1
00:00:01,600 --> 00:00:05,680
Speaker 1: Because he's dying. Okay, our guest has a bad cold,

2
00:00:06,160 --> 00:00:08,520
Speaker 1: and if he pauses, it's because he's dying. So we're

3
00:00:08,520 --> 00:00:11,360
Speaker 1: supposed to keep that in mind. The fabulous Jeffrey Jeffrey

4
00:00:11,400 --> 00:00:14,400
Speaker 1: Fowler of The Washington Post. He writes, there tech column

5
00:00:14,440 --> 00:00:17,880
Speaker 1: has a rhino virus. Ladies and gentlemen, Jeffrey, welcome and

6
00:00:17,880 --> 00:00:20,200
Speaker 1: none less. Sorry to hear you're feeling down. Thanks for

7
00:00:20,200 --> 00:00:23,080
Speaker 1: having I'm always feel better when I talk to you, guys.

8
00:00:23,079 --> 00:00:28,960
Speaker 1: You're very You're very kind. Better your your piece on Alexa,

9
00:00:29,200 --> 00:00:34,040
Speaker 1: Amazon's the various echoed devices listening in is amazing. It

10
00:00:34,200 --> 00:00:38,240
Speaker 1: is a blockbuster. Where do we begin? I thought it

11
00:00:38,360 --> 00:00:40,960
Speaker 1: only listened when we said Alexa and woke it up

12
00:00:40,960 --> 00:00:43,919
Speaker 1: and asked it to play the old ninety seven's. Well,

13
00:00:43,920 --> 00:00:47,000
Speaker 1: it turns out that Alexa listens lots of times. And

14
00:00:47,080 --> 00:00:48,559
Speaker 1: I think the big thing that I think a lot

15
00:00:48,600 --> 00:00:52,360
Speaker 1: of people don't realize is that Amazon keeps recordings of

16
00:00:52,560 --> 00:00:55,600
Speaker 1: every time one of these echo speakers or another device

17
00:00:55,600 --> 00:00:58,640
Speaker 1: that's Alexa in it, every time the microphone fires up,

18
00:00:58,840 --> 00:01:01,920
Speaker 1: and it stores those indefinitely. So what I did is

19
00:01:01,960 --> 00:01:04,440
Speaker 1: I went through I've had these devices in my life

20
00:01:04,440 --> 00:01:07,039
Speaker 1: for the last four years. So I went through my archives,

21
00:01:07,120 --> 00:01:10,880
Speaker 1: four years of snippets of me saying, you know, Alexa

22
00:01:11,080 --> 00:01:14,919
Speaker 1: sev spaghetti timer, Alexa, make a fart noise, Alexa, Um,

23
00:01:15,360 --> 00:01:17,399
Speaker 1: you know, tell me the weather. And I listened to

24
00:01:17,440 --> 00:01:19,959
Speaker 1: what was there, and I found thousands and thousands of

25
00:01:19,959 --> 00:01:22,000
Speaker 1: thousands of fragments of my life, some of which are

26
00:01:22,520 --> 00:01:25,119
Speaker 1: you know, totally you know you'd expect. Another stuff which

27
00:01:25,200 --> 00:01:28,959
Speaker 1: really surprised me, including dozens of times when the thing

28
00:01:29,000 --> 00:01:32,559
Speaker 1: it just went off on its own because it went rogue,

29
00:01:32,840 --> 00:01:35,080
Speaker 1: and it recorded when it when it went rogue, It recorded,

30
00:01:35,520 --> 00:01:39,440
Speaker 1: you know, my family talking about medication. It recorded uh,

31
00:01:39,680 --> 00:01:44,120
Speaker 1: random episodes of Downton Abbey. It recorded um, even a

32
00:01:44,160 --> 00:01:45,960
Speaker 1: friend who was sitting on my couch next to the

33
00:01:46,000 --> 00:01:49,520
Speaker 1: speaker working on a business deal and recorded her um

34
00:01:49,880 --> 00:01:52,960
Speaker 1: talking about that too. So the the big picture here

35
00:01:53,160 --> 00:01:56,200
Speaker 1: is that we don't really have a handle on when

36
00:01:56,200 --> 00:01:58,360
Speaker 1: these things are recording and the way that Amazon wants

37
00:01:58,440 --> 00:02:01,480
Speaker 1: us to think. And meanwhile Amazon is doing this crazy

38
00:02:01,520 --> 00:02:06,120
Speaker 1: thing of keeping all these recordings anyway into perpetuity. What's

39
00:02:06,160 --> 00:02:10,000
Speaker 1: the rationale for keeping the stuff? You? I think in

