04:10,600 --> 00:04:15,440
Speaker 3: Goodle might have information that could be helpful in building

59
00:04:15,480 --> 00:04:20,160
Speaker 3: out the investigation. So we sought to get our hands

60
00:04:20,600 --> 00:04:24,360
Speaker 3: on those search warrants. We knew that was the best

61
00:04:24,360 --> 00:04:28,120
Speaker 3: way that we could see what cops were actually requesting.

62
00:04:28,880 --> 00:04:33,039
Speaker 2: Unfortunately, for US, search warrants are public records, so we

63
00:04:33,120 --> 00:04:37,800
Speaker 2: could actually visit the courthouses and collect them for our

64
00:04:37,839 --> 00:04:38,880
Speaker 2: reporting purposes.

65
00:04:40,160 --> 00:04:44,240
Speaker 1: Julia and a police department gets a warrant for Google

66
00:04:44,320 --> 00:04:46,880
Speaker 1: to give the information. What exactly are they asking for?

67
00:04:46,960 --> 00:04:49,640
Speaker 1: What kind of information do they want Google to give?

68
00:04:50,480 --> 00:04:53,920
Speaker 3: In the case of location warrants, this is all about

69
00:04:54,040 --> 00:04:58,560
Speaker 3: where a specific device was at at a particular moment

70
00:04:58,800 --> 00:05:02,880
Speaker 3: in time. Little is able to pull in information from

71
00:05:03,480 --> 00:05:09,040
Speaker 3: why Fi hotspots, the phones, GPS, sell tower connections, and

72
00:05:09,080 --> 00:05:13,159
Speaker 3: that gives them a pretty good ability to estimate where

73
00:05:13,200 --> 00:05:18,200
Speaker 3: a particular device was at a given time within several

74
00:05:18,240 --> 00:05:19,640
Speaker 3: feet actually.

75
00:05:20,400 --> 00:05:22,800
Speaker 1: And so Davey, it seems like they're casting a pretty

76
00:05:22,920 --> 00:05:25,320
Speaker 1: wide net here they are.

77
00:05:26,360 --> 00:05:29,520
Speaker 2: The way it works is the police will give let's

78
00:05:29,560 --> 00:05:34,560
Speaker 2: say four coordinates. These are just x y coordinates on

79
00:05:34,640 --> 00:05:38,480
Speaker 2: a map, and then with these four points, they'll draw

80
00:05:38,640 --> 00:05:43,880
Speaker 2: a shape around a certain area and we'll say, okay, Google,

81
00:05:44,000 --> 00:05:46,680
Speaker 2: give me all of the devices that have been in

82
00:05:46,760 --> 00:05:51,960
Speaker 2: this polygon over a certain amount of time. So some

83
00:05:52,040 --> 00:05:55,800
Speaker 2: of the warrants we've seen have covered just minutes. So

84
00:05:56,680 --> 00:05:59,719
Speaker 2: when the police know when a crime has occurred, they

85
00:05:59,720 --> 00:06:03,240
Speaker 2: can say, okay, within a five minute period, tell us

86
00:06:03,240 --> 00:06:06,200
Speaker 2: who was in this area. But we've also seen law

87
00:06:06,279 --> 00:06:11,000
Speaker 2: enforcement agencies going to Google asking them for how many

88
00:06:11,000 --> 00:06:14,520
Speaker 2: devices were in this area for up to a week.

89
00:06:15,480 --> 00:06:18,880
Speaker 1: Julia, can you give us some examples of times where

90
00:06:19,000 --> 00:06:22,920
Speaker 1: police departments went to Google and said, we want this information.

91
00:06:23,720 --> 00:06:29,200
Speaker 3: In one case, a woman's purse was stolen in the

92
00:06:29,240 --> 00:06:34,320
Speaker 3: parking lot of a target in Scottsdale, Arizona. In that case,

93
00:06:34,800 --> 00:06:39,880
Speaker 3: the police sought a warrant for Doodle's location data information

94
00:06:40,080 --> 00:06:43,159
Speaker 3: for people who were in that parking lot at the

95
00:06:43,240 --> 00:06:48,880
Speaker 3: time that the woman's purse was drabbed. In another case

96
00:06:49,040 --> 00:06:55,760
Speaker 3: that we found out of Raleigh, North Carolina, a detective

97
00:06:56,160 --> 00:06:59,880
Speaker 3: was warming up his car in the morning and he

98
00:07:00,600 --> 00:07:04,920
Speaker 3: ran into his apartment to get something, leaving the car running,

99
00:07:05,520 --> 00:07:09,800
Speaker 3: and when he came back, the vehicle had vanished with

100
00:07:09,920 --> 00:07:15,440
Speaker 3: some police dear inside, including a police radio. And so

101
00:07:15,960 --> 00:07:21,320
Speaker 3: the Raleigh Police Department obtained one warrant for Doodle's location

102
00:07:21,600 --> 00:07:24,880
Speaker 3: data to see who had been in the neighborhood at

103
00:07:24,880 --> 00:07:26,000
Speaker 3: the time of the trime.

104
00:07:26,880 --> 00:07:29,880
Speaker 1: And Julia, what did the Raleigh Police Department say about

105
00:07:29,880 --> 00:07:30,320
Speaker 1: the case?

106
00:07:30,920 --> 00:07:33,960
Speaker 3: The Raleigh Police Department said that they ultimately did not

107
00:07:34,120 --> 00:07:37,880
Speaker 3: recover that radio that they were seeking to find. They

108
00:07:37,920 --> 00:07:40,680
Speaker 3: also said they always strive to recover stolen property.

