1
00:00:04,240 --> 00:00:07,240
Speaker 1: Welcome to tech Stuff, a production of I Heart Radios,

2
00:00:07,320 --> 00:00:14,000
Speaker 1: How Stuff Works. Hey there, and welcome to tech Stuff.

3
00:00:14,040 --> 00:00:16,720
Speaker 1: I'm your host, Jonathan Strickland. I'm an executive producer with

4
00:00:16,760 --> 00:00:18,560
Speaker 1: How Stuff Works and I Heart Radio and I Love

5
00:00:18,640 --> 00:00:22,680
Speaker 1: all Things Tech. And Oz and Kara from Sleepwalkers have

6
00:00:23,160 --> 00:00:25,799
Speaker 1: come back to join me yet again. Spoiler alert. We

7
00:00:25,840 --> 00:00:28,000
Speaker 1: actually haven't left anywhere. We just sat here the whole

8
00:00:28,040 --> 00:00:30,319
Speaker 1: time and we're recording two episodes back to back. But

9
00:00:30,680 --> 00:00:33,839
Speaker 1: if you haven't heard our previous discussion, which is kind

10
00:00:33,840 --> 00:00:37,199
Speaker 1: of a high level discussion about AI and the potential

11
00:00:37,280 --> 00:00:41,000
Speaker 1: dangers and sort of the the various messaging we've received

12
00:00:41,000 --> 00:00:43,879
Speaker 1: about AI and the warning signs and the promises, you

13
00:00:43,880 --> 00:00:46,680
Speaker 1: should listen to that episode first. If you haven't subscribed

14
00:00:46,720 --> 00:00:49,839
Speaker 1: to Sleepwalkers, you should absolutely do that too, because the

15
00:00:49,840 --> 00:00:53,200
Speaker 1: show is amazing. And today we're gonna talk a little

16
00:00:53,200 --> 00:00:56,800
Speaker 1: bit more about how different parts of the world are

17
00:00:56,840 --> 00:01:01,200
Speaker 1: treating AI, whether it's from a government perfective or a

18
00:01:01,240 --> 00:01:06,280
Speaker 1: business perspective, The technological you know, development of AI, where

19
00:01:06,360 --> 00:01:10,760
Speaker 1: is that actually happening the fastest, And the answer to

20
00:01:10,800 --> 00:01:14,600
Speaker 1: that might surprise you if you haven't been paying close

21
00:01:14,640 --> 00:01:18,280
Speaker 1: attention to news around the world and we're going to

22
00:01:18,400 --> 00:01:21,640
Speaker 1: dive into all of that. So, without any further Ado,

23
00:01:22,080 --> 00:01:25,280
Speaker 1: welcome back to the show, Os and Kara, Thank you

24
00:01:25,319 --> 00:01:29,640
Speaker 1: so much. Jonathan, Hi Jonathan again. Hi again. Yea. And

25
00:01:30,080 --> 00:01:33,200
Speaker 1: while we were between shows, we were just you know,

26
00:01:33,840 --> 00:01:37,319
Speaker 1: talking about Lady Gaga and Shakespeare as you are want

27
00:01:37,400 --> 00:01:40,360
Speaker 1: to do. Uh, this is kind of what we technologists

28
00:01:40,440 --> 00:01:44,039
Speaker 1: technology podcasters tend to to really, you know, kind of

29
00:01:44,080 --> 00:01:47,080
Speaker 1: revel in when we're not on on Mike or at

30
00:01:47,120 --> 00:01:51,200
Speaker 1: least not recording. But I wanted to talk first about

31
00:01:52,400 --> 00:01:59,880
Speaker 1: where do we see really aggressive you know, moving forward

32
00:02:00,040 --> 00:02:02,720
Speaker 1: on on technology and AI, Like, where are we seeing

33
00:02:02,720 --> 00:02:07,960
Speaker 1: the most development in AI? Because a lot of people

34
00:02:08,000 --> 00:02:10,560
Speaker 1: think of Silicon Valley as sort of the place like

35
00:02:10,639 --> 00:02:15,280
Speaker 1: it's it's the breeding ground for all technologies. But as

36
00:02:15,320 --> 00:02:18,200
Speaker 1: it turns out, that's that's really a very narrow view,

37
00:02:18,960 --> 00:02:24,680
Speaker 1: and it it's ignoring a giant superpower that is pouring

38
00:02:24,720 --> 00:02:33,399
Speaker 1: a lot of resources into AI development. Right China. Yeah, sorry,

39
00:02:33,440 --> 00:02:37,400
Speaker 1: that China is is um is leading the charge on AI.

40
00:02:37,480 --> 00:02:39,720
Speaker 1: And one of the guests we had on the show

41
00:02:39,880 --> 00:02:44,280
Speaker 1: on Sleepwalkers is a guy called Kaifu Lee who was

42
00:02:44,320 --> 00:02:47,160
Speaker 1: part of the team at Apple that developed the technology

43
00:02:47,280 --> 00:02:50,920
Speaker 1: behind SyRI in the nineties, went on to run Google

44
00:02:51,040 --> 00:02:54,320
Speaker 1: China and is now one of the biggest technology investors

45
00:02:54,320 --> 00:02:57,600
Speaker 1: in China through a fund called Signivation Ventures, which is

46
00:02:57,600 --> 00:03:01,920
Speaker 1: about all sorts of different technology ventures in China, including

47
00:03:01,960 --> 00:03:04,720
Speaker 1: several unicorns a billion dollar companies, one of which is

48
00:03:05,080 --> 00:03:08,400
Speaker 1: called meg v which does facial recognition technology. Kai Fu

49
00:03:08,480 --> 00:03:13,200
Speaker 1: Lee recently wrote a book called AI Superpowers, China, Silicon

50
00:03:13,280 --> 00:03:16,480
Speaker 1: Valley and the New World Order. New World Order is

51
00:03:16,560 --> 00:03:20,239
Speaker 1: quite a resonant phrase, shall we say, But the thesis

