1
00:00:14,280 --> 00:00:17,239
Speaker 1: Welcome to Tech Stuff. I'm Os Voloshan here with Cara

2
00:00:17,360 --> 00:00:20,360
Speaker 1: Price High as Hi Cara. So we've talked on the

3
00:00:20,400 --> 00:00:23,720
Speaker 1: show a fair bit about facial recognition technology and the

4
00:00:23,800 --> 00:00:28,320
Speaker 1: ubiquity of data scraping by private companies. It does sometimes

5
00:00:28,360 --> 00:00:30,840
Speaker 1: feel a bit like we are living in a kind

6
00:00:30,840 --> 00:00:34,959
Speaker 1: of science fiction movie about the population of people who

7
00:00:35,000 --> 00:00:36,479
Speaker 1: are being technologically surveiled.

8
00:00:37,280 --> 00:00:39,120
Speaker 2: I think the movie is called Minority Report, and we

9
00:00:39,200 --> 00:00:40,960
Speaker 2: live in it right now.

10
00:00:41,000 --> 00:00:43,479
Speaker 1: And over the past few months, it's become impossible to

11
00:00:43,520 --> 00:00:47,879
Speaker 1: ignore how government agencies are partnering with private companies to

12
00:00:48,000 --> 00:00:51,240
Speaker 1: help in the efforts to do mass deportations.

13
00:00:51,720 --> 00:00:57,279
Speaker 2: Yeah, it's very disconcerting to read about how government agencies

14
00:00:57,320 --> 00:01:01,480
Speaker 2: are using facial recognition technology to detain deput tees.

15
00:01:01,960 --> 00:01:04,360
Speaker 1: Yeah, and it's not just facial recognition. We'll get into

16
00:01:04,400 --> 00:01:07,160
Speaker 1: all the different technologies that are being used today with

17
00:01:07,280 --> 00:01:11,240
Speaker 1: our friend Joseph Cox from four oho form Media. The

18
00:01:11,240 --> 00:01:14,679
Speaker 1: amount of data that immigration and customs enforcement can scrape

19
00:01:14,760 --> 00:01:18,399
Speaker 1: and utilize is pretty mind blowing. And on the other

20
00:01:18,440 --> 00:01:21,000
Speaker 1: side of it, people have developed products to help track

21
00:01:21,319 --> 00:01:24,960
Speaker 1: ice activity, but those products and apps have been removed

22
00:01:25,280 --> 00:01:27,480
Speaker 1: from the Apple App Store, and also I think from

23
00:01:27,520 --> 00:01:30,560
Speaker 1: the Android app stores, which limits how much people are

24
00:01:30,600 --> 00:01:31,560
Speaker 1: able to use these tools.

25
00:01:31,760 --> 00:01:35,120
Speaker 2: You know, how private tech companies are responding to ICE

26
00:01:35,319 --> 00:01:39,160
Speaker 2: does put a spotlight on the relationship between Silicon Valley

27
00:01:39,200 --> 00:01:41,360
Speaker 2: and this presidential administration.

28
00:01:42,080 --> 00:01:46,360
Speaker 1: Joseph said, it's really astonishing what a different environment it

29
00:01:46,440 --> 00:01:49,840
Speaker 1: is now versus the first Trump administration in that respect,

30
00:01:50,000 --> 00:01:52,640
Speaker 1: which we'll get into but without further ado, here's my

31
00:01:52,720 --> 00:01:56,400
Speaker 1: conversation with Joseph Cox. Thanks for joining us today.

32
00:01:56,240 --> 00:01:58,360
Speaker 3: Joseph, yeah, absolutely, thank you for having me.

33
00:01:58,640 --> 00:02:02,400
Speaker 1: So I've got to be questions what's going on with

34
00:02:02,480 --> 00:02:05,680
Speaker 1: ICE in technology and how worried should we all be

35
00:02:06,320 --> 00:02:07,120
Speaker 1: about what's going on?

36
00:02:08,000 --> 00:02:14,120
Speaker 3: So ICE is buying essentially every surveillance technology available under

37
00:02:14,120 --> 00:02:18,239
Speaker 3: the sun. A very brief overview is facial recognition technology,

38
00:02:18,360 --> 00:02:23,680
Speaker 3: social media surveillance, phone location data, spyware for hacking into

39
00:02:23,720 --> 00:02:27,560
Speaker 3: phones remotely fake cell phone towers as well, if you

40
00:02:27,639 --> 00:02:30,400
Speaker 3: can imagine it, and if a law enforcement agency has

41
00:02:30,440 --> 00:02:34,240
Speaker 3: had it before in the US, ICE has almost certainly

42
00:02:34,360 --> 00:02:37,520
Speaker 3: this point purchased it or adapted it, or is at

43
00:02:37,560 --> 00:02:41,360
Speaker 3: least thinking about buying it. As for your second question,

44
00:02:42,000 --> 00:02:46,320
Speaker 3: we actually just published an article today about ICE and

45
00:02:46,360 --> 00:02:50,600
Speaker 3: Customs and Border Protection officials using a facial recognition app

46
00:02:51,120 --> 00:02:55,120
Speaker 3: in the field on people that stops to verify their citizenship.

47
00:02:55,639 --> 00:02:59,120
Speaker 3: And a lawmaker who's also been following the issue told

48
00:02:59,160 --> 00:03:03,000
Speaker 3: me that ICE believes a result from this app can

49
00:03:03,080 --> 00:03:07,040
Speaker 3: override a birth certificate in determining whether someone is a

50
00:03:07,120 --> 00:03:10,360
Speaker 3: citizen of the United States or not. That is obviously

51
00:03:10,400 --> 00:03:13,720
Speaker 3: incredibly alarming that you're going to take the results from

52
00:03:13,720 --> 00:03:16,320
Speaker 3: a piece of technology and you're going to trust that

53
00:03:16,360 --> 00:03:21,360
Speaker 3: and a notoriously inaccurate technology sometimes, and you're going to

54
00:03:21,400 --> 00:03:25,480
Speaker 3: trust that over the piece of paper that says that

55
00:03:25,560 --> 00:03:27,720
Speaker 3: somebody is a US citizen.