109
00:07:43,120 --> 00:07:46,760
Speaker 1: Dave, even though we all have Google accounts. These warrants

110
00:07:46,800 --> 00:07:48,880
Speaker 1: don't cover you know, tens of millions of people. They

111
00:07:48,920 --> 00:07:53,160
Speaker 1: only cover whatever devices may have been in a particular

112
00:07:53,200 --> 00:07:56,360
Speaker 1: area at a particular time around a crime scene. Is

113
00:07:56,360 --> 00:07:56,720
Speaker 1: that right?

114
00:07:57,840 --> 00:08:02,040
Speaker 2: Yes, that is correct. But the pool of people that

115
00:08:02,120 --> 00:08:05,080
Speaker 2: the cops in Google are drawing this data from is

116
00:08:05,400 --> 00:08:09,680
Speaker 2: a huge pool, and that can be, you know, incredibly invasive.

117
00:08:10,200 --> 00:08:14,160
Speaker 2: Your stuff can get ensnared in these warrants, Julia.

118
00:08:14,240 --> 00:08:17,920
Speaker 1: So Google gets all of these warrants telling them they

119
00:08:17,920 --> 00:08:21,160
Speaker 1: have to give up information. How does the company deal

120
00:08:21,200 --> 00:08:24,160
Speaker 1: with this? What's their position about all of these requestsing

121
00:08:24,280 --> 00:08:27,520
Speaker 1: you have to give up the user data of your customers.

122
00:08:28,360 --> 00:08:33,360
Speaker 3: I think these requests put Goodle in a difficult spot because,

123
00:08:33,600 --> 00:08:37,160
Speaker 3: on the one hand, these are court orders. They're not

124
00:08:37,400 --> 00:08:41,880
Speaker 3: just optional, as a former Doodle leadal director told us.

125
00:08:42,559 --> 00:08:46,319
Speaker 3: But at the same time, the company really values safeguarding

126
00:08:46,480 --> 00:08:49,520
Speaker 3: its users privacy, and so it has to watch this

127
00:08:49,679 --> 00:08:54,640
Speaker 3: line between helping law enforcement with their investigations and also

128
00:08:55,160 --> 00:08:58,920
Speaker 3: not letting any detective rifle through any user's data.

129
00:09:00,000 --> 00:09:05,680
Speaker 1: Google gets a warrant to hand over some information, what

130
00:09:05,800 --> 00:09:06,280
Speaker 1: do they do?

131
00:09:06,920 --> 00:09:12,320
Speaker 3: Goodle receives thousands of requests from law enforcement in the

132
00:09:12,440 --> 00:09:16,080
Speaker 3: United States each year, and so they've come up with

133
00:09:16,200 --> 00:09:21,200
Speaker 3: some formulas to respond to the common requests to give

134
00:09:21,320 --> 00:09:26,360
Speaker 3: their legal specialists a roadmap to follow. Law enforcement will

135
00:09:26,480 --> 00:09:30,920
Speaker 3: request information about devices that were present near the time

136
00:09:30,960 --> 00:09:34,680
Speaker 3: of the crime, and Goodle will simply provide a list

137
00:09:34,880 --> 00:09:40,920
Speaker 3: of anonymous device ideas that do not reveal which person

138
00:09:41,400 --> 00:09:46,600
Speaker 3: was tied to that device. Law enforcement will carefully go

139
00:09:46,760 --> 00:09:51,680
Speaker 3: over this list and they'll identify a subset of devices

140
00:09:51,920 --> 00:09:56,720
Speaker 3: that they would like expanded information about to see where

141
00:09:56,760 --> 00:10:01,440
Speaker 3: the device was before and a after the time of

142
00:10:01,480 --> 00:10:05,600
Speaker 3: the trime to help them decide who they really want

143
00:10:05,640 --> 00:10:10,920
Speaker 3: to focus on. And Google will provide account information that's

144
00:10:11,000 --> 00:10:14,520
Speaker 3: usually the name that the account was registered under and

145
00:10:14,600 --> 00:10:18,480
Speaker 3: an email address for that final batch of users.

146
00:10:20,679 --> 00:10:24,440
Speaker 1: Davy Julius said earlier that Google gets thousands of these

147
00:10:24,720 --> 00:10:28,760
Speaker 1: requests for information a year. Just how many do they get?

148
00:10:29,080 --> 00:10:33,840
Speaker 2: So Google says that it receives more than sixty thousand

149
00:10:34,080 --> 00:10:38,040
Speaker 2: search warrants and this was true in the US last year.

150
00:10:38,679 --> 00:10:42,440
Speaker 2: That's more than double the number from twenty nineteen. And

151
00:10:42,520 --> 00:10:46,480
Speaker 2: their company provides at least some information in about eighty

152
00:10:46,520 --> 00:10:47,679
Speaker 2: percent of cases.

153
00:10:48,440 --> 00:10:51,319
Speaker 1: You said, there's this group of people at Google whose

154
00:10:51,400 --> 00:10:53,320
Speaker 1: job it is to do this? Who are they?

155
00:10:53,920 --> 00:10:57,360
Speaker 2: It's a team known as the Legal Investigation Support Team

156
00:10:57,559 --> 00:11:01,480
Speaker 2: or LIS. Their whole job is through view search warrants

157
00:11:01,520 --> 00:11:05,840
Speaker 2: and subpoenas, and they try to push back on as

158
00:11:05,840 --> 00:11:09,640
Speaker 2: many as they can using what they have at their disposal.