52
00:03:20,320 --> 00:03:23,200
Speaker 1: of the book is simply that China are doing AI

53
00:03:23,360 --> 00:03:26,040
Speaker 1: a lot better, a lot more aggressively, and with a

54
00:03:26,080 --> 00:03:29,520
Speaker 1: lot more promise than we are in the US. And

55
00:03:29,560 --> 00:03:33,520
Speaker 1: that's for two reasons. Number one, China, unlike the US,

56
00:03:33,760 --> 00:03:40,080
Speaker 1: is a centralized country, a centralized government who have said absolutely,

57
00:03:40,480 --> 00:03:44,600
Speaker 1: with no hesitation, AI is our biggest priority. In seventeen,

58
00:03:44,640 --> 00:03:47,839
Speaker 1: the Communist Party released what they called the New Generation

59
00:03:48,080 --> 00:03:52,960
Speaker 1: Artificial Intelligence Development Plan, and the first paragraph read AI

60
00:03:53,040 --> 00:03:57,000
Speaker 1: has become a new focus of international competition. AI is

61
00:03:57,000 --> 00:04:00,280
Speaker 1: a strategic technology that will lead the future. The was

62
00:04:00,400 --> 00:04:03,320
Speaker 1: major developed countries are taking the development of AI as

63
00:04:03,320 --> 00:04:09,640
Speaker 1: a major strategy to enhance national competitiveness and protect national security.

64
00:04:09,720 --> 00:04:11,840
Speaker 1: So China have been on this for two years, I

65
00:04:11,880 --> 00:04:14,360
Speaker 1: mean for longer than two years, but they're taking immensity

66
00:04:14,440 --> 00:04:18,279
Speaker 1: seriously and we're not. We did have our own presidential

67
00:04:18,320 --> 00:04:20,880
Speaker 1: executive order in April of this year, but it didn't

68
00:04:20,880 --> 00:04:23,600
Speaker 1: come with any funding a kaifu. Second point, which I'm

69
00:04:23,600 --> 00:04:25,720
Speaker 1: sure we'll get onto, is that the China also have

70
00:04:25,760 --> 00:04:28,880
Speaker 1: a much richer data set, which is the power behind

71
00:04:28,920 --> 00:04:33,040
Speaker 1: the throne of AI. Yeah, something that I think a

72
00:04:33,360 --> 00:04:37,280
Speaker 1: lot of Americans in particular aren't aware of is that

73
00:04:38,000 --> 00:04:41,320
Speaker 1: when you look over at China and China's efforts to

74
00:04:42,160 --> 00:04:45,680
Speaker 1: to own AI essentially, I mean there there's they're saying,

75
00:04:46,279 --> 00:04:51,440
Speaker 1: we are going to be established as the the primary

76
00:04:52,200 --> 00:04:55,760
Speaker 1: source of AI development by twenty thirty and they're they're

77
00:04:55,760 --> 00:04:58,919
Speaker 1: well on their way to doing that already. Um. You

78
00:04:58,960 --> 00:05:02,239
Speaker 1: look at the the B A T that's the three

79
00:05:02,400 --> 00:05:05,599
Speaker 1: big companies that collectively are valued at more than a

80
00:05:05,640 --> 00:05:09,479
Speaker 1: trillion dollars, bay Do, Ali, Baba and ten Cent, and

81
00:05:10,200 --> 00:05:15,560
Speaker 1: those are already enormous corporations that have deep ties to

82
00:05:16,440 --> 00:05:19,920
Speaker 1: the state government of China than that are working very

83
00:05:19,920 --> 00:05:22,440
Speaker 1: hard in AI. But beyond that, even though you know,

84
00:05:22,480 --> 00:05:24,680
Speaker 1: you might think, well, those are maybe three big companies,

85
00:05:24,680 --> 00:05:28,440
Speaker 1: but how is that that that big of a of

86
00:05:28,120 --> 00:05:30,800
Speaker 1: a of a you know, a force on its own.

87
00:05:31,560 --> 00:05:37,920
Speaker 1: They also invest heavily in lots of startup companies, including unicorns. Uh.

88
00:05:37,960 --> 00:05:42,480
Speaker 1: They in fact, they out of the six hundred billion

89
00:05:42,560 --> 00:05:45,440
Speaker 1: dollars of unicorns out of China from a couple of

90
00:05:45,520 --> 00:05:49,680
Speaker 1: years ago, they made up fifty percent of the investment

91
00:05:49,800 --> 00:05:54,080
Speaker 1: into those companies. Then beyond that, they're investing in companies

92
00:05:54,279 --> 00:05:56,719
Speaker 1: outside of China and in other parts of the world,

93
00:05:56,760 --> 00:06:00,400
Speaker 1: including the United States. Uh. Ten Cent, for example, has

94
00:06:00,480 --> 00:06:05,960
Speaker 1: been investing heavily in video games over in the West Coast,

95
00:06:06,800 --> 00:06:10,560
Speaker 1: I'm along with lots of other different companies out there.

96
00:06:10,680 --> 00:06:15,760
Speaker 1: So not only are you talking about a country that

97
00:06:15,839 --> 00:06:19,840
Speaker 1: has an enormous population and therefore an enormous source of

98
00:06:19,880 --> 00:06:24,279
Speaker 1: information on its own, it has spread out around the globe,

99
00:06:24,360 --> 00:06:29,600
Speaker 1: so it's gathering information from everywhere. So it's it's really

100
00:06:29,880 --> 00:06:34,520
Speaker 1: this incredibly pervasive system to gather the fuel that is

101
00:06:34,560 --> 00:06:38,680
Speaker 1: going to power artificial intelligence. Uh. And meanwhile, you also

102
00:06:38,800 --> 00:06:45,480
Speaker 1: have very smart people running very sophisticated laboratories working on

103
00:06:45,520 --> 00:06:49,120
Speaker 1: the next generation of of algorithms and applications of AI.