40
00:02:10,040 --> 00:02:12,480
Speaker 1: a lot of cases, and this supplies across the Board

41
00:02:12,480 --> 00:02:15,240
Speaker 1: and Silicon Valley. These companies want to take data. They

42
00:02:15,280 --> 00:02:17,200
Speaker 1: want to take our voices and a lot of other

43
00:02:17,360 --> 00:02:20,200
Speaker 1: data about things going on in our homes too, because

44
00:02:20,240 --> 00:02:22,640
Speaker 1: they can. There aren't really any laws against it. There

45
00:02:22,639 --> 00:02:26,720
Speaker 1: aren't really any industry best practices that people are following. Um,

46
00:02:26,760 --> 00:02:28,840
Speaker 1: So they're doing it in the hopes that maybe they'll

47
00:02:28,880 --> 00:02:31,000
Speaker 1: be able to benefit from it in the future. When

48
00:02:31,000 --> 00:02:34,960
Speaker 1: you ask these companies directly, they say, oh, with specifically

49
00:02:34,960 --> 00:02:38,160
Speaker 1: with the voice, it's to better train our voice algorithms

50
00:02:38,200 --> 00:02:42,320
Speaker 1: so that we understand more active and more different things.

51
00:02:42,360 --> 00:02:45,200
Speaker 1: But you know what, that's that's actually a false choice

52
00:02:45,280 --> 00:02:47,480
Speaker 1: because one of the things I discovered in reporting this

53
00:02:47,960 --> 00:02:52,720
Speaker 1: is that Amazon's art rival Google last year stopped automatically

54
00:02:52,800 --> 00:02:55,400
Speaker 1: keeping these recordings. Google said, you know what, we actually

55
00:02:55,440 --> 00:02:58,880
Speaker 1: don't need this data to make assistant work better. In arguably,

56
00:02:58,919 --> 00:03:01,639
Speaker 1: Assistant is better than a Alexa in many ways. So

57
00:03:02,040 --> 00:03:04,400
Speaker 1: Amazon right now is giving his customers really a false,

58
00:03:04,480 --> 00:03:08,239
Speaker 1: false choice between privacy and function. How did you get

59
00:03:08,280 --> 00:03:11,440
Speaker 1: access to this stuff? And I think of that because

60
00:03:11,480 --> 00:03:16,040
Speaker 1: my son has his Alexa in his room, and uh,

61
00:03:16,320 --> 00:03:18,000
Speaker 1: you know, I don't have a way to check the

62
00:03:18,000 --> 00:03:20,160
Speaker 1: search history like I do on a computer. It'd be

63
00:03:20,240 --> 00:03:22,840
Speaker 1: kind of interesting to hear what all things he's Alexa

64
00:03:22,960 --> 00:03:25,680
Speaker 1: to do on his own. Here's the crazy part. You

65
00:03:25,800 --> 00:03:28,000
Speaker 1: do have access to it. I think just most people

66
00:03:28,040 --> 00:03:30,560
Speaker 1: don't realize it. If you've got the Alexa app, or

67
00:03:30,600 --> 00:03:32,680
Speaker 1: you go to a web browser, you pull up UM

68
00:03:32,720 --> 00:03:35,520
Speaker 1: your Alexa privacy settings. So I've got links and instructions

69
00:03:35,680 --> 00:03:37,920
Speaker 1: and how to do all this stuff in my Washington

70
00:03:37,920 --> 00:03:39,800
Speaker 1: Post call and we will have a link to your

71
00:03:39,920 --> 00:03:42,200
Speaker 1: fabulous piece at Armstrong and Getty dot com. Go on,

72
00:03:42,280 --> 00:03:45,360
Speaker 1: please cool. You can go through and um and listen

73
00:03:45,560 --> 00:03:48,280
Speaker 1: to the recordings. They also Amazon also gives you the

74
00:03:48,360 --> 00:03:52,120
Speaker 1: power to delete those recordings, but you can't tell it

75
00:03:52,200 --> 00:03:56,000
Speaker 1: to stop recording again in the future. The stuff you've

76
00:03:56,000 --> 00:03:58,800
Speaker 1: got that is really interesting. I am not going to

77
00:03:58,800 --> 00:04:02,400
Speaker 1: tell my son that I have everything that he searches

78
00:04:02,480 --> 00:04:05,080
Speaker 1: on for the rest of his life, so but that's cool.