159
00:11:09,880 --> 00:11:14,400
Speaker 2: So let's say someone misspelled a Gmail account and it's

160
00:11:14,920 --> 00:11:20,160
Speaker 2: gamil dot com instead of Gmail dot com. Once these

161
00:11:20,360 --> 00:11:24,800
Speaker 2: warrants are served to Google, Google does have an obligation

162
00:11:25,080 --> 00:11:30,760
Speaker 2: to respond, a legal obligation, And so when they go

163
00:11:30,840 --> 00:11:33,920
Speaker 2: through these warrants and try to sort through them and

164
00:11:34,000 --> 00:11:37,439
Speaker 2: scrutinize them, they are trying to make sure they're doing

165
00:11:37,480 --> 00:11:41,520
Speaker 2: it just as precisely as possible, you know, to the

166
00:11:41,600 --> 00:11:45,600
Speaker 2: letter of the law, to make sure that they can

167
00:11:46,280 --> 00:11:50,680
Speaker 2: only give up people's information when it is absolutely necessary.

168
00:11:50,520 --> 00:11:53,160
Speaker 3: And I would say also the minimum amount of information

169
00:11:53,400 --> 00:11:56,280
Speaker 3: necessary to comply with this request.

170
00:11:56,920 --> 00:12:00,760
Speaker 1: So Google is only one of many services are tracking

171
00:12:00,840 --> 00:12:03,440
Speaker 1: us and gathering up our information. What about Apple and

172
00:12:03,520 --> 00:12:06,760
Speaker 1: Apple Maps, which also helps you get around and tracks

173
00:12:06,800 --> 00:12:09,160
Speaker 1: your location. Do they also give up this kind of

174
00:12:09,160 --> 00:12:11,280
Speaker 1: information in response to warrants?

175
00:12:11,800 --> 00:12:14,920
Speaker 3: Law enforcement experts have told us that Google is the

176
00:12:14,920 --> 00:12:18,800
Speaker 3: only company that can provide this detailed list of which

177
00:12:18,840 --> 00:12:23,000
Speaker 3: devices were present at a diven time, and Apple in

178
00:12:23,040 --> 00:12:29,439
Speaker 3: its transparency reports discloses that it does receive some geofence warrants,

179
00:12:29,520 --> 00:12:32,720
Speaker 3: as these location warrants are known, but they say that

180
00:12:32,800 --> 00:12:36,360
Speaker 3: they are not capable of complying with those warrants.

181
00:12:36,880 --> 00:12:39,960
Speaker 2: But it doesn't actually matter because a lot of people

182
00:12:40,040 --> 00:12:43,760
Speaker 2: who have iPhones have Google apps on them. So if

183
00:12:43,800 --> 00:12:46,200
Speaker 2: you think about I have to get to a friend's

184
00:12:46,200 --> 00:12:49,440
Speaker 2: birthday party and how do I get there, I'll fire

185
00:12:49,520 --> 00:12:52,959
Speaker 2: up Google Maps and see the best subway route, and

186
00:12:53,000 --> 00:12:58,640
Speaker 2: that information is something that Google can see and does have.

187
00:12:59,640 --> 00:13:03,360
Speaker 1: Google sent Bloomberg a statement in response to questions about

188
00:13:03,360 --> 00:13:06,960
Speaker 1: these warrants about user data. They said with all law

189
00:13:07,080 --> 00:13:10,920
Speaker 1: enforcement demands, including reverse warrants, they have a rigorous process

190
00:13:11,000 --> 00:13:14,640
Speaker 1: designed to protect the privacy of their users while supporting

191
00:13:14,640 --> 00:13:18,080
Speaker 1: the important work of law enforcement. They said they examine

192
00:13:18,120 --> 00:13:22,040
Speaker 1: each demand for legal validity consistent with developing case law,

193
00:13:22,240 --> 00:13:26,199
Speaker 1: and they routinely push back against overbroad or otherwise inappropriate

194
00:13:26,240 --> 00:13:30,640
Speaker 1: demands for user data, including objecting to some demands entirely.

195
00:13:31,280 --> 00:13:34,520
Speaker 1: They said for years they've supported legislative reform to government

196
00:13:34,600 --> 00:13:40,160
Speaker 1: practices around access to user data, including proposing substantive guardrails

197
00:13:40,160 --> 00:13:43,800
Speaker 1: on access to data and limits on the overuse of

198
00:13:43,840 --> 00:13:48,040
Speaker 1: GAG orders, which they say contribute to a lack of transparency.

199
00:13:49,559 --> 00:13:52,600
Speaker 1: I'll talk more with Julia and Davy a little later.

200
00:13:52,960 --> 00:13:56,400
Speaker 1: After the break, A detective tells us how police use

201
00:13:56,480 --> 00:14:07,960
Speaker 1: the data Google hands over. Now, let's get a look

202
00:14:08,000 --> 00:14:11,480
Speaker 1: at how law enforcement uses this information in real life.

203
00:14:11,840 --> 00:14:15,199
Speaker 1: Detective Travis Stobb is an officer with the Mesa, Arizona

204
00:14:15,240 --> 00:14:20,360
Speaker 1: Police Department, which has sought search warrants for Google data. Travis,

205
00:14:20,400 --> 00:14:22,080
Speaker 1: can you tell me what it's like to be a

206
00:14:22,120 --> 00:14:25,880
Speaker 1: detective working a case, especially when you're running out of

207
00:14:25,920 --> 00:14:29,200
Speaker 1: avenues to pursue and you just can't quite find the

208
00:14:29,240 --> 00:14:30,520
Speaker 1: person who committed the crime.