56
00:03:28,320 --> 00:03:31,360
Speaker 1: What is the reliability of facial recognition software today?

57
00:03:31,840 --> 00:03:35,160
Speaker 3: So I don't know about this specific technology it's called

58
00:03:35,640 --> 00:03:40,880
Speaker 3: Mobile Fortify, but broadly, over the years, as companies and

59
00:03:40,920 --> 00:03:45,120
Speaker 3: governments have been deploying and developing facial recognition technology, we've

60
00:03:45,120 --> 00:03:49,000
Speaker 3: seen there are racial biases. Often because the data sets

61
00:03:49,000 --> 00:03:51,160
Speaker 3: of all of these photos of faces used to train

62
00:03:51,520 --> 00:03:55,360
Speaker 3: the AI or the machine learning algorithms, they don't have

63
00:03:55,400 --> 00:03:57,800
Speaker 3: as many black people or brown people in them, so

64
00:03:57,800 --> 00:04:01,640
Speaker 3: they can actually be very inaccurate at directly identifying people.

65
00:04:01,880 --> 00:04:04,080
Speaker 3: We've seen a ton of coverage in the New York Times,

66
00:04:04,080 --> 00:04:07,600
Speaker 3: for example, where people have been arrested by mistake because

67
00:04:07,640 --> 00:04:10,400
Speaker 3: they believe one black man is actually another one based

68
00:04:10,440 --> 00:04:13,360
Speaker 3: on the facial recognition technology. Again, I don't know the

69
00:04:13,360 --> 00:04:17,200
Speaker 3: accuracy of this specific tool that ICE is using. They

70
00:04:17,240 --> 00:04:21,800
Speaker 3: all do vary, but this bias, and especially racial bias,

71
00:04:21,839 --> 00:04:26,040
Speaker 3: has been our longstanding criticism of facial recognition over the

72
00:04:26,120 --> 00:04:28,160
Speaker 3: past five, six, even ten years.

73
00:04:28,320 --> 00:04:30,920
Speaker 1: I want to get in start of the specific tools, Joseph,

74
00:04:31,000 --> 00:04:34,120
Speaker 1: but before we get there, I actually several years ago

75
00:04:34,200 --> 00:04:36,800
Speaker 1: hosted a podcast on a very different topic, which was

76
00:04:36,839 --> 00:04:39,520
Speaker 1: called Forgotten Women at Fuarez, and it was about the

77
00:04:40,120 --> 00:04:44,240
Speaker 1: femicides and murders of women on the Texas Mexico border.

78
00:04:44,560 --> 00:04:47,000
Speaker 1: And one of the interesting things that came out of

79
00:04:47,360 --> 00:04:51,080
Speaker 1: spending time there was how everyone on the US side

80
00:04:51,120 --> 00:04:56,240
Speaker 1: of the border talks about the twenty five miles inland

81
00:04:56,279 --> 00:04:59,680
Speaker 1: from the border as this kind of zone where the

82
00:04:59,760 --> 00:05:03,480
Speaker 1: law that govern law enforcement don't really apply. There's all

83
00:05:03,520 --> 00:05:06,880
Speaker 1: these kind of special exceptions for border enforcement and stuff,

84
00:05:07,200 --> 00:05:09,760
Speaker 1: and that therefore that area became a kind of early

85
00:05:09,880 --> 00:05:15,599
Speaker 1: testing ground for both actions that might be considered extra judicial,

86
00:05:16,120 --> 00:05:20,480
Speaker 1: but also for technologies that the wider population wouldn't accept

87
00:05:20,600 --> 00:05:25,200
Speaker 1: if they were deployed on the wider population. Are we

88
00:05:25,320 --> 00:05:29,120
Speaker 1: now seeing the metastasis of the border area and the

89
00:05:29,120 --> 00:05:33,320
Speaker 1: technology used to surveil it and control it across the

90
00:05:33,320 --> 00:05:33,960
Speaker 1: whole country?

91
00:05:34,560 --> 00:05:37,880
Speaker 3: Yeah, I think absolutely. I think a concrete example would

92
00:05:37,960 --> 00:05:42,640
Speaker 3: be the use of surveillance drones. Many people will associate

93
00:05:42,680 --> 00:05:47,240
Speaker 3: these very large predator drones with the US War on terror,

94
00:05:47,279 --> 00:05:50,599
Speaker 3: firing health fire missiles in the Iraq, Afghanistan, Yemen, that

95
00:05:50,680 --> 00:05:53,960
Speaker 3: sort of thing. Those same drones or be unarmed have

96
00:05:54,080 --> 00:05:57,200
Speaker 3: been surveiling the US Mexico border and the Canadian border

97
00:05:57,240 --> 00:05:59,960
Speaker 3: as well for years and years and years. What happened

98
00:06:00,200 --> 00:06:04,440
Speaker 3: now more regularly and more frequently is that DHS will

99
00:06:04,440 --> 00:06:07,800
Speaker 3: fly those same high powered surveillance drones again that were

100
00:06:07,839 --> 00:06:11,200
Speaker 3: designed for war zones above US cities. The most recent

101
00:06:11,240 --> 00:06:15,600
Speaker 3: example of that was the La anti Ice protests and

102
00:06:15,760 --> 00:06:18,880
Speaker 3: DHS flew those same drones above the city. They even

103
00:06:18,920 --> 00:06:22,479
Speaker 3: published some of the footage in social media posts. And

104
00:06:22,560 --> 00:06:26,400
Speaker 3: we see it in all sorts of different technology as well,

105
00:06:26,600 --> 00:06:30,159
Speaker 3: maybe downloading the contents of your phone, which is very

106
00:06:30,240 --> 00:06:33,400
Speaker 3: easy at a border where people have fewer rights obviously

107
00:06:33,440 --> 00:06:37,200
Speaker 3: because they're crossing a border or entering or leaving. We're

108
00:06:37,279 --> 00:06:40,760
Speaker 3: absolutely going to see that sort of thing proliferating. And

109
00:06:40,800 --> 00:06:43,600
Speaker 3: that's one of the main worries from the surveillance experts

110
00:06:43,640 --> 00:06:46,160
Speaker 3: I speak to, is this trickle down effect of well,

111
00:06:46,200 --> 00:06:49,640
Speaker 3: this technology starts in an immigration context, then it goes

112
00:06:49,680 --> 00:06:53,359
Speaker 3: to domestic immigration enforcement, and then it ends up with

113
00:06:53,400 --> 00:06:55,320
Speaker 3: local cops or local sheriff's office.