104
00:06:49,240 --> 00:06:52,520
Speaker 1: So you've got like the perfect storm over there. Yeah. Absolutely.

105
00:06:52,600 --> 00:06:55,120
Speaker 1: And I also think, I mean, just in terms of

106
00:06:55,200 --> 00:06:57,880
Speaker 1: what you're talking about, you know, Tencent being involved in

107
00:06:57,960 --> 00:07:01,240
Speaker 1: Fortnite if I'm correct, Yes, yeah, I was just doing

108
00:07:01,520 --> 00:07:04,400
Speaker 1: was doing the floss right over here. Actually, that's right.

109
00:07:05,080 --> 00:07:09,840
Speaker 1: And also the ownership of a little known dating app

110
00:07:09,880 --> 00:07:13,520
Speaker 1: called grinder Um, which the US is actually trying to

111
00:07:14,280 --> 00:07:18,200
Speaker 1: get back just because of the national security threat that

112
00:07:18,320 --> 00:07:24,119
Speaker 1: is involved in the Chinese owning uh so much user

113
00:07:24,200 --> 00:07:30,360
Speaker 1: data from United States citizens, including U S military personnel. Um.

114
00:07:30,840 --> 00:07:37,640
Speaker 1: I think I think what's alarming about this is that

115
00:07:37,960 --> 00:07:42,600
Speaker 1: we is that we're sort of entering into this new

116
00:07:42,680 --> 00:07:49,600
Speaker 1: territory of you know, war and competition and war being

117
00:07:49,680 --> 00:07:53,280
Speaker 1: less about you know, guns and bombs and more. And

118
00:07:53,520 --> 00:07:56,080
Speaker 1: this is not these are not my words I've read this,

119
00:07:57,120 --> 00:08:00,760
Speaker 1: but more about you know, bits and bites, so to speak.

120
00:08:00,840 --> 00:08:03,480
Speaker 1: I think you read that from Secret Ar State Mike Pompeo.

121
00:08:03,560 --> 00:08:07,480
Speaker 1: That's right, I'm quoting Mike Pompeio. Happily, I think, um,

122
00:08:07,640 --> 00:08:11,000
Speaker 1: which actually he said in regards to Huawei, which is

123
00:08:11,120 --> 00:08:14,520
Speaker 1: obviously in the news quite a bit recently. Um, you know.

124
00:08:14,600 --> 00:08:18,880
Speaker 1: I think that the there's there's the technology race, which is,

125
00:08:18,920 --> 00:08:22,680
Speaker 1: you know, sort of who's going to have who's going

126
00:08:22,720 --> 00:08:26,960
Speaker 1: to advance quickly and in the best way. But it's

127
00:08:27,000 --> 00:08:31,840
Speaker 1: also about sort of who owns data and who has

128
00:08:31,920 --> 00:08:35,199
Speaker 1: access to what data. I think the China, the Chinese

129
00:08:35,200 --> 00:08:39,960
Speaker 1: government does not really have a problem uh spying on

130
00:08:40,000 --> 00:08:42,720
Speaker 1: its citizens. I don't even think we'd call it spying necessarily.

131
00:08:42,760 --> 00:08:46,040
Speaker 1: I think it's sweeping, um is what is what they

132
00:08:46,080 --> 00:08:49,920
Speaker 1: tend to do, and um with that data can make

133
00:08:50,000 --> 00:08:55,400
Speaker 1: some pretty uh you know, chilling accusations about people who

134
00:08:55,600 --> 00:08:59,520
Speaker 1: they think are dissenting against the government. Um. You know,

135
00:08:59,559 --> 00:09:03,520
Speaker 1: in the United States, we are very free with the

136
00:09:03,600 --> 00:09:05,800
Speaker 1: data we give away because I don't think enough people

137
00:09:05,840 --> 00:09:08,360
Speaker 1: think about how much data they're giving away in any

138
00:09:08,400 --> 00:09:13,839
Speaker 1: given day. But the US government hopefully will remain so

139
00:09:14,080 --> 00:09:19,880
Speaker 1: is not as what's the word pervasive, pervasive and aggressive

140
00:09:19,920 --> 00:09:24,320
Speaker 1: about collecting said data and using it to um, you know,

141
00:09:24,400 --> 00:09:28,839
Speaker 1: in prison there citizens. So I don't know, I think

142
00:09:29,200 --> 00:09:31,560
Speaker 1: China can kind of get away with more and maybe

143
00:09:31,559 --> 00:09:35,520
Speaker 1: that's why the Chinese government is allowing Chinese businesses to

144
00:09:35,520 --> 00:09:38,600
Speaker 1: get away with more than the US government. Ill saying

145
00:09:38,600 --> 00:09:40,720
Speaker 1: it's worth pointing out in the US, you know, we

146
00:09:40,760 --> 00:09:44,280
Speaker 1: have this phenomenous surveillance capitalism, So there's a bunch of

147
00:09:44,280 --> 00:09:46,880
Speaker 1: big firms who take all the data we give them,

148
00:09:47,400 --> 00:09:50,400
Speaker 1: use it to model us, make predictions about us, and

149
00:09:50,480 --> 00:09:55,040
Speaker 1: sell us more stuff. In China, it's surveillance statecraft, and

150
00:09:55,960 --> 00:10:00,360
Speaker 1: the data is not siloed between Amazon, Google, face Book

151
00:10:00,600 --> 00:10:02,920
Speaker 1: and others. UM. It's tends to be in the hands

152
00:10:02,920 --> 00:10:05,960
Speaker 1: of the big technology companies in China, you know who,

153
00:10:06,120 --> 00:10:09,280
Speaker 1: given that they're not fully state owned, but much more

154
00:10:09,320 --> 00:10:11,760
Speaker 1: hand in pocket with the government than our technology companies

155
00:10:11,800 --> 00:10:15,400
Speaker 1: are share their collected data much more widely, which allows

156
00:10:15,400 --> 00:10:18,920
Speaker 1: them to make better predictions about what might happen and