209
00:14:31,280 --> 00:14:35,640
Speaker 4: Well, every case is different, obviously, and there comes with

210
00:14:35,680 --> 00:14:38,520
Speaker 4: it a certain frustration when you're not able to get

211
00:14:38,560 --> 00:14:41,000
Speaker 4: a break in the case, and you feel a lot

212
00:14:41,000 --> 00:14:45,440
Speaker 4: of pressure both externally and internally, especially in these instances

213
00:14:45,520 --> 00:14:48,680
Speaker 4: where you have someone who continues to engage in an

214
00:14:48,800 --> 00:14:51,720
Speaker 4: armed robbery after armed robbery and you want to get

215
00:14:51,760 --> 00:14:56,000
Speaker 4: them caught before someone gets hurt, or just stop them

216
00:14:56,000 --> 00:14:59,600
Speaker 4: from doing what they're doing. And so the ability to

217
00:15:00,240 --> 00:15:02,640
Speaker 4: leverage this Google data was a.

218
00:15:02,560 --> 00:15:05,640
Speaker 5: Welcome addition to our toolkit for sure.

219
00:15:06,480 --> 00:15:09,160
Speaker 1: So tell me about a case where you use this

220
00:15:09,240 --> 00:15:13,000
Speaker 1: warrant and what happened me personally.

221
00:15:13,080 --> 00:15:16,200
Speaker 4: It's still under investigation, but it was a late night

222
00:15:16,280 --> 00:15:21,520
Speaker 4: gun store burglary where a device is inside of a

223
00:15:21,560 --> 00:15:24,960
Speaker 4: business that's empty, that's right next to the gun store,

224
00:15:25,520 --> 00:15:28,200
Speaker 4: and so that produces a very viable lead for us

225
00:15:28,200 --> 00:15:31,360
Speaker 4: to follow up on. And I've assisted other agencies as

226
00:15:31,400 --> 00:15:35,760
Speaker 4: well with analyzing the results from their Google warrants, in

227
00:15:35,800 --> 00:15:41,000
Speaker 4: particular some road rage type shootings where some aspect of

228
00:15:41,040 --> 00:15:43,720
Speaker 4: the crime is captured on video, and you can compare

229
00:15:43,720 --> 00:15:46,480
Speaker 4: the video surveillance with the data you get from Google

230
00:15:46,760 --> 00:15:52,640
Speaker 4: to potentially identify the suspect. Unfortunately, my own personal experience

231
00:15:52,880 --> 00:15:54,720
Speaker 4: hasn't been as successful as others.

232
00:15:55,280 --> 00:15:58,880
Speaker 5: I strike out far more often than I succeed.

233
00:15:59,840 --> 00:16:02,160
Speaker 1: You walk us through the process of what it's like

234
00:16:02,320 --> 00:16:04,560
Speaker 1: to try to get this information from Google.

235
00:16:05,120 --> 00:16:08,040
Speaker 4: The current process right now begins with a search warrant

236
00:16:08,280 --> 00:16:13,640
Speaker 4: to obtain device IDs that are within a set geographical area.

237
00:16:14,400 --> 00:16:18,680
Speaker 4: So when we request this data from Google, we're asking

238
00:16:18,720 --> 00:16:21,520
Speaker 4: them to search within a geo fence, and that is

239
00:16:21,800 --> 00:16:27,600
Speaker 4: a geographical area defined by latitude and longitude, typically four points.

240
00:16:27,800 --> 00:16:32,360
Speaker 4: It's a square around a given area. We explain the

241
00:16:32,400 --> 00:16:36,400
Speaker 4: circumstances of the crime and what we're hoping to obtain

242
00:16:36,520 --> 00:16:40,440
Speaker 4: from Google and how it will assist our investigation.

243
00:16:41,520 --> 00:16:43,320
Speaker 1: How easy is it to get a warrant like that

244
00:16:43,400 --> 00:16:45,800
Speaker 1: from a judge.

245
00:16:44,840 --> 00:16:47,880
Speaker 5: Here in Maricopa County. We have to really work it.

246
00:16:48,360 --> 00:16:52,560
Speaker 4: There's been some recent court rulings in case law that

247
00:16:53,000 --> 00:16:58,000
Speaker 4: shine a spotlight on these kind of warrants some privacy concerns,

248
00:16:58,560 --> 00:17:02,400
Speaker 4: and so we work with those commissioners at our initial

249
00:17:02,400 --> 00:17:05,320
Speaker 4: appearance court at the Search Warrant Center to really make

250
00:17:05,359 --> 00:17:08,760
Speaker 4: sure that we're keeping our scope narrow, that we're not

251
00:17:09,160 --> 00:17:13,800
Speaker 4: being overly broad, that we are doing everything we can

252
00:17:14,000 --> 00:17:18,040
Speaker 4: to ensure that we're not infringing upon the privacy of

253
00:17:18,080 --> 00:17:20,160
Speaker 4: those persons who are not involved.

254
00:17:19,800 --> 00:17:21,520
Speaker 5: In the crime but are simply in the area.

255
00:17:22,200 --> 00:17:24,560
Speaker 4: So we work really hard to make sure that we

256
00:17:24,720 --> 00:17:27,760
Speaker 4: tailor the geo fence in such a manner that it

257
00:17:27,920 --> 00:17:31,600
Speaker 4: zeros in on the area where the crime was committed.