79
00:04:05,160 --> 00:04:08,560
Speaker 1: Jeffrey Fowler is the Washington Post technology columnists. Jeff we

80
00:04:08,640 --> 00:04:10,640
Speaker 1: ever tell you about the time I discovered that I

81
00:04:10,640 --> 00:04:13,920
Speaker 1: could see what windows my tabs my son had open

82
00:04:14,000 --> 00:04:17,760
Speaker 1: on his iPhone. I could see his searches on the Internet,

83
00:04:18,800 --> 00:04:20,320
Speaker 1: and I called it. He's in the air and told

84
00:04:20,360 --> 00:04:22,479
Speaker 1: him and he's a college kid. Oh yeah. At the

85
00:04:22,520 --> 00:04:25,040
Speaker 1: time he was yeah, young college man, that was college

86
00:04:25,040 --> 00:04:27,240
Speaker 1: boys search history. But then it occurred to me he

87
00:04:27,240 --> 00:04:29,200
Speaker 1: could see mine as well, and it's it was hard

88
00:04:29,200 --> 00:04:32,520
Speaker 1: to say was more horrified. But you know, this is

89
00:04:32,560 --> 00:04:35,760
Speaker 1: a really good example of this moment we're in with technology. Right.

90
00:04:35,800 --> 00:04:38,800
Speaker 1: We love the conveniences and the capabilities of these things,

91
00:04:38,800 --> 00:04:41,240
Speaker 1: but we haven't really figured out in our minds the

92
00:04:41,320 --> 00:04:43,920
Speaker 1: right way to think about how they're following us and

93
00:04:44,040 --> 00:04:47,440
Speaker 1: who all might get to hear or see or look

94
00:04:47,480 --> 00:04:50,359
Speaker 1: at what we're doing. Well, if we always assume the

95
00:04:50,400 --> 00:04:52,960
Speaker 1: worst would be pretty close to right, wouldn't we. Well,

96
00:04:52,960 --> 00:04:54,640
Speaker 1: then we all have to live in a you know,

97
00:04:54,720 --> 00:04:57,719
Speaker 1: in a bunker underground. Well you just have to. You

98
00:04:57,800 --> 00:04:59,680
Speaker 1: just have to accept it. But if you assume their

99
00:04:59,720 --> 00:05:03,440
Speaker 1: list and cataloging it and or or taking your information

100
00:05:03,560 --> 00:05:06,920
Speaker 1: selling it, it's almost always the case. Well, here's the thing.

101
00:05:06,960 --> 00:05:10,160
Speaker 1: I'm a tech optivist. I actually think that data can

102
00:05:10,279 --> 00:05:13,760
Speaker 1: make our homes run better, more efficiently, more power efficient.

103
00:05:14,000 --> 00:05:16,479
Speaker 1: I think having a voice assistant, uh, you know, computer

104
00:05:16,520 --> 00:05:18,120
Speaker 1: that's around you all the time, kind of like on

105
00:05:18,200 --> 00:05:20,560
Speaker 1: Star Trek. I think that's actually a really useful thing.

106
00:05:20,800 --> 00:05:23,880
Speaker 1: I think just that these companies right now are so

107
00:05:24,000 --> 00:05:26,240
Speaker 1: used to doing like Facebook was able to do and

108
00:05:26,279 --> 00:05:29,000
Speaker 1: get away with murder, that they're just going on these

109
00:05:29,000 --> 00:05:32,560
Speaker 1: wild data graphs to collect all this information and they shouldn't.

110
00:05:32,640 --> 00:05:34,520
Speaker 1: And we need to kind of have you know, we

111
00:05:34,560 --> 00:05:36,800
Speaker 1: need to revolt as consumers and say this isn't right.

112
00:05:37,080 --> 00:05:39,360
Speaker 1: So we need to make different choices. In the case

113
00:05:39,440 --> 00:05:42,240
Speaker 1: of voice assistance, you could you have choices with Google

114
00:05:42,360 --> 00:05:44,920
Speaker 1: or Apple, um. And we need to then maybe think

115
00:05:44,960 --> 00:05:48,320
Speaker 1: about laws that, you know, stop some of this. In fact,

116
00:05:48,600 --> 00:05:52,159
Speaker 1: California right now, the Assembly is debating a bill that

117
00:05:52,200 --> 00:05:55,400
Speaker 1: would stop Amazon from keeping all those recordings automatically. They

118
00:05:55,480 --> 00:05:57,160
Speaker 1: have to get our permission before they do it. Well

119
00:05:57,200 --> 00:06:00,320
Speaker 1: they should. It's like the wild West of technology. That

120
00:06:00,400 --> 00:06:02,920
Speaker 1: was ironic. I'm going to run for Congress just like

121
00:06:02,920 --> 00:06:05,840
Speaker 1: this about cliches, this stay in practice sometimes. Hey, one

122
00:06:05,880 --> 00:06:08,200
Speaker 1: thing you point out in your peace Jeffrey, is that, um,

123
00:06:08,440 --> 00:06:11,360
Speaker 1: it's it's far from just the Amazon Echo or whatever.