157
00:10:19,400 --> 00:10:22,679
Speaker 1: be determinative about the outcome of their citizens. I do

158
00:10:22,800 --> 00:10:25,240
Speaker 1: think it's worth pointing out a lot of this conversation

159
00:10:25,320 --> 00:10:27,880
Speaker 1: about the difference between AI in China and the United

160
00:10:27,920 --> 00:10:30,920
Speaker 1: States that we have on this side of the Pacific

161
00:10:31,480 --> 00:10:38,040
Speaker 1: um is filtered through our liberal individual worldview, which states

162
00:10:38,200 --> 00:10:42,360
Speaker 1: that you know, free will, the ability to control one's

163
00:10:42,400 --> 00:10:46,200
Speaker 1: own outcomes, um, the you know, the the recognition of

164
00:10:46,200 --> 00:10:50,040
Speaker 1: oneself as an individual are the utmost goods, and that

165
00:10:50,120 --> 00:10:53,840
Speaker 1: worldview simply isn't shared throughout most of China. Now you

166
00:10:53,880 --> 00:10:56,760
Speaker 1: can say that's that will change, you know, society opens up.

167
00:10:57,000 --> 00:10:59,360
Speaker 1: You can say there's an inevitable progress, if you want

168
00:10:59,400 --> 00:11:01,640
Speaker 1: to call it, program us towards our worldview. But the

169
00:11:01,640 --> 00:11:04,040
Speaker 1: fact is that isn't the world view in China. And

170
00:11:04,160 --> 00:11:08,000
Speaker 1: many Chinese citizens are used to living in one on

171
00:11:08,000 --> 00:11:12,480
Speaker 1: one party states since um, since mal z Dong or

172
00:11:12,480 --> 00:11:15,000
Speaker 1: where it was in the nineteen forties, and so there's

173
00:11:15,000 --> 00:11:16,600
Speaker 1: a there's a there's a there's a sense of being

174
00:11:16,600 --> 00:11:22,840
Speaker 1: accustomed to the belief that harmony and the furthering of

175
00:11:22,960 --> 00:11:27,520
Speaker 1: the state's goals rise um rising is a rising tide

176
00:11:27,520 --> 00:11:31,240
Speaker 1: which raises all boats. And so you know, I think

177
00:11:31,240 --> 00:11:33,360
Speaker 1: we have this desire to see and of course it

178
00:11:33,400 --> 00:11:36,120
Speaker 1: has happened in China. There's been Channeman's Square, you know.

179
00:11:36,120 --> 00:11:38,600
Speaker 1: There there there have been organic protest movements. But we

180
00:11:38,679 --> 00:11:43,079
Speaker 1: do have the desire to impose our absolute belief in

181
00:11:43,120 --> 00:11:45,520
Speaker 1: the importance of the individual and free will on the

182
00:11:45,559 --> 00:11:48,120
Speaker 1: rest of the world. And you know, I'm not sure

183
00:11:48,280 --> 00:11:51,120
Speaker 1: that if you ask the average Chinese citizen, you know,

184
00:11:51,360 --> 00:11:56,080
Speaker 1: do you resent being surveiled when the most highest number

185
00:11:56,120 --> 00:11:58,360
Speaker 1: of people have been lifted out of poverty in the

186
00:11:58,360 --> 00:12:01,719
Speaker 1: fastest time of any country tree in history. I think

187
00:12:01,720 --> 00:12:04,680
Speaker 1: the answer might be, you know, perhaps we prefer more freedoms,

188
00:12:04,720 --> 00:12:08,200
Speaker 1: but this isn't the worst thing in the world. Also,

189
00:12:08,360 --> 00:12:12,080
Speaker 1: China is an ethnically largely ethnically homogeneous society. So for that,

190
00:12:12,160 --> 00:12:15,760
Speaker 1: for the average Han Chinese, that the trade off for this,

191
00:12:15,760 --> 00:12:19,880
Speaker 1: this lifting out of poverty and the national pride which

192
00:12:19,920 --> 00:12:22,800
Speaker 1: is which is absolutely on the rise in China, giving

193
00:12:22,800 --> 00:12:25,160
Speaker 1: a much stronger sense of national identity, I would argue

194
00:12:25,160 --> 00:12:27,240
Speaker 1: than we have here, the tradeoff may be worth it.

195
00:12:27,600 --> 00:12:31,680
Speaker 1: What's very chilling, Beyond chilling is the treatment of the

196
00:12:31,760 --> 00:12:35,120
Speaker 1: non Han Chinese in China, the tibet the Tibetans, and

197
00:12:35,200 --> 00:12:39,360
Speaker 1: specifically the Weakers, who are really the people who experienced

198
00:12:39,400 --> 00:12:43,040
Speaker 1: the hard end of this surveillance state in China. Hey, guys,

199
00:12:43,040 --> 00:12:46,280
Speaker 1: it's Jonathan from the future. I'm just popping into interrupt

200
00:12:46,400 --> 00:12:49,360
Speaker 1: here because as it turned out, we got so into

201
00:12:49,360 --> 00:12:52,600
Speaker 1: this conversation I totally forgot to put in a break.