258
00:17:32,359 --> 00:17:34,520
Speaker 1: Has a judge ever turned you down when you ask

259
00:17:34,560 --> 00:17:37,080
Speaker 1: for one of these warrants or asked you to refine

260
00:17:37,119 --> 00:17:37,840
Speaker 1: the information.

261
00:17:38,680 --> 00:17:42,520
Speaker 4: Yes, probably about half of the Google geofence warrants that

262
00:17:42,600 --> 00:17:48,960
Speaker 4: I've written have undergone a refinement process after being submitted

263
00:17:48,960 --> 00:17:52,359
Speaker 4: the first time round. And again it's just trying to

264
00:17:52,480 --> 00:17:55,520
Speaker 4: zero in strictly on the information that we want to

265
00:17:55,560 --> 00:17:59,480
Speaker 4: obtain and keep out that extraneous information.

266
00:18:00,480 --> 00:18:02,479
Speaker 1: All right, So you get your warrant from the judge

267
00:18:02,800 --> 00:18:05,000
Speaker 1: and you take it to Google. What are you asking

268
00:18:05,080 --> 00:18:07,920
Speaker 1: Google for in a typical case, like how do you say, hey,

269
00:18:07,960 --> 00:18:09,159
Speaker 1: I want information?

270
00:18:10,000 --> 00:18:15,880
Speaker 4: So Google retains all this data on any device that

271
00:18:16,080 --> 00:18:20,720
Speaker 4: has an Android platform or a Google app on it,

272
00:18:20,920 --> 00:18:24,640
Speaker 4: such as the Gmail app or Google Maps. And what

273
00:18:24,680 --> 00:18:28,399
Speaker 4: we're essentially asking from Google is to provide us the

274
00:18:28,920 --> 00:18:33,280
Speaker 4: data that they keep stored within their servers that provides

275
00:18:33,480 --> 00:18:37,439
Speaker 4: locational data for devices that are running some sort of

276
00:18:37,840 --> 00:18:43,800
Speaker 4: Google platform or app. And their locational data is based

277
00:18:43,960 --> 00:18:47,040
Speaker 4: off of a variety of different things, whether it's the

278
00:18:47,119 --> 00:18:50,359
Speaker 4: GPS on the phone, or it's connecting to Wi Fi,

279
00:18:50,600 --> 00:18:53,720
Speaker 4: it's connecting to a cell phone tower, and it comes

280
00:18:53,720 --> 00:18:57,320
Speaker 4: with varying degrees of accuracy. But ultimately what we want

281
00:18:57,359 --> 00:19:01,600
Speaker 4: from them is any device that is within this pre

282
00:19:01,720 --> 00:19:05,600
Speaker 4: defined geographical boundary, and that is the geofence.

283
00:19:06,320 --> 00:19:08,080
Speaker 5: If we have a.

284
00:19:07,560 --> 00:19:11,320
Speaker 4: Murder that's committed inside of a residence, I typically four

285
00:19:11,359 --> 00:19:15,800
Speaker 4: point the corners of that residence, the property itself, and

286
00:19:15,880 --> 00:19:19,000
Speaker 4: if we have, say a getaway path, we know that

287
00:19:19,040 --> 00:19:22,240
Speaker 4: the suspect fled down the street in a certain direction,

288
00:19:22,640 --> 00:19:25,920
Speaker 4: then I'll four point the sidewalk in the street where

289
00:19:25,960 --> 00:19:30,800
Speaker 4: he fled. And we're narrowing those geographical boundaries the geo

290
00:19:30,880 --> 00:19:34,800
Speaker 4: fence so as not to overlap houses of people who

291
00:19:34,840 --> 00:19:37,399
Speaker 4: are uninvolved. We're looking just for that area where the

292
00:19:37,520 --> 00:19:38,600
Speaker 4: suspect was.

293
00:19:39,760 --> 00:19:43,320
Speaker 1: So you go to Google with this information within this

294
00:19:43,600 --> 00:19:45,639
Speaker 1: geo fence, where does it go from there?

295
00:19:46,200 --> 00:19:49,639
Speaker 4: When they returned the information to us, they identified the

296
00:19:49,680 --> 00:19:55,320
Speaker 4: devices through an anonymized alphanumeric code. It's very long and

297
00:19:55,400 --> 00:19:59,960
Speaker 4: it contains absolutely zero identifying information about the account hold.

298
00:20:00,640 --> 00:20:03,240
Speaker 4: And this is to protect the privacy of those devices

299
00:20:03,240 --> 00:20:07,000
Speaker 4: that will inevitably capture that are uninvolved. So it's upon

300
00:20:07,080 --> 00:20:10,159
Speaker 4: us as detectives to study all the devices that have

301
00:20:10,200 --> 00:20:13,919
Speaker 4: been provided and determine which ones are most likely connected

302
00:20:13,960 --> 00:20:17,240
Speaker 4: to the crime that we're investigating, while filtering out those

303
00:20:17,320 --> 00:20:22,119
Speaker 4: that we believe are uninvolved. At that point, we write

304
00:20:22,200 --> 00:20:27,760
Speaker 4: a second search warrant for further locational data on those

305
00:20:27,800 --> 00:20:32,639
Speaker 4: devices that we believe are most likely involved, and that information,

306
00:20:32,960 --> 00:20:36,880
Speaker 4: when it's returned to us from Google, will show locational

307
00:20:37,000 --> 00:20:40,520
Speaker 4: data for those devices an hour prior to and an

308
00:20:40,520 --> 00:20:45,159
Speaker 4: hour after the timeframe set within an original warrant, and

309
00:20:45,200 --> 00:20:45,760
Speaker 4: it will give.