114
00:06:55,800 --> 00:06:59,520
Speaker 1: Or you hear Trump talking about having the military practiced

115
00:06:59,600 --> 00:07:04,080
Speaker 1: on pops in sanctuary cities. Makes me think of Anderil,

116
00:07:04,360 --> 00:07:07,560
Speaker 1: the weapons manufacturer who are doing all these drones and

117
00:07:07,600 --> 00:07:10,880
Speaker 1: autonomous weapons that are being exported to Taiwan in preparation

118
00:07:11,000 --> 00:07:14,400
Speaker 1: for a potential conflict there. But I believe that Anderil

119
00:07:14,480 --> 00:07:17,080
Speaker 1: back in twenty sixteen developed a lot of these technologies,

120
00:07:17,120 --> 00:07:20,920
Speaker 1: these weapons of wars, again starting off with border contracts.

121
00:07:20,960 --> 00:07:25,520
Speaker 1: So this intersection of the tech the military and using

122
00:07:25,600 --> 00:07:29,240
Speaker 1: military methods to control domestic populations, I mean, as just

123
00:07:29,720 --> 00:07:31,720
Speaker 1: a set of ideas, it's quite terrifying.

124
00:07:32,080 --> 00:07:37,640
Speaker 3: Yeah, Andreill runs these AI powered towers all across the

125
00:07:37,760 --> 00:07:41,160
Speaker 3: US Mexico border and the US Canadian border as well.

126
00:07:41,280 --> 00:07:45,280
Speaker 3: They're capable of all sorts of things like AI object recognition,

127
00:07:45,720 --> 00:07:50,480
Speaker 3: tracking people, tracking vehicles, alerting DHS officials, all of that

128
00:07:50,520 --> 00:07:53,440
Speaker 3: sort of thing. And there is just a big shift

129
00:07:53,600 --> 00:07:57,560
Speaker 3: now in Silicon Valley in the tech companies where they

130
00:07:57,600 --> 00:08:01,640
Speaker 3: are more explicitly going to work with the government. Whereas

131
00:08:01,640 --> 00:08:07,680
Speaker 3: before Facebook, Instagram, Meta, Twitter or whoever uber, they were

132
00:08:07,760 --> 00:08:10,640
Speaker 3: very much separate from the government. We're making consumer goods,

133
00:08:10,680 --> 00:08:14,640
Speaker 3: We're making an app for whatever reason. Tech companies today

134
00:08:15,440 --> 00:08:19,000
Speaker 3: are not only leaning into they're very proud to work

135
00:08:19,000 --> 00:08:20,960
Speaker 3: with the government. And Andrew has been a leader in that,

136
00:08:21,040 --> 00:08:24,480
Speaker 3: And obviously other examples would be palanted, for example, where

137
00:08:24,800 --> 00:08:28,880
Speaker 3: these companies say, this is our purpose is to protect

138
00:08:29,040 --> 00:08:32,760
Speaker 3: the United States. Now, of course, how people understand what

139
00:08:32,840 --> 00:08:35,880
Speaker 3: protecting the United States might actually be is obviously going

140
00:08:35,960 --> 00:08:38,920
Speaker 3: to vary, especially when we talk about surveillance technology and

141
00:08:38,960 --> 00:08:41,240
Speaker 3: how that's being used and who is being used against.

142
00:08:42,080 --> 00:08:46,440
Speaker 1: You mentioned facial recognition. Clearview AI always seems to come

143
00:08:46,520 --> 00:08:49,160
Speaker 1: up sort of as the big batter wolf of this

144
00:08:49,760 --> 00:08:53,320
Speaker 1: topic area. Can you explain a bit about how that

145
00:08:53,440 --> 00:08:56,640
Speaker 1: technology or that company is being used by or partnered with,

146
00:08:56,760 --> 00:08:57,280
Speaker 1: the government.

147
00:08:57,960 --> 00:09:01,319
Speaker 3: Yes, so we just revealed that I paid clearview AI

148
00:09:01,559 --> 00:09:06,080
Speaker 3: another five million dollars for access to its technology. Ice

149
00:09:06,080 --> 00:09:09,040
Speaker 3: has been contracting with Clearview for years. It's something of

150
00:09:09,120 --> 00:09:12,480
Speaker 3: the established player in the government or law enforcement facial

151
00:09:12,520 --> 00:09:17,640
Speaker 3: recognition space. What Clearview did several years ago was basically

152
00:09:17,720 --> 00:09:22,880
Speaker 3: scrape a ton of photos of people's faces from the

153
00:09:23,040 --> 00:09:26,320
Speaker 3: World Wide Web, so they're then no profiles, are social

154
00:09:26,320 --> 00:09:29,560
Speaker 3: media profiles. And they made this massive database which is

155
00:09:29,600 --> 00:09:32,880
Speaker 3: now something like ten billion images, and they allowed people

156
00:09:33,200 --> 00:09:35,520
Speaker 3: to upload a photo and we will then show you

157
00:09:35,760 --> 00:09:39,040
Speaker 3: similar photos, probably of the same person, and you can

158
00:09:39,080 --> 00:09:41,760
Speaker 3: then figure out who this person is. And when it

159
00:09:41,800 --> 00:09:44,120
Speaker 3: was revealed by The New York Times several years ago,

160
00:09:44,559 --> 00:09:48,560
Speaker 3: that was really crossing the line, like Facebook could have

161
00:09:48,600 --> 00:09:52,160
Speaker 3: done this, and although they had facial recognition technology, it

162
00:09:52,200 --> 00:09:54,240
Speaker 3: was to tag people, you know, when you tag your

163
00:09:54,240 --> 00:09:58,160
Speaker 3: friends and your photos. It wasn't to unmask strangers. Google

164
00:09:58,200 --> 00:10:00,439
Speaker 3: could have done it. They said, this is two dangerous,

165
00:10:00,480 --> 00:10:04,760
Speaker 3: we don't want to Clearview cross that line. And incredibly

166
00:10:04,800 --> 00:10:08,040
Speaker 3: controversial all the time, still controversial now, but it's almost

167
00:10:08,080 --> 00:10:12,520
Speaker 3: normal now. There are Clearview contracts every single day essentially

168
00:10:12,600 --> 00:10:16,160
Speaker 3: where local cops are already using that, sheriff's offices are

169
00:10:16,240 --> 00:10:18,880
Speaker 3: or using that, and ICE is absolutely using it as well.