202
00:12:53,080 --> 00:12:56,760
Speaker 1: So let's take a quick break even while Jonathan in

203
00:12:56,800 --> 00:12:59,800
Speaker 1: the past and ours and Kara continue their conversation. We'll

204
00:12:59,800 --> 00:13:09,880
Speaker 1: get back to that in just a second, Hey, guys,

205
00:13:10,040 --> 00:13:12,199
Speaker 1: Jonathan from the future. Again, we're going to get right

206
00:13:12,200 --> 00:13:15,680
Speaker 1: back into the episode. Oz was just talking about China

207
00:13:15,920 --> 00:13:19,680
Speaker 1: and it's use of technology and the creation of a

208
00:13:19,720 --> 00:13:24,160
Speaker 1: surveillance state, and we're gonna pick up with my response. Yeah,

209
00:13:24,200 --> 00:13:26,360
Speaker 1: and as we mentioned in the last episode, you know,

210
00:13:26,360 --> 00:13:29,240
Speaker 1: we were talking about about bias and how that can

211
00:13:30,160 --> 00:13:33,800
Speaker 1: be unintentionally inserted into a system and how that can

212
00:13:33,840 --> 00:13:37,800
Speaker 1: cause harm. But you could also intentionally create a biased

213
00:13:37,880 --> 00:13:43,559
Speaker 1: system specifically in order to keep tabs on particular populations

214
00:13:44,080 --> 00:13:49,840
Speaker 1: that that are, you know, minorities, and yeah, the and

215
00:13:49,920 --> 00:13:56,000
Speaker 1: obviously that could lead to truly horrific inhumane practices, leading

216
00:13:56,000 --> 00:13:59,839
Speaker 1: all the way up to even genocide. Yes, I mean,

217
00:14:00,559 --> 00:14:05,400
Speaker 1: right now what we're seeing is sort of um imprisonment

218
00:14:05,559 --> 00:14:08,440
Speaker 1: in re education camps. But I think it's important to

219
00:14:08,440 --> 00:14:11,880
Speaker 1: note and this is again, you know, piggybacking off of

220
00:14:11,880 --> 00:14:14,400
Speaker 1: what Oz was saying. You know, the Chinese Communist Party

221
00:14:14,640 --> 00:14:18,320
Speaker 1: has often used surveillance as a means for control. The

222
00:14:18,320 --> 00:14:21,960
Speaker 1: differences is that artificial intelligence enables a kind of surveillance

223
00:14:21,960 --> 00:14:24,800
Speaker 1: that we haven't seen before, which is, you know, in China,

224
00:14:25,160 --> 00:14:32,760
Speaker 1: there is basically a very large operation that they call Egypt,

225
00:14:32,800 --> 00:14:37,960
Speaker 1: which is the Integrated Joint up Integrated Joint Operations. And

226
00:14:38,000 --> 00:14:41,480
Speaker 1: what that does is it's a database that is sweeping

227
00:14:41,560 --> 00:14:47,800
Speaker 1: information from you know, basically every source imaginable WiFi, you know,

228
00:14:47,960 --> 00:14:50,840
Speaker 1: visitor management system so you know, for example, in America,

229
00:14:50,880 --> 00:14:52,640
Speaker 1: what we call you know, when you walk into a

230
00:14:52,640 --> 00:14:56,560
Speaker 1: building and register your name to visit an office. You know, um,

231
00:14:56,680 --> 00:15:01,440
Speaker 1: wecha um we chat conversations, you know, uh, when you

232
00:15:01,520 --> 00:15:03,360
Speaker 1: leave the country, when you come back in the country,

233
00:15:03,560 --> 00:15:08,720
Speaker 1: all of these things are being swept into um a

234
00:15:08,840 --> 00:15:11,200
Speaker 1: larger I don't I don't know what the word is

235
00:15:11,200 --> 00:15:12,800
Speaker 1: for it. It's not server, I don't know how to

236
00:15:12,840 --> 00:15:15,600
Speaker 1: say it. It's a system. It's a system. It's a

237
00:15:15,600 --> 00:15:20,160
Speaker 1: system that it's then making decisions and predictions about who

238
00:15:20,240 --> 00:15:24,240
Speaker 1: is basically doing right and doing wrong. And we just

239
00:15:24,360 --> 00:15:27,280
Speaker 1: simply don't have We don't have anything like that in

240
00:15:27,280 --> 00:15:29,680
Speaker 1: the United States right now. I mean, certainly we can

241
00:15:30,200 --> 00:15:33,000
Speaker 1: there there are companies that can use our information in

242
00:15:33,040 --> 00:15:35,640
Speaker 1: the United States to make predictions about us, to sell

243
00:15:35,720 --> 00:15:40,160
Speaker 1: us things, to basically set our insurance premiums. But as

244
00:15:40,240 --> 00:15:45,240
Speaker 1: far as just an integrated system that is making decisions

245
00:15:45,280 --> 00:15:48,960
Speaker 1: about its people and then also using it to imprison

246
00:15:49,000 --> 00:15:52,200
Speaker 1: its people. There's it's just unprecedented. And you know, the

247
00:15:52,280 --> 00:15:56,960
Speaker 1: Human Rights Watch is basically calling it a humanitarian crisis,

248
00:15:56,960 --> 00:15:58,480
Speaker 1: which I which I think it is when you think

249
00:15:58,520 --> 00:16:02,400
Speaker 1: about you know, um, your own country basically spying on

250
00:16:02,440 --> 00:16:07,040
Speaker 1: your conversations. Um, really, I mean obviously without your consent.

251
00:16:07,080 --> 00:16:09,400
Speaker 1: People are stopped all the time in China and their

252
00:16:09,400 --> 00:16:11,840
Speaker 1: phones are taken and read through like that's a very

253
00:16:11,880 --> 00:16:15,440
Speaker 1: that's a sort of normal day, um, and then using

254
00:16:15,480 --> 00:16:18,480
Speaker 1: that information to put you in a re education camp.

255
00:16:18,520 --> 00:16:21,760
Speaker 1: I mean, I think if most Americans knew that, which

256
00:16:21,760 --> 00:16:26,560
Speaker 1: I don't think they do, um, they would be alarmed.

257
00:16:26,920 --> 00:16:30,040
Speaker 1: But let's be clear. You know, we're we're very far

258
00:16:30,200 --> 00:16:33,120
Speaker 1: from having a re education camps in the US, but

259
00:16:33,240 --> 00:16:35,960
Speaker 1: we do, and it's not on a state level or

260
00:16:36,120 --> 00:16:39,640
Speaker 1: a national level. We do make decisions about people's outcomes

261
00:16:39,640 --> 00:16:42,800
Speaker 1: with stuff like credit schools. I mean the credit schools.

262
00:16:42,840 --> 00:16:47,680
Speaker 1: You know in China it's it's it's ethnically it's explicitly ethnic. Right.