310
00:20:45,680 --> 00:20:47,760
Speaker 5: Us locational data that falls outside of.

311
00:20:47,760 --> 00:20:50,399
Speaker 4: The geo fence, and we can use this to maybe

312
00:20:50,440 --> 00:20:54,800
Speaker 4: determine where people go and see if these devices are

313
00:20:54,960 --> 00:20:57,879
Speaker 4: in fact involved in the manner that we originally thought.

314
00:20:58,520 --> 00:21:01,399
Speaker 4: It's further clarifying and information, but at that point we

315
00:21:01,480 --> 00:21:06,000
Speaker 4: still have no account information, no identifying information on who

316
00:21:06,080 --> 00:21:09,359
Speaker 4: the actual owner is of that device. It's not until

317
00:21:09,359 --> 00:21:12,040
Speaker 4: we get the data back from that second search warrant

318
00:21:12,240 --> 00:21:14,960
Speaker 4: and we can further narrow down the scope.

319
00:21:14,600 --> 00:21:15,680
Speaker 5: Of what we're looking for.

320
00:21:16,000 --> 00:21:18,480
Speaker 4: And then we write a third search warrant, and that

321
00:21:18,600 --> 00:21:22,960
Speaker 4: search warrant is the one where we get actual account information,

322
00:21:23,040 --> 00:21:26,679
Speaker 4: account data, we identify the person who is associated to

323
00:21:26,720 --> 00:21:28,000
Speaker 4: that phone.

324
00:21:28,119 --> 00:21:31,600
Speaker 1: So has access to this kind of information. Changed the

325
00:21:31,600 --> 00:21:34,320
Speaker 1: way your investigations into crimes work.

326
00:21:35,160 --> 00:21:39,680
Speaker 4: Not really, it's for me in particular, As I said before,

327
00:21:40,400 --> 00:21:41,080
Speaker 4: I haven't.

328
00:21:40,800 --> 00:21:42,720
Speaker 5: Had a ton of luck with it.

329
00:21:42,720 --> 00:21:46,080
Speaker 4: It's another tool on our toolkit, and it's a welcome one,

330
00:21:46,440 --> 00:21:49,000
Speaker 4: but it doesn't change the fundamentals really of how we

331
00:21:49,080 --> 00:21:52,320
Speaker 4: proceed with our investigations. It just gives us another avenue

332
00:21:52,359 --> 00:21:55,560
Speaker 4: to turn to when we're coming up dry on more

333
00:21:55,600 --> 00:22:00,520
Speaker 4: traditional investigative techniques. We want to use its kind of

334
00:22:00,560 --> 00:22:05,359
Speaker 4: a measure of last resort. There's other quicker avenues of

335
00:22:05,440 --> 00:22:08,439
Speaker 4: investigation that can be taken at the beginning, and we

336
00:22:08,480 --> 00:22:11,520
Speaker 4: want to go through and exhaust those first. We're not

337
00:22:11,560 --> 00:22:15,280
Speaker 4: necessarily wanting to jump onto the geofense warrant right away

338
00:22:15,760 --> 00:22:19,320
Speaker 4: when we could produce a suspect through other methods a

339
00:22:19,359 --> 00:22:22,520
Speaker 4: lot quicker because there is a time delay in obtaining

340
00:22:22,520 --> 00:22:26,959
Speaker 4: this information from Google, processing it, submitting the second warrant,

341
00:22:27,160 --> 00:22:28,960
Speaker 4: waiting it takes a little bit.

342
00:22:29,680 --> 00:22:32,080
Speaker 1: Are there any kind of crimes where you just don't

343
00:22:32,240 --> 00:22:33,800
Speaker 1: use this to afore.

344
00:22:34,280 --> 00:22:38,840
Speaker 4: We've been cautious in how we apply this, making sure

345
00:22:38,880 --> 00:22:42,119
Speaker 4: that we use it for serious crimes only. We would

346
00:22:42,200 --> 00:22:46,080
Speaker 4: not want to be using this for misdemeanor crimes, lower

347
00:22:46,160 --> 00:22:50,120
Speaker 4: level offenses. You know, we're looking at felony crimes. We're

348
00:22:50,119 --> 00:22:54,359
Speaker 4: looking at Part one crimes, which are going to be burglaries, arsons,

349
00:22:54,440 --> 00:23:00,000
Speaker 4: sex assaults, armed robberies, murders, the serious crimes. That's when

350
00:23:00,080 --> 00:23:02,600
Speaker 4: we're going to use an investigative technique like this.

351
00:23:04,920 --> 00:23:07,480
Speaker 1: You talked a lot about how Google takes all these

352
00:23:07,480 --> 00:23:11,120
Speaker 1: measures to annimize the data. But as someone who's used

353
00:23:11,119 --> 00:23:14,440
Speaker 1: this and seen it used, do you have privacy concerns

354
00:23:14,480 --> 00:23:16,480
Speaker 1: about this to a certain extent?