124
00:06:11,560 --> 00:06:15,640
Speaker 1: We're all being encouraged to surround ourselves with that sort

125
00:06:15,680 --> 00:06:19,440
Speaker 1: of device smart doorbells, security cameras, all sort of it's

126
00:06:19,440 --> 00:06:23,360
Speaker 1: all the refrigerators for God's sake. Yeah, totally. And this

127
00:06:23,360 --> 00:06:25,440
Speaker 1: is where the real wild West comes in, because again,

128
00:06:25,480 --> 00:06:27,520
Speaker 1: I'm a tech column. This is my job to try

129
00:06:27,520 --> 00:06:29,599
Speaker 1: all this new stuff and live with it and report

130
00:06:29,640 --> 00:06:32,200
Speaker 1: backs from the future about what it's all like and

131
00:06:32,360 --> 00:06:34,800
Speaker 1: what I discovered when I went down the same past,

132
00:06:34,839 --> 00:06:36,640
Speaker 1: the same rabbit hole of trying to figure out what

133
00:06:36,640 --> 00:06:39,760
Speaker 1: what data were those things collecting? What I found is

134
00:06:39,800 --> 00:06:43,440
Speaker 1: that most of them were keeping the data about what's

135
00:06:43,480 --> 00:06:46,440
Speaker 1: going on in my home indefinitely. So for example, uh,

136
00:06:46,560 --> 00:06:48,719
Speaker 1: I have a Nest thermostat. Well, it turns out the

137
00:06:48,760 --> 00:06:52,320
Speaker 1: Nest thermostat checks back in with its maker, which is Google,

138
00:06:52,520 --> 00:06:55,760
Speaker 1: every fifteen minutes and reports not only the temperature in

139
00:06:55,800 --> 00:06:58,840
Speaker 1: your house, but also whether a human being walked in

140
00:06:58,920 --> 00:07:02,039
Speaker 1: front of the thermostat that during that period. So going

141
00:07:02,080 --> 00:07:04,920
Speaker 1: back for years and years and years, Google knows every

142
00:07:04,920 --> 00:07:07,480
Speaker 1: time I get up for a midnight snack because actually

143
00:07:07,640 --> 00:07:11,720
Speaker 1: passed by the thermostat. That's weird, that's breaky. I'm not

144
00:07:11,760 --> 00:07:14,960
Speaker 1: sure it matters, but I think it probably does. Back

145
00:07:14,960 --> 00:07:18,520
Speaker 1: to the yeah, back, yeah, back to the Amazon thing.

146
00:07:18,560 --> 00:07:20,880
Speaker 1: Did you see me rhyme and rhymer reason when it

147
00:07:21,040 --> 00:07:24,320
Speaker 1: switched on and started recording like your friend talking about business?

148
00:07:24,320 --> 00:07:27,200
Speaker 1: Are you talking about medicine? It was really weird. I

149
00:07:27,280 --> 00:07:29,280
Speaker 1: could not figure it out. It just seemed like sometimes

150
00:07:29,280 --> 00:07:32,320
Speaker 1: there are things that that that that set it off

151
00:07:32,360 --> 00:07:34,560
Speaker 1: to go to go rogue. Now Amazon told me that

152
00:07:34,640 --> 00:07:38,559
Speaker 1: they have improved UM the accuracy of the wake word

153
00:07:38,640 --> 00:07:41,960
Speaker 1: ALEXA like fift in the last year. UM. I have

154
00:07:42,000 --> 00:07:43,800
Speaker 1: no way to verify with it. That's treat what they

155
00:07:43,800 --> 00:07:46,240
Speaker 1: say that they're they're working on it. UM. But the

156
00:07:46,680 --> 00:07:49,240
Speaker 1: other element of this, but I think it has gotten

157
00:07:49,240 --> 00:07:51,320
Speaker 1: a lot of people going, you know, with their eyebrows

158
00:07:51,360 --> 00:07:53,280
Speaker 1: really up is We've learned in the last month as

159
00:07:53,320 --> 00:07:56,880
Speaker 1: well that Amazon has employees who actually listened to some

160
00:07:56,960 --> 00:08:03,240
Speaker 1: of these recordings. Could couldn't Could an Amazon employee figure out, hey,

161
00:08:03,400 --> 00:08:05,400
Speaker 1: Jack Armstrong's got one of these things, I want to

162
00:08:05,440 --> 00:08:07,240
Speaker 1: see what he's doing in the afternoon and flip it on.