170
00:10:19,080 --> 00:10:22,560
Speaker 3: With this latest contract. Ice spent this five million dollars

171
00:10:22,640 --> 00:10:25,480
Speaker 3: in part because it believes it can use facial recognition

172
00:10:25,520 --> 00:10:29,199
Speaker 3: technology to unmask the people that it believes are quote,

173
00:10:29,200 --> 00:10:30,760
Speaker 3: assaulting officers.

174
00:10:31,640 --> 00:10:36,120
Speaker 1: So that's facial recognition. There's also location tracking. Talk about flock.

175
00:10:36,720 --> 00:10:40,240
Speaker 3: Flock is an interesting company. A lot of people may

176
00:10:40,280 --> 00:10:43,000
Speaker 3: have actually heard about it because it is proliferated around

177
00:10:43,040 --> 00:10:46,480
Speaker 3: local communities. And what they do is they sell these

178
00:10:46,559 --> 00:10:51,640
Speaker 3: cameras to police departments or homeowner associations or neighborhoods or

179
00:10:51,679 --> 00:10:56,040
Speaker 3: city governments or whatever. And these cameras they sit stationary

180
00:10:56,240 --> 00:11:00,160
Speaker 3: at the road and they continuously scan every vehicle that

181
00:11:00,200 --> 00:11:02,360
Speaker 3: goes by. That will be the model of the car,

182
00:11:02,679 --> 00:11:06,240
Speaker 3: the brand, the color, and most importantly the license plate.

183
00:11:06,640 --> 00:11:10,680
Speaker 3: And this creates this vast surveillance network that law enforcement

184
00:11:10,720 --> 00:11:15,040
Speaker 3: officers can tap into without a warrant. Typically you enter

185
00:11:15,720 --> 00:11:18,920
Speaker 3: xyz license plate, it shows you all the locations this

186
00:11:19,040 --> 00:11:20,520
Speaker 3: vehicle has been and.

187
00:11:20,480 --> 00:11:22,000
Speaker 1: Do you have to be law enforce where anyone can

188
00:11:22,080 --> 00:11:22,320
Speaker 1: use this?

189
00:11:22,559 --> 00:11:25,920
Speaker 3: If your homeowners association, it's going to be the authorized

190
00:11:26,000 --> 00:11:30,200
Speaker 3: user of that particular homeowners association or that neighborhood or something.

191
00:11:30,200 --> 00:11:32,480
Speaker 3: It's similar to a ring camera. In a way, but

192
00:11:32,520 --> 00:11:33,920
Speaker 3: you're not going to be able to access all of

193
00:11:33,920 --> 00:11:34,240
Speaker 3: the data.

194
00:11:34,280 --> 00:11:36,760
Speaker 1: I couldn't if I was a stalker and know my

195
00:11:36,960 --> 00:11:39,920
Speaker 1: ex's number plate and therefore track them around the country

196
00:11:39,960 --> 00:11:42,240
Speaker 1: as an individual person. But as a member of law enforcement,

197
00:11:42,280 --> 00:11:42,600
Speaker 1: I could.

198
00:11:43,120 --> 00:11:46,160
Speaker 3: Yes. Essentially yes, and law enforcement have access to this

199
00:11:46,240 --> 00:11:49,640
Speaker 3: sort of additional capability, which is the National Lookup Tool,

200
00:11:49,920 --> 00:11:53,480
Speaker 3: where a cop, say in Texas, can search cameras in

201
00:11:53,600 --> 00:11:57,400
Speaker 3: Illinois or California. And that's not even a hypothetic example.

202
00:11:57,440 --> 00:12:00,920
Speaker 3: We actually reported that a law enforcement office in Texas

203
00:12:01,280 --> 00:12:04,679
Speaker 3: searched for a woman who self administered an abortion, and

204
00:12:04,720 --> 00:12:09,200
Speaker 3: it searched cameras all over the country, and crucially, we

205
00:12:09,360 --> 00:12:13,160
Speaker 3: found cops were performing lookups for ICE in this tool.

206
00:12:13,200 --> 00:12:15,680
Speaker 3: We got these spreadsheets has said all the reasons why,

207
00:12:15,960 --> 00:12:18,240
Speaker 3: and you scroll through them and some just say ICE,

208
00:12:18,440 --> 00:12:22,520
Speaker 3: some say immigration or immigration enforcement. Immigration protest was another

209
00:12:22,559 --> 00:12:27,040
Speaker 3: one as well, and that was already rarely interesting enough,

210
00:12:27,160 --> 00:12:30,640
Speaker 3: and Flock made serious changes after we revealed that, But

211
00:12:31,080 --> 00:12:33,559
Speaker 3: we then also learned that ICE actually had direct access

212
00:12:33,600 --> 00:12:35,520
Speaker 3: as well. They didn't even have to go through the cops,

213
00:12:35,559 --> 00:12:39,040
Speaker 3: and that's apparently stopped now, But ICE were getting access

214
00:12:39,040 --> 00:12:42,199
Speaker 3: to that system through a side door via local police,

215
00:12:42,240 --> 00:12:45,080
Speaker 3: but then also just direct access as well, so you've.

216
00:12:44,960 --> 00:12:49,320
Speaker 1: Got facial rehognition, license plate tracking, and then we have

217
00:12:49,480 --> 00:12:52,360
Speaker 1: this sort of system that sits even above, which is

218
00:12:52,400 --> 00:12:56,199
Speaker 1: the investigative case management you mentioned Palentier earlier. This is

219
00:12:56,200 --> 00:12:57,280
Speaker 1: this a Palentier product.