355
00:23:16,520 --> 00:23:19,800
Speaker 4: I do you know my personal beliefs, Hue towards a

356
00:23:19,840 --> 00:23:23,200
Speaker 4: certain amount of privacy of the individual is very important

357
00:23:23,200 --> 00:23:26,359
Speaker 4: to me. I take it very seriously, and so when

358
00:23:26,680 --> 00:23:31,440
Speaker 4: the courts have privacy concerns, I take those seriously. It's

359
00:23:31,480 --> 00:23:35,320
Speaker 4: reassuring to me that the information that comes back to

360
00:23:35,400 --> 00:23:39,359
Speaker 4: us from Google, it's anonymized and it doesn't contain any

361
00:23:39,400 --> 00:23:44,119
Speaker 4: personal information whatsoever about the person whose device it is

362
00:23:44,560 --> 00:23:49,200
Speaker 4: that we've captured in this geofence. Even more to the point,

363
00:23:49,520 --> 00:23:53,160
Speaker 4: the anonymized ideas that they assign to these devices, it's

364
00:23:53,240 --> 00:23:56,960
Speaker 4: only good for ninety days and then it's deleted and

365
00:23:57,000 --> 00:24:01,520
Speaker 4: there's no record of it. The pace at which technology advances,

366
00:24:01,920 --> 00:24:04,199
Speaker 4: the courts can't keep up with it, and so it

367
00:24:04,280 --> 00:24:07,280
Speaker 4: sits upon us to make sure that we're doing all

368
00:24:07,320 --> 00:24:10,520
Speaker 4: that we can to minimize those issues and I think

369
00:24:10,560 --> 00:24:14,359
Speaker 4: in this regard the search warrants and the way Google

370
00:24:14,400 --> 00:24:16,680
Speaker 4: disseminates that data, I think it's done a really good

371
00:24:16,760 --> 00:24:22,000
Speaker 4: job of protecting the privacy of individuals in regards to

372
00:24:22,040 --> 00:24:23,399
Speaker 4: this investigative technique.

373
00:24:24,040 --> 00:24:27,280
Speaker 1: Detective, appreciate your time, Thanks for giving us this inside view.

374
00:24:28,119 --> 00:24:29,400
Speaker 5: Thank you very much for having me on.

375
00:24:30,680 --> 00:24:33,800
Speaker 1: When we come back, Bloomberg's Julia Love and Davey Alba

376
00:24:33,920 --> 00:24:37,320
Speaker 1: tell us about the pushback in some places against the

377
00:24:37,440 --> 00:24:49,880
Speaker 1: use of these warrants. Julia, do we have any idea

378
00:24:50,080 --> 00:24:53,600
Speaker 1: about whether this works. Does all this information in this

379
00:24:53,720 --> 00:24:57,840
Speaker 1: very complicated system lead to crimes being solved at higher rate?

380
00:24:58,600 --> 00:25:03,760
Speaker 3: We have certainly heard success stories, but we also found

381
00:25:04,119 --> 00:25:08,600
Speaker 3: many cases where this information did not pan out for

382
00:25:08,720 --> 00:25:14,320
Speaker 3: law enforcement. With the taste of that stolen purse in Scottsdale, Arizona,

383
00:25:14,520 --> 00:25:18,280
Speaker 3: the detective told me that he decided not to proceed

384
00:25:18,440 --> 00:25:21,119
Speaker 3: because he didn't think that this tool was going to

385
00:25:21,160 --> 00:25:22,440
Speaker 3: pan out for him.

386
00:25:22,880 --> 00:25:27,640
Speaker 2: We've talked to a police officer who said that they

387
00:25:27,720 --> 00:25:31,080
Speaker 2: were able to identify two suspects in a murder at

388
00:25:31,119 --> 00:25:31,960
Speaker 2: a local park.

389
00:25:32,560 --> 00:25:35,640
Speaker 3: We spoked with one detective in Arizona who told us

390
00:25:35,680 --> 00:25:41,119
Speaker 3: about a victory that his colleague had with these geofence warrants,

391
00:25:41,240 --> 00:25:45,760
Speaker 3: and it was a case of catalytic converter theft in

392
00:25:46,040 --> 00:25:49,920
Speaker 3: a parking structure, and using a geofence warrant, they were

393
00:25:49,960 --> 00:25:53,840
Speaker 3: able to find a device that was present near the

394
00:25:53,880 --> 00:25:57,159
Speaker 3: power box at the exact moment that the power was

395
00:25:57,240 --> 00:25:59,320
Speaker 3: shut off in the parking structure.

396
00:26:01,119 --> 00:26:03,719
Speaker 1: Davy, we talked earlier about how a lot of people's

397
00:26:03,800 --> 00:26:06,720
Speaker 1: data can wind up in one of these warrants and

398
00:26:06,920 --> 00:26:10,840
Speaker 1: these requests for information. Have they ever led to people

399
00:26:10,880 --> 00:26:12,480
Speaker 1: being wrongfully arrested?

400
00:26:13,480 --> 00:26:17,280
Speaker 2: Yes, we do know of at least one crime from

401
00:26:17,320 --> 00:26:21,880
Speaker 2: twenty eighteen where a man in Arizona was wrongly arrested

402
00:26:21,920 --> 00:26:26,560
Speaker 2: for murder based on the location of his phone. What

403
00:26:26,840 --> 00:26:31,600
Speaker 2: actually happened in that case was the man had left

404
00:26:31,640 --> 00:26:36,560
Speaker 2: his phone in a vehicle that a friend took and

405
00:26:36,600 --> 00:26:39,480
Speaker 2: then the fact that the phone was in the car

406
00:26:39,880 --> 00:26:44,240
Speaker 2: sort of pinpointed him as a possible suspect. But he

407
00:26:44,320 --> 00:26:47,200
Speaker 2: actually was not present in that case, and he was

408
00:26:47,520 --> 00:26:49,560
Speaker 2: wrongly accused and arrested.