163
00:08:07,320 --> 00:08:09,640
Speaker 1: You think they have access to do that. I think

164
00:08:09,800 --> 00:08:13,600
Speaker 1: doing that remotely would be difficult. But I have not

165
00:08:13,680 --> 00:08:16,160
Speaker 1: seen any evidence that it has happened. But but it

166
00:08:16,200 --> 00:08:19,000
Speaker 1: could been theory. Anything that's connected to the internet could

167
00:08:19,040 --> 00:08:22,400
Speaker 1: be UM. But you know, Amazon has until now sort

168
00:08:22,440 --> 00:08:25,960
Speaker 1: of rested on the the the the argument that, look,

169
00:08:26,040 --> 00:08:27,760
Speaker 1: we've got a button on the top of this thing,

170
00:08:27,800 --> 00:08:30,200
Speaker 1: and you can press and it turns the little light

171
00:08:30,320 --> 00:08:32,520
Speaker 1: ring red, and that means we're not recording anymore. As

172
00:08:32,520 --> 00:08:34,960
Speaker 1: far as I understand, that still does work. But of

173
00:08:35,000 --> 00:08:37,800
Speaker 1: course like that kills the function of the device, So

174
00:08:37,920 --> 00:08:42,040
Speaker 1: again we're given this sort of fall. Yeah, it's kind

175
00:08:42,040 --> 00:08:43,720
Speaker 1: of like the if you bothered by don't use Google.

176
00:08:43,720 --> 00:08:47,720
Speaker 1: Well that's pretty unrealistic. Um. Uh so you you right

177
00:08:47,760 --> 00:08:50,200
Speaker 1: for the Washington Post, which is owned by the guy

178
00:08:50,200 --> 00:08:52,160
Speaker 1: who owns Amazon. Do you get any pushback on it

179
00:08:52,240 --> 00:08:57,000
Speaker 1: into stuff? Absolutely not. The review all technology with the

180
00:08:57,040 --> 00:09:00,880
Speaker 1: same critical and often snarky if you don't wait, well

181
00:09:00,880 --> 00:09:03,880
Speaker 1: we barely have forty seconds for this, Jeffrey. But what

182
00:09:03,920 --> 00:09:07,520
Speaker 1: do you think of Google's big to do yesterday? Yeah?

183
00:09:07,559 --> 00:09:10,360
Speaker 1: It was Google. I oh, that's their big annual conference. Um.

184
00:09:10,400 --> 00:09:12,800
Speaker 1: Two thoughts. One, you know, going to these things and

185
00:09:13,040 --> 00:09:15,160
Speaker 1: hearing Google talking about all their fancy new products, you'd

186
00:09:15,200 --> 00:09:17,720
Speaker 1: never know they were an advertising company takes all of

187
00:09:17,720 --> 00:09:21,280
Speaker 1: its money based on you know, mining our data to

188
00:09:21,360 --> 00:09:23,040
Speaker 1: try to sell us ads. They didn't bring that up

189
00:09:23,080 --> 00:09:26,080
Speaker 1: at all. Um, they talked yep, they talked a lot

190
00:09:26,120 --> 00:09:28,679
Speaker 1: about privacy though, and they said that essentially in certain

191
00:09:28,679 --> 00:09:31,079
Speaker 1: ways they've got religion on it, and they're working on

192
00:09:31,120 --> 00:09:35,680
Speaker 1: things to do processing on devices instead of in their servers.

193
00:09:35,760 --> 00:09:38,600
Speaker 1: So um, clearly the message is now the Silicon Valley

194
00:09:38,640 --> 00:09:40,920
Speaker 1: that like, we're not happy about the state of things,

195
00:09:40,920 --> 00:09:45,880
Speaker 1: will will assume you're lying till further notice. Google. Jeffrey Fowler,

196
00:09:46,000 --> 00:09:48,440
Speaker 1: the Washington Post tech columnists. We have a link to

197
00:09:48,520 --> 00:09:50,640
Speaker 1: his great piece at Armstrong and Gutty dot com, which

198
00:09:50,679 --> 00:09:53,559
Speaker 1: has a link to listen to what Alexa has been

199
00:09:53,600 --> 00:09:57,600
Speaker 1: listening to in your home. Jeffrey, you are so good, terrific.

200
00:09:57,760 --> 00:09:58,559
Speaker 1: Nice job. Thanks