220
00:12:57,960 --> 00:13:03,360
Speaker 3: So it started as DHS project and it's basically a

221
00:13:03,400 --> 00:13:06,719
Speaker 3: centralized database and you look into it and you can

222
00:13:07,000 --> 00:13:10,319
Speaker 3: look up a person that is their name, social Security number,

223
00:13:10,600 --> 00:13:14,480
Speaker 3: maybe identifying features like tattoos or scars, that sort of thing.

224
00:13:14,880 --> 00:13:18,920
Speaker 3: Palenteer then started work on that, and this was years

225
00:13:18,920 --> 00:13:20,520
Speaker 3: and years ago. They've been doing this for something like

226
00:13:20,600 --> 00:13:22,920
Speaker 3: ten years at this point. But then with the second

227
00:13:22,920 --> 00:13:28,000
Speaker 3: Trump administration, something changed and it became clear that ICE

228
00:13:28,040 --> 00:13:31,480
Speaker 3: and DHS more broadly needed this tool to be a

229
00:13:31,559 --> 00:13:34,440
Speaker 3: bit more powerful, which is where they brought in Palenteer

230
00:13:34,520 --> 00:13:37,439
Speaker 3: to actually take over the development, not just the maintenance,

231
00:13:37,480 --> 00:13:42,240
Speaker 3: of this project. And that's how we come to immigration OS, which,

232
00:13:42,480 --> 00:13:45,840
Speaker 3: according to material I got leaked from inside Palenteer, is

233
00:13:45,920 --> 00:13:48,880
Speaker 3: designed to make it much easier to find the physical

234
00:13:48,960 --> 00:13:53,040
Speaker 3: location of aliens. Of course, the US government's terms for

235
00:13:53,240 --> 00:13:58,440
Speaker 3: immigrants essentially we don't know all of the different data

236
00:13:58,440 --> 00:14:02,720
Speaker 3: that's going into this system alan to their entire the

237
00:14:02,840 --> 00:14:06,400
Speaker 3: entire purpose of their existence is to match together all

238
00:14:06,400 --> 00:14:08,520
Speaker 3: of this different sorts of data from all over the

239
00:14:08,520 --> 00:14:12,319
Speaker 3: place and make it intelligible and readable for government customers.

240
00:14:12,520 --> 00:14:15,480
Speaker 3: So we can assume that Immigration OS is probably bringing

241
00:14:15,559 --> 00:14:18,240
Speaker 3: data from all over the place into an interface that

242
00:14:18,480 --> 00:14:21,600
Speaker 3: ICE or DHS can use to I mean, essentially track

243
00:14:21,640 --> 00:14:23,320
Speaker 3: down people at least recording some TIRA.

244
00:14:23,440 --> 00:14:28,120
Speaker 1: I got interesting. One of my dear friends and board

245
00:14:28,120 --> 00:14:32,080
Speaker 1: members at Kaleidoscope is a man called Stuart Carl who's

246
00:14:32,120 --> 00:14:34,600
Speaker 1: a First Amendment lawyer. He was actually the general counselor

247
00:14:34,680 --> 00:14:36,720
Speaker 1: for the Wall Street Journal for many years, and now

248
00:14:36,760 --> 00:14:42,239
Speaker 1: he teaches at Columbia Journalism School, and he advised students

249
00:14:43,120 --> 00:14:47,040
Speaker 1: basically to avoid posting on social media who are not

250
00:14:47,160 --> 00:14:50,640
Speaker 1: US citizens about political matters at least until after they graduate.

251
00:14:50,920 --> 00:14:53,360
Speaker 1: I think his point was get your degree that you've

252
00:14:53,400 --> 00:14:55,560
Speaker 1: come for, and then decide where you want to draw

253
00:14:55,600 --> 00:14:57,720
Speaker 1: your line. He got a lot of blowback for that,

254
00:14:57,840 --> 00:15:00,720
Speaker 1: and there was debate and a York Time story about it.

255
00:15:00,120 --> 00:15:04,200
Speaker 1: But this idea of second guessing what you post on

256
00:15:04,240 --> 00:15:07,440
Speaker 1: social media in the current political environment, I mean, how

257
00:15:07,480 --> 00:15:12,680
Speaker 1: programmatic is social media scrapian analysis versus opportunistic when they

258
00:15:12,720 --> 00:15:14,440
Speaker 1: already want to build a case against somebody.

259
00:15:14,960 --> 00:15:17,760
Speaker 3: Yeah, they do it in several different ways, but broadly,

260
00:15:18,320 --> 00:15:22,600
Speaker 3: you have right wing agitators who are constantly going through

261
00:15:22,640 --> 00:15:26,480
Speaker 3: social media Twitter, Instagram. Whatever they see something they don't like,

262
00:15:26,880 --> 00:15:30,720
Speaker 3: they flag it to the administration. And we have several

263
00:15:31,000 --> 00:15:35,320
Speaker 3: very high profile right wing social media personalities who are

264
00:15:35,360 --> 00:15:37,640
Speaker 3: clearly doing that, and they're emitting as much on their

265
00:15:37,680 --> 00:15:41,840
Speaker 3: own social media. The other side are these technological tools

266
00:15:42,320 --> 00:15:45,240
Speaker 3: which ICE has, Customs and Border Protection has, and the

267
00:15:45,280 --> 00:15:48,560
Speaker 3: State Department I believe has as well, and these are

268
00:15:48,800 --> 00:15:52,840
Speaker 3: broadly speaking AI powers. They say they can make decisions

269
00:15:52,960 --> 00:15:56,360
Speaker 3: or inferences based on somebody's profile, what they've posted, the

270
00:15:56,440 --> 00:15:58,400
Speaker 3: location they've done it from, and all of that sort

271
00:15:58,400 --> 00:16:02,840
Speaker 3: of thing. Again, already a black box. We're also trying

272
00:16:02,840 --> 00:16:05,320
Speaker 3: to look at a black box from the outside, so

273
00:16:05,360 --> 00:16:07,960
Speaker 3: it's doubly opaque about what is actually going on. But

274
00:16:08,040 --> 00:16:10,640
Speaker 3: what we do know is that these government agencies SUS

275
00:16:10,640 --> 00:16:14,400
Speaker 3: spend millions and millions of dollars on social media surveillance tools.