409
00:26:50,280 --> 00:26:51,840
Speaker 1: And what happened in that case.

410
00:26:52,200 --> 00:26:54,160
Speaker 3: Police ultimately arrested another.

411
00:26:53,960 --> 00:26:59,359
Speaker 1: Suspect and he was ultimately cleared of charges. Yes, Davy

412
00:26:59,440 --> 00:27:02,040
Speaker 1: is anyone pushing back against these kinds of warrants, saying

413
00:27:02,040 --> 00:27:05,240
Speaker 1: that Google shouldn't have the handover users' data if the

414
00:27:05,280 --> 00:27:06,040
Speaker 1: police asked for it.

415
00:27:07,160 --> 00:27:12,640
Speaker 2: Yeah, we have spoken to law professors and privacy advocates

416
00:27:13,280 --> 00:27:19,800
Speaker 2: who say that this practice is incredibly invasive and that

417
00:27:19,960 --> 00:27:23,960
Speaker 2: Google should not have this data in the first place.

418
00:27:24,840 --> 00:27:31,119
Speaker 3: There's an important legal battle happening in Colorado. A group

419
00:27:31,200 --> 00:27:36,320
Speaker 3: of teenagers were charged with murder for an arson case

420
00:27:36,400 --> 00:27:40,040
Speaker 3: that happened there a few years ago. The police found

421
00:27:40,080 --> 00:27:45,119
Speaker 3: them through Doodle's search data for people who searched the

422
00:27:45,240 --> 00:27:48,879
Speaker 3: address of the house around the time of the fire.

423
00:27:49,800 --> 00:27:54,040
Speaker 3: But one of the defendant's lawyers are arguing that this

424
00:27:54,359 --> 00:28:00,520
Speaker 3: was an illegal search, and so that is posing the

425
00:28:00,600 --> 00:28:05,800
Speaker 3: first known challenge to the validity of this technique in

426
00:28:05,840 --> 00:28:06,360
Speaker 3: the country.

427
00:28:08,359 --> 00:28:11,439
Speaker 1: Davey, as you continue to report this story, what are

428
00:28:11,480 --> 00:28:13,480
Speaker 1: you looking for? Where do you think it goes from here?

429
00:28:14,240 --> 00:28:20,000
Speaker 2: I think that we will see more scrutiny on these practices,

430
00:28:20,760 --> 00:28:24,960
Speaker 2: and we may see more public outcry from folks who

431
00:28:25,000 --> 00:28:30,199
Speaker 2: are concerned that they could get ensnared in these warrants.

432
00:28:30,920 --> 00:28:36,000
Speaker 2: You know, there is that question of how the public

433
00:28:36,040 --> 00:28:41,360
Speaker 2: will react to the increased practice of using warrants to

434
00:28:42,000 --> 00:28:46,800
Speaker 2: locate suspects. I think there's a question about how the

435
00:28:46,920 --> 00:28:51,920
Speaker 2: law will evolve to include the fact that this is

436
00:28:51,960 --> 00:28:56,480
Speaker 2: now possible. So, for instance, in the case of wiretapping,

437
00:28:57,320 --> 00:29:00,360
Speaker 2: that is a police technique that can only be used

438
00:29:00,400 --> 00:29:05,200
Speaker 2: for very serious cases like kidnapping, murder, drug trafficking, and

439
00:29:05,240 --> 00:29:09,040
Speaker 2: the like. But there's still no federal law that exists

440
00:29:09,680 --> 00:29:14,520
Speaker 2: for regulating this kind of technique where cops can go

441
00:29:14,600 --> 00:29:18,000
Speaker 2: to Google for location and search data, So it'll be

442
00:29:18,200 --> 00:29:20,760
Speaker 2: interesting to see how that evolves.

443
00:29:21,840 --> 00:29:25,120
Speaker 1: Davey Julia, thanks so much for coming on the show.

444
00:29:25,480 --> 00:29:27,320
Speaker 2: Thanks for having us. It was a pleasure.

445
00:29:27,880 --> 00:29:28,240
Speaker 5: Thank you.

446
00:29:29,520 --> 00:29:31,280
Speaker 1: Thanks for listening to us Here at The Big Take.

447
00:29:31,440 --> 00:29:34,400
Speaker 1: It's a daily podcast from Bloomberg and iHeartRadio. For more

448
00:29:34,440 --> 00:29:38,440
Speaker 1: shows from iHeartRadio, visit the iHeartRadio app, Apple Podcasts, or

449
00:29:38,440 --> 00:29:41,000
Speaker 1: wherever you listen, and we'd love to hear from you.

450
00:29:41,080 --> 00:29:44,480
Speaker 1: Email us questions or comments to Big Take at Bloomberg

451
00:29:44,480 --> 00:29:48,120
Speaker 1: dot Nex. The supervising producer of The Big Take is

452
00:29:48,200 --> 00:29:52,400
Speaker 1: Vicky Bergolina. Our senior producer is Catherine Fink. This episode

453
00:29:52,440 --> 00:29:56,440
Speaker 1: was produced by Sam Gabauer and Christine Driscoll. Rafiel mcli

454
00:29:56,600 --> 00:30:00,120
Speaker 1: Is our engineer. Our original music was composed by Leo Sidrin.

455
00:30:00,480 --> 00:30:04,040
Speaker 1: I'm Westkesova. We'll be back tomorrow with another big take.