263
00:16:47,880 --> 00:16:50,560
Speaker 1: Well again in China they wouldn't say that, but effectively

264
00:16:50,560 --> 00:16:53,680
Speaker 1: it's explicitly ethnic against the wagers. In the US, we

265
00:16:53,720 --> 00:16:56,160
Speaker 1: have credit schools and guess what most people who grow

266
00:16:56,200 --> 00:16:58,240
Speaker 1: up with a certain amount of privilege know what their

267
00:16:58,240 --> 00:17:01,800
Speaker 1: credit school is, understand the principle of the credit school.

268
00:17:02,080 --> 00:17:04,080
Speaker 1: Get a credit card as early as they can, start

269
00:17:04,080 --> 00:17:06,439
Speaker 1: building their credit, making monthly payoffs, and then when it

270
00:17:06,480 --> 00:17:08,760
Speaker 1: comes time to get a mortgage and buy a house

271
00:17:08,800 --> 00:17:11,359
Speaker 1: and move to a nice neighborhood, guess what, All the

272
00:17:11,359 --> 00:17:13,720
Speaker 1: pieces are in place. But but for for many people

273
00:17:13,720 --> 00:17:15,840
Speaker 1: who don't grow up with the privilege, that the credit

274
00:17:15,880 --> 00:17:18,320
Speaker 1: school comes as a complete surprise at a certain point

275
00:17:18,359 --> 00:17:20,480
Speaker 1: in life that you know, there's even notion of having

276
00:17:20,560 --> 00:17:22,840
Speaker 1: bad credit. All of a sudden, you know you've got

277
00:17:22,880 --> 00:17:26,119
Speaker 1: bad credit without ever having realized that you had this

278
00:17:26,160 --> 00:17:28,280
Speaker 1: credit school you're supposed to be working on. And guess what.

279
00:17:28,600 --> 00:17:30,400
Speaker 1: You can't move out of your neighborhood, You can't get

280
00:17:30,440 --> 00:17:32,720
Speaker 1: what you want, You can't buy your children and things

281
00:17:32,720 --> 00:17:35,280
Speaker 1: they need. So we do have this here. It's not

282
00:17:35,400 --> 00:17:41,040
Speaker 1: state policy, but we've we've effectually outsourced this predictive technology

283
00:17:41,080 --> 00:17:42,760
Speaker 1: about what people are going to do in the future

284
00:17:42,800 --> 00:17:45,600
Speaker 1: to private corporations who who use it to profit from

285
00:17:45,720 --> 00:17:48,520
Speaker 1: us rather than to control us. But we shouldn't beat

286
00:17:48,560 --> 00:17:51,440
Speaker 1: we shouldn't beat China too hard with this stick when

287
00:17:51,440 --> 00:17:54,600
Speaker 1: we have certain analogous practices here in the US. Sure,

288
00:17:54,840 --> 00:17:57,280
Speaker 1: I mean you could even I'm sure there are plenty

289
00:17:57,280 --> 00:18:01,040
Speaker 1: of people who do argue that when we the practices

290
00:18:01,080 --> 00:18:06,440
Speaker 1: of certain actual state level organizations in the United States,

291
00:18:06,960 --> 00:18:09,479
Speaker 1: that we should be concerned about what sort of systems

292
00:18:09,480 --> 00:18:11,960
Speaker 1: they might be using. I'm thinking specifically of the n

293
00:18:12,080 --> 00:18:15,840
Speaker 1: s A because it wasn't that long ago when we

294
00:18:15,840 --> 00:18:18,600
Speaker 1: were having enormous headlines about the n s A and

295
00:18:18,640 --> 00:18:21,840
Speaker 1: its practices of trying to have you know, essentially listening

296
00:18:21,920 --> 00:18:28,280
Speaker 1: points just outside of major UH communications channels, whether it

297
00:18:28,400 --> 00:18:32,960
Speaker 1: was Internet service providers or the telecommunications industry in general,

298
00:18:33,320 --> 00:18:36,080
Speaker 1: and you start thinking, well, if you start applying these

299
00:18:36,160 --> 00:18:40,359
Speaker 1: kind of AI surveillance UH programs, which is you know,

300
00:18:40,600 --> 00:18:45,360
Speaker 1: in essay, is all about trying to detect communications UH

301
00:18:45,480 --> 00:18:49,480
Speaker 1: in that everyone gets lumped into and not just bad actors,

302
00:18:49,920 --> 00:18:52,600
Speaker 1: then you start getting those concerns. And you know, we

303
00:18:52,600 --> 00:18:55,560
Speaker 1: we saw plenty of that even without the AI element

304
00:18:55,960 --> 00:18:58,000
Speaker 1: when the n s A stories were breaking a couple

305
00:18:58,000 --> 00:19:01,040
Speaker 1: of years ago, even just to the point where we

306
00:19:01,040 --> 00:19:04,400
Speaker 1: were seeing people in the n s A behaving poorly,

307
00:19:04,560 --> 00:19:11,000
Speaker 1: like using the information gathered to track down like old relationship,

308
00:19:11,800 --> 00:19:15,080
Speaker 1: you know, like old boyfriends and girlfriends, not an ethical

309
00:19:15,160 --> 00:19:18,320
Speaker 1: use of your power if you're if you're looking in

310
00:19:18,400 --> 00:19:22,199
Speaker 1: on communications. So we know that it doesn't have to

311
00:19:22,240 --> 00:19:27,280
Speaker 1: be a an official state line policy for this to

312
00:19:27,560 --> 00:19:32,480
Speaker 1: either be uh misused in an unauthorized way or put

313
00:19:32,520 --> 00:19:35,760
Speaker 1: to use in a way that maybe it's not immoral,

314
00:19:35,880 --> 00:19:38,359
Speaker 1: but you could at least argue it's a moral with

315
00:19:38,400 --> 00:19:42,080
Speaker 1: a lot of the the business practices, because that's morality