276
00:16:14,440 --> 00:16:17,280
Speaker 3: We just don't know exactly how it goes from post

277
00:16:17,600 --> 00:16:20,400
Speaker 3: to visa being taken away. But the threat is absolutely

278
00:16:20,400 --> 00:16:22,920
Speaker 3: there that if you say something that the administration does

279
00:16:23,000 --> 00:16:26,600
Speaker 3: not agree with, there's every chance that some sort of

280
00:16:26,640 --> 00:16:30,840
Speaker 3: pretense will be used to revoke you your immigration status.

281
00:16:31,160 --> 00:16:36,200
Speaker 1: You know. As this sort of ice crisis royals the country.

282
00:16:36,240 --> 00:16:39,320
Speaker 1: There are certain courts that are fighting back. I mean

283
00:16:39,360 --> 00:16:45,080
Speaker 1: recently in Chicago, one of the judges basically hauled in

284
00:16:45,160 --> 00:16:49,640
Speaker 1: the local head of border patrol and said, you know,

285
00:16:49,680 --> 00:16:51,400
Speaker 1: you have to come and report to me every single

286
00:16:51,480 --> 00:16:54,160
Speaker 1: day at six pm and what you're doing. You can't

287
00:16:54,200 --> 00:16:56,840
Speaker 1: throw tear gas into crowds where children are about to

288
00:16:56,920 --> 00:16:58,640
Speaker 1: arrive with the warning, etc.

289
00:16:58,880 --> 00:16:59,040
Speaker 3: Etc.

290
00:16:59,400 --> 00:17:03,320
Speaker 1: And essentially kind of insisted on some kind of role

291
00:17:03,560 --> 00:17:08,639
Speaker 1: for judiciary in managing the behavior or the illegal behavior

292
00:17:08,720 --> 00:17:12,040
Speaker 1: or the potentially illegal behavior of some officials. How's that

293
00:17:12,080 --> 00:17:14,160
Speaker 1: playing out in the tech realm, That's a.

294
00:17:14,200 --> 00:17:17,720
Speaker 3: Very good question, actually, Yeah. And with the court stuff, obviously,

295
00:17:17,760 --> 00:17:20,960
Speaker 3: that is the American system working as intenders, where you

296
00:17:20,960 --> 00:17:24,600
Speaker 3: have all of these different branches of government and all

297
00:17:24,640 --> 00:17:26,560
Speaker 3: of these other things, like the courts, and they provide

298
00:17:26,560 --> 00:17:31,800
Speaker 3: these checks and balances. That doesn't exist really for the

299
00:17:31,840 --> 00:17:37,480
Speaker 3: private surveillance tech social media industry. We have had impact

300
00:17:37,720 --> 00:17:41,600
Speaker 3: through our reporting. Again, our reporting or FLOCK ended up

301
00:17:41,640 --> 00:17:46,080
Speaker 3: with the company removing California, Illinois, and I think a

302
00:17:46,119 --> 00:17:49,440
Speaker 3: couple more states from its National Lookup TOL, meaning ICE

303
00:17:49,480 --> 00:17:52,160
Speaker 3: can no longer access those through the police, and then

304
00:17:52,280 --> 00:17:56,040
Speaker 3: FLOCK also stopped. It's more direct sharing with ICE and

305
00:17:56,040 --> 00:17:59,119
Speaker 3: Customs and Border Protection as well. There has been some

306
00:17:59,240 --> 00:18:04,200
Speaker 3: pressure from Democrats on House committees looking into the facial

307
00:18:04,240 --> 00:18:07,560
Speaker 3: recognition side as well after our reporting. That being said,

308
00:18:08,119 --> 00:18:12,000
Speaker 3: accountability looks very very different in the second of Trump administration,

309
00:18:12,400 --> 00:18:16,200
Speaker 3: where you have masked agents, you have officials physically assaulting

310
00:18:16,240 --> 00:18:18,840
Speaker 3: people in the corridors of court rooms in New York

311
00:18:18,880 --> 00:18:22,159
Speaker 3: and then two days later they're back at work. Accountability

312
00:18:22,160 --> 00:18:25,679
Speaker 3: looks very very different here. And for the tech stuff,

313
00:18:26,480 --> 00:18:31,440
Speaker 3: the accountability comes from the outside. It's not really from government,

314
00:18:31,480 --> 00:18:33,760
Speaker 3: because of course the government wants to work for these companies.

315
00:18:34,000 --> 00:18:36,920
Speaker 3: It can come from inside where people are clearly annoyed

316
00:18:36,920 --> 00:18:40,240
Speaker 3: inside Pale or Flock and then they'll provide information to me.

317
00:18:41,160 --> 00:18:44,440
Speaker 3: But it looks very very different to as you say,

318
00:18:44,520 --> 00:18:46,960
Speaker 3: taking a senior official in front of a judge and

319
00:18:47,000 --> 00:18:49,720
Speaker 3: having him report every day. It's just a completely different environment.

320
00:18:56,119 --> 00:18:59,879
Speaker 1: After the break the people fighting back against the ICE

321
00:19:00,200 --> 00:19:10,639
Speaker 1: surveillance state stay with us. There is also resistance tech.

322
00:19:11,320 --> 00:19:14,360
Speaker 1: I'm very curious to hear about ice block. I mean,

323
00:19:14,720 --> 00:19:18,040
Speaker 1: our producer Eliza made an interesting point that, you know,

324
00:19:18,080 --> 00:19:22,480
Speaker 1: after the California fires, there was this great interest in

325
00:19:22,920 --> 00:19:26,680
Speaker 1: funding fire watching apps and early warning systems, and there's

326
00:19:26,680 --> 00:19:31,400
Speaker 1: also an explosion of apps around helping communities avoid ice

327
00:19:31,480 --> 00:19:34,199
Speaker 1: round apps and stuff that it will be fair, they

328
00:19:34,200 --> 00:19:37,159
Speaker 1: didn't receive the same funding attention. Talk a little bit

329
00:19:37,160 --> 00:19:40,760
Speaker 1: about these apps and the resistance to them.