316
00:19:42,200 --> 00:19:45,679
Speaker 1: is not consideration when you're looking at how do we

317
00:19:45,920 --> 00:19:49,840
Speaker 1: make more revenue, how do we make a greater profit? Um,

318
00:19:50,160 --> 00:19:52,560
Speaker 1: And and you know you're just you're essentially you're checking

319
00:19:52,560 --> 00:19:55,159
Speaker 1: off boxes saying, right, here's how we can make this

320
00:19:55,240 --> 00:19:59,359
Speaker 1: more efficient, have a lower cost to us, a greater

321
00:19:59,600 --> 00:20:02,520
Speaker 1: payoff off in the long run. And so we start

322
00:20:02,560 --> 00:20:04,640
Speaker 1: to see how exactly as you were saying, Oz that

323
00:20:05,080 --> 00:20:08,040
Speaker 1: that while you it is easy to point to another

324
00:20:08,080 --> 00:20:13,560
Speaker 1: country and say these policies are clearly uh harmful to

325
00:20:13,640 --> 00:20:16,880
Speaker 1: people and are therefore bad, we also need to make

326
00:20:16,920 --> 00:20:20,399
Speaker 1: sure we're reflecting on the environment that we ourselves are

327
00:20:20,440 --> 00:20:23,960
Speaker 1: in when we I'm sorry, go ahead, oh no, I

328
00:20:24,000 --> 00:20:26,359
Speaker 1: was just gonna make one comment, you know, and for example,

329
00:20:26,359 --> 00:20:30,720
Speaker 1: in the state of Arizona, if you applied for a

330
00:20:30,840 --> 00:20:33,520
Speaker 1: driver's license in the state of Arizona, your photo was

331
00:20:33,600 --> 00:20:38,360
Speaker 1: then put into a database that was being used by

332
00:20:38,480 --> 00:20:41,480
Speaker 1: that was being used for facial recognition, right, and that

333
00:20:41,560 --> 00:20:45,159
Speaker 1: was without driver's consent. Basically, law enforcement came back and

334
00:20:45,200 --> 00:20:47,159
Speaker 1: was like, whoa we you know, we think people know

335
00:20:47,200 --> 00:20:49,000
Speaker 1: that this is going on. You know, that was their

336
00:20:49,000 --> 00:20:51,960
Speaker 1: best answer. It wasn't like, oh, there was you know,

337
00:20:52,040 --> 00:20:54,560
Speaker 1: fine print that people didn't read. It was basically like, well,

338
00:20:54,600 --> 00:20:58,600
Speaker 1: we thought people knew, and so you know, I we

339
00:20:58,720 --> 00:21:02,240
Speaker 1: can't exonerate yeah, sort of our own our home turf,

340
00:21:02,280 --> 00:21:04,680
Speaker 1: because there's certainly I don't even know if you would

341
00:21:04,680 --> 00:21:09,480
Speaker 1: call this misuse. It just seems like exploitation um for

342
00:21:10,640 --> 00:21:14,960
Speaker 1: gain for for basically gains of I think police departments

343
00:21:15,000 --> 00:21:19,719
Speaker 1: that are seeing this technology, recognizing how powerful it is,

344
00:21:19,800 --> 00:21:23,760
Speaker 1: but also realizing that it's something that needs to be

345
00:21:23,800 --> 00:21:26,960
Speaker 1: regulated and not knowing how or not really caring. And

346
00:21:27,000 --> 00:21:29,480
Speaker 1: I think Jonathan, you used the great example of the

347
00:21:29,560 --> 00:21:31,879
Speaker 1: n S a UM, which I think tie is very

348
00:21:31,920 --> 00:21:34,320
Speaker 1: neatly to what we're talking about. Edward Snowden had a

349
00:21:34,400 --> 00:21:38,560
Speaker 1: very haunting phrase which was turnkey tyranny to describe basically,

350
00:21:38,560 --> 00:21:42,280
Speaker 1: once you build infrastructure for something, um, anything can happen.

351
00:21:42,400 --> 00:21:44,960
Speaker 1: And the technological infrastructure we have here in the US

352
00:21:45,040 --> 00:21:48,119
Speaker 1: for surveillance and social control is I mean, we have

353
00:21:48,200 --> 00:21:52,160
Speaker 1: few cameras, you know, there are more distinctions between companies,

354
00:21:52,200 --> 00:21:55,280
Speaker 1: but effectively the infrastructure exists to do what's happening in

355
00:21:55,359 --> 00:21:59,440
Speaker 1: China here. And that's very frightening because you know, as

356
00:21:59,480 --> 00:22:03,360
Speaker 1: we know for Henry Ford's what wonderful pressure to build

357
00:22:03,480 --> 00:22:05,840
Speaker 1: roads in the US. Guess what you build the roads,

358
00:22:05,880 --> 00:22:07,840
Speaker 1: people are going to drive cars and not take the train.

359
00:22:07,960 --> 00:22:10,920
Speaker 1: So once this infrastructure exists, and you add to that,

360
00:22:11,440 --> 00:22:15,600
Speaker 1: you know a leader who doesn't respect norms or wartime,

361
00:22:16,320 --> 00:22:18,919
Speaker 1: all of a sudden the barriers that we think are

362
00:22:19,000 --> 00:22:23,160
Speaker 1: so solid to protect that infrastructure being weaponized against us

363
00:22:23,520 --> 00:22:26,040
Speaker 1: start to road very very quickly. And I think that

364
00:22:26,040 --> 00:22:28,200
Speaker 1: that's the moment we're in right now. And that's another

365
00:22:28,200 --> 00:22:30,920
Speaker 1: reason why he wants to call this podcast sleepwalkers, because

366
00:22:31,280 --> 00:22:34,640
Speaker 1: we don't insist on those norms and legal protections. Bit

367
00:22:34,680 --> 00:22:37,280
Speaker 1: by bit, the infrastructure will have its own logic and

368
00:22:37,600 --> 00:22:41,280
Speaker 1: one emergency after another. Remember the Patriot Act will allow

369
00:22:41,359 --> 00:22:44,639
Speaker 1: this technology to be used against us in ways that

370
00:22:44,680 --> 00:22:48,160
Speaker 1: we currently find sickening and horrifying and terrifying. In China,

371
00:22:48,680 --> 00:22:52,960
Speaker 1: they could easily come home. Yeah. Yeah, it's a sobering point.

372
00:22:53,160 --> 00:22:55,680
Speaker 1: And on that point, I think we're going to take

373
00:22:55,680 --> 00:22:57,919
Speaker 1: a quick break so I can suck my thumb in

374
00:22:57,920 --> 00:23:11,280
Speaker 1: the corner. Okay, alright, Pruny thomicide. Now we've talked about

375
00:23:11,600 --> 00:23:14,600
Speaker 1: some of the differences between say, China and the United States.

376
00:23:14,840 --> 00:23:17,719
Speaker 1: One of the things I thought was interesting is the

377
00:23:17,760 --> 00:23:21,399
Speaker 1: idea that in China you have you have sort of

378
00:23:21,440 --> 00:23:29,199
Speaker 1: a very concrete strategy in place, right, a top down strategy,

379
00:23:29,800 --> 00:23:33,080
Speaker 1: and it's it's pretty much all the companies that are

380
00:23:33,160 --> 00:23:36,199
Speaker 1: working on this strategy are in alignment with it to

381
00:23:36,520 --> 00:23:39,800
Speaker 1: some degree or another. Some are very much in step

382
00:23:40,160 --> 00:23:44,600
Speaker 1: with the state government. Others are to a lesser degree perhaps,

383
00:23:44,680 --> 00:23:47,720
Speaker 1: but still you know, following along the strategy. Meanwhile, in

384
00:23:47,720 --> 00:23:52,720
Speaker 1: the United States, it's much more of this competitive this

385
00:23:52,720 --> 00:23:56,680
Speaker 1: this classic capitalist idea of competition in the space where

386
00:23:56,720 --> 00:23:59,280
Speaker 1: you have all these different pockets that are all trying

387
00:23:59,320 --> 00:24:03,280
Speaker 1: to own a I themselves competing against each other. So

388
00:24:03,320 --> 00:24:06,520
Speaker 1: we're getting lots of interesting innovation, but not nearly at

389
00:24:06,560 --> 00:24:09,639
Speaker 1: the same speed or scale as we're seeing in China.

390
00:24:09,800 --> 00:24:12,800
Speaker 1: Is that is that more or less a correct assumption

391
00:24:12,880 --> 00:24:15,439
Speaker 1: or am I way off base here? No? I think

392
00:24:15,480 --> 00:24:20,200
Speaker 1: that's completely fair. You know, this year, the government that

393
00:24:20,320 --> 00:24:25,560
Speaker 1: President Trump did announce an executive order on AI in February,

394
00:24:26,400 --> 00:24:28,040
Speaker 1: and so that was that was seen as a kind

395
00:24:28,040 --> 00:24:31,639
Speaker 1: of related response to what China are doing. And that

396
00:24:31,720 --> 00:24:36,040
Speaker 1: executive order that the President issued had four major components.

397
00:24:36,440 --> 00:24:39,439
Speaker 1: Number one, to set AI as a national priority of

398
00:24:39,480 --> 00:24:42,760
Speaker 1: the United States. Number two, interestingly, to get better at

399
00:24:42,880 --> 00:24:47,640
Speaker 1: data sharing between the government and private enterprise. Number three,

400
00:24:47,640 --> 00:24:50,640
Speaker 1: it is the ethical guidelines on how AI is used

401
00:24:50,640 --> 00:24:53,720
Speaker 1: in terms of surveillance and military and number four to

402
00:24:53,840 --> 00:24:56,200
Speaker 1: make sure that the United States is doing the best

403
00:24:56,240 --> 00:24:59,159
Speaker 1: it can to educate the next generation of engineers and

404
00:24:59,240 --> 00:25:02,680
Speaker 1: AI scientists. Now, those are all good things, apart from

405
00:25:02,680 --> 00:25:06,760
Speaker 1: maybe two number two, the data sharing um. But guess

406
00:25:06,800 --> 00:25:09,320
Speaker 1: how much funding that executive order came with. I'm going

407
00:25:09,320 --> 00:25:12,760
Speaker 1: to guess that was a big old goose egg. Zero dollars,

408
00:25:13,600 --> 00:25:18,280
Speaker 1: zero dollars. Whereas shen Zen, which is a province in China,

409
00:25:18,720 --> 00:25:21,639
Speaker 1: is spending fifteen billion dollars this year. That's one of

410
00:25:21,680 --> 00:25:25,520
Speaker 1: the many, many provinces in China spending not even federal

411
00:25:25,560 --> 00:25:28,880
Speaker 1: money but state money on AI. So in that context

412
00:25:28,920 --> 00:25:31,080
Speaker 1: you can say, oh, you know, we're getting up to speed,

413
00:25:31,480 --> 00:25:34,560
Speaker 1: we're responding, but you know you've got to put your

414
00:25:34,560 --> 00:25:37,040
Speaker 1: money where your mouth is and we're not doing that. Yeah.

415
00:25:37,240 --> 00:25:40,600
Speaker 1: And on a on a related note, you know, when

416
00:25:40,640 --> 00:25:45,119
Speaker 1: we talk about things like the regulations, the laws in place,

417
00:25:45,200 --> 00:25:48,280
Speaker 1: like how do we how do we then create policies

418
00:25:48,880 --> 00:25:54,960
Speaker 1: that ensure that we're using AI responsibly and productively and

419
00:25:55,040 --> 00:25:58,840
Speaker 1: not in ways that are harmful or destructive, um at

420
00:25:58,880 --> 00:26:01,399
Speaker 1: least to our own set aisions if and hopefully not

421
00:26:01,480 --> 00:26:05,480
Speaker 1: to anyone at all. We tend