330
00:19:41,320 --> 00:19:44,399
Speaker 3: Yeah, so ice block is the most prominent one. It

331
00:19:44,440 --> 00:19:47,399
Speaker 3: was launched earlier this year. People only really started to

332
00:19:47,440 --> 00:19:51,240
Speaker 3: pay attention in around June when CNN covered the app.

333
00:19:51,680 --> 00:19:54,800
Speaker 3: And all ice block does is it allows people to

334
00:19:54,960 --> 00:19:58,680
Speaker 3: report sightings of ICE officials in their local proximity. It

335
00:19:58,800 --> 00:20:02,040
Speaker 3: uses the location data off your iPhone. It was only

336
00:20:02,119 --> 00:20:04,800
Speaker 3: on iPhone. It wasn't on Android for privacy reasons, is

337
00:20:04,840 --> 00:20:07,920
Speaker 3: what the developer told me. And if you see an

338
00:20:07,920 --> 00:20:10,280
Speaker 3: ICE official or some sort of ICE raid within five

339
00:20:10,320 --> 00:20:12,720
Speaker 3: miles of you or whatever, you can report it and

340
00:20:12,760 --> 00:20:16,240
Speaker 3: then other people around the area will receive a push alert.

341
00:20:16,280 --> 00:20:22,080
Speaker 3: It's rarely similar to ring the neighborhood app or even

342
00:20:22,320 --> 00:20:24,520
Speaker 3: a citizen or something like that. All the fire watching

343
00:20:24,560 --> 00:20:28,320
Speaker 3: apps you mentioned as well. Time goes on. The DOJ

344
00:20:28,480 --> 00:20:31,520
Speaker 3: does not like that CNN has reported this and amplified it.

345
00:20:31,840 --> 00:20:34,400
Speaker 3: The DOJ even says it's trying to find a way

346
00:20:34,440 --> 00:20:37,600
Speaker 3: to charge CNN. Obviously that didn't come to anything at

347
00:20:37,680 --> 00:20:41,080
Speaker 3: least yet. But fast forward to September and there is

348
00:20:41,160 --> 00:20:44,440
Speaker 3: a horrible shooting at an ICE facility where a gunman

349
00:20:44,840 --> 00:20:48,600
Speaker 3: shoots at the facility and he hits free detainees. I

350
00:20:48,640 --> 00:20:53,000
Speaker 3: believe two died and one survived. When authorities looking to

351
00:20:53,040 --> 00:20:56,520
Speaker 3: that person, they look at his phone and they find

352
00:20:56,520 --> 00:21:01,400
Speaker 3: that that person was searching for ice spotting apps explicitly

353
00:21:01,840 --> 00:21:06,320
Speaker 3: ice block. The DOJ and Pambondi, the Attorney General then

354
00:21:06,520 --> 00:21:10,200
Speaker 3: uses that as a justification to go to Apple and say, look,

355
00:21:10,240 --> 00:21:13,600
Speaker 3: you have to remove this app. This is encouraging violence

356
00:21:14,040 --> 00:21:17,640
Speaker 3: against ICE officials. Putting aside the fact that obviously ICE

357
00:21:17,640 --> 00:21:19,360
Speaker 3: officials are going to be an ICE facility, I don't

358
00:21:19,359 --> 00:21:22,520
Speaker 3: think you need an app to determine that. But Apple

359
00:21:22,880 --> 00:21:26,040
Speaker 3: gives in and it removes not just that app, but

360
00:21:26,080 --> 00:21:29,520
Speaker 3: a host of other ones at the explicit request of

361
00:21:29,960 --> 00:21:32,919
Speaker 3: the US Department of Justice. Now Apple removes and Google

362
00:21:32,920 --> 00:21:36,879
Speaker 3: removes apps every single day for doing sketchy or weird

363
00:21:36,960 --> 00:21:42,359
Speaker 3: or misleading stuff. It is an entirely different thing for

364
00:21:42,400 --> 00:21:45,520
Speaker 3: the US government to go and say remove this app,

365
00:21:45,760 --> 00:21:49,680
Speaker 3: especially when that app was actually facilitating First Amendment protected

366
00:21:49,720 --> 00:21:52,879
Speaker 3: speech as well. This is not illegal to report where

367
00:21:53,359 --> 00:21:56,160
Speaker 3: law enforcement are. That's a core tenant of the First Amendment.

368
00:21:56,880 --> 00:21:58,560
Speaker 1: Yeah, it's kind of extraordinary. I want to help people

369
00:21:58,680 --> 00:22:03,679
Speaker 1: look back on this time. I mean, obviously, people who

370
00:22:03,720 --> 00:22:07,160
Speaker 1: are rich and powerful like getting richer and more powerful.

371
00:22:07,280 --> 00:22:12,880
Speaker 1: But the capitulation every step of the way of private companies,

372
00:22:13,240 --> 00:22:18,400
Speaker 1: especially tech companies, to these requests and mandates, it's quite extraordinary.

373
00:22:18,440 --> 00:22:22,000
Speaker 1: I wonder how it will be judged by history.

374
00:22:22,760 --> 00:22:25,720
Speaker 3: Yeah. I remember in the first Trump administration, I feel

375
00:22:25,720 --> 00:22:29,800
Speaker 3: like there was so much more pushback from tech companies,

376
00:22:29,880 --> 00:22:32,600
Speaker 3: you know, even surveillance companies as well. I think Amazon

377
00:22:32,680 --> 00:22:37,600
Speaker 3: stopped selling its facial recognition technology to law enforcement agencies

378
00:22:37,600 --> 00:22:41,080
Speaker 3: for a period of time as well. That has completely

379
00:22:41,160 --> 00:22:45,200
Speaker 3: gone Now. It is absolute capitulation to the government. It

380
00:22:45,320 --> 00:22:48,720
Speaker 3: is doing whatever they want. It is not really pushing back.

381
00:22:48,760 --> 00:22:51,120
Speaker 3: I mean Apple removes that app. You know, a few

382
00:22:51,119 --> 00:22:53,399
Speaker 3: weeks earlier or whatever it was, Tim Cook was giving

383
00:22:53,680 --> 00:22:57,680
Speaker 3: Trump a golden statue of the Apple logo or whatever.

384
00:22:58,080 --> 00:23:02,080
Speaker 3: Another example is that DHS right now is using a

385
00:23:02,080 --> 00:23:05,880
Speaker 3: lot of memes from the Halo video game franchise which

386
00:23:05,920 --> 00:23:11,359
Speaker 3: are dehumanizing immigrants and basically comparing them to like an

387
00:23:11,400 --> 00:23:15,840
Speaker 3: alien parasite. Microsoft hasn't responded to that. Microsoft hasn't, you know,

388
00:23:16,480 --> 00:23:19,240
Speaker 3: asked them to take it down. It's an entirely different

389
00:23:19,320 --> 00:23:22,520
Speaker 3: environment now than it was in the first Trump administration

390
00:23:22,560 --> 00:23:25,360
Speaker 3: where it comes to tech and even just business right

391
00:23:25,480 --> 00:23:26,840
Speaker 3: and private companies.

392
00:23:27,480 --> 00:23:30,360
Speaker 1: So who's fighting back? I mean, there's Joshua Aaron who

393
00:23:30,480 --> 00:23:33,560
Speaker 1: founded the ice block app. Where is their resistance to this?

394
00:23:33,640 --> 00:23:36,320
Speaker 1: And also what the normal people do in their daily

395
00:23:36,359 --> 00:23:39,600
Speaker 1: lives to protect themselves from the reach of some of

396
00:23:39,600 --> 00:23:41,920
Speaker 1: these technologies and keep themselves safe.

397
00:23:42,240 --> 00:23:44,400
Speaker 3: Yeah, I mean there's a lot of different things people

398
00:23:44,440 --> 00:23:47,680
Speaker 3: are doing. There are other apps similar to ice Block.

399
00:23:47,760 --> 00:23:50,680
Speaker 3: There was one called eyes up. There was just cataloging

400
00:23:51,280 --> 00:23:54,280
Speaker 3: videos of ice abuses. I wasn't even tracking the physical

401
00:23:54,320 --> 00:23:58,200
Speaker 3: location of people. That developer who he only gave me

402
00:23:58,359 --> 00:24:01,679
Speaker 3: his first name, Mark, but he's doing work in his

403
00:24:01,720 --> 00:24:02,080
Speaker 3: own way.

404
00:24:02,080 --> 00:24:02,359
Speaker 2: There.

405
00:24:02,560 --> 00:24:05,440
Speaker 3: You even have states like New York State is basically

406
00:24:05,440 --> 00:24:07,919
Speaker 3: building it's equivalent of that, which is, hey, if you

407
00:24:08,040 --> 00:24:12,600
Speaker 3: have videos of abuses by ICE, send them to us,

408
00:24:12,840 --> 00:24:15,960
Speaker 3: Like there will be consequences for this later. So people

409
00:24:16,040 --> 00:24:19,000
Speaker 3: right now are thinking about, well, how do we collect

410
00:24:19,080 --> 00:24:24,000
Speaker 3: evidence for accountability later? Like it can't happen now. There's

411
00:24:24,080 --> 00:24:26,920
Speaker 3: no real way to hold a ICE agent who maybe

412
00:24:27,000 --> 00:24:30,719
Speaker 3: detains a US citizen unlawfully to account because they're masks,

413
00:24:30,760 --> 00:24:33,600
Speaker 3: and there's this complete culture of we're just going to

414
00:24:33,640 --> 00:24:36,119
Speaker 3: do whatever we want. But people are preparing for that.

415
00:24:36,440 --> 00:24:38,119
Speaker 3: As for the second part of your questions and what

416
00:24:38,160 --> 00:24:41,199
Speaker 3: people can do to protect their privacy, I mean, it

417
00:24:41,240 --> 00:24:45,280
Speaker 3: really depends on the technology, but I would sort of

418
00:24:45,280 --> 00:24:47,119
Speaker 3: suggest some of the normal stuff when it comes to

419
00:24:47,160 --> 00:24:50,560
Speaker 3: tracking mobile phones, which is, keep the phone fully up

420
00:24:50,560 --> 00:24:56,360
Speaker 3: to date, uninstall sketchy apps, don't give location permissions. Of course,

421
00:24:56,440 --> 00:24:58,600
Speaker 3: there's lots more you could do, and it's going to

422
00:24:58,640 --> 00:25:01,040
Speaker 3: depend upon your individe your use case. So you a

423
00:25:01,160 --> 00:25:04,640
Speaker 3: lawyer defending immigrants, so you're a journalist, are you an

424
00:25:04,640 --> 00:25:07,800
Speaker 3: activists whatever? So people should do their own research. But

425
00:25:07,920 --> 00:25:10,480
Speaker 3: there are things that people can do if they think

426
00:25:10,480 --> 00:25:11,119
Speaker 3: it's appropriate.

427
00:25:13,480 --> 00:25:15,200
Speaker 1: Joseph thank you, Thank you so much.

428
00:25:41,760 --> 00:25:43,560
Speaker 2: That's it for this week for tech Stuff.

429
00:25:43,560 --> 00:25:46,439
Speaker 1: I'm Cara Price and I'm os Vlocian. This episode was

430
00:25:46,440 --> 00:25:50,320
Speaker 1: produced by Eliza Dennis, Tyler Hill and Melissa Slaughter. It

431
00:25:50,400 --> 00:25:53,720
Speaker 1: was executive produced by me Cara Price, Julia Nutter, and

432
00:25:53,800 --> 00:25:57,920
Speaker 1: Kate Osborne for Kaleidisco and Katrina nor velve iHeart Podcasts.

433
00:25:58,640 --> 00:26:02,400
Speaker 1: Jack Insley mixed this up. Kyle Murdoch wrote our theme song.

434
00:26:02,720 --> 00:26:04,920
Speaker 2: Join us on Friday for the Week in Tech, where

435
00:26:04,960 --> 00:26:06,560
Speaker 2: we'll run through the headlines.

436
00:26:06,119 --> 00:26:08,879
Speaker 1: You need to follow, and please do rate and review

437
00:26:08,920 --> 00:26:11,240
Speaker 1: the show and reach out to us at tech Stuff

438
00:26:11,280 --> 00:26:14,160
Speaker 1: podcast at gmail dot com. We want to hear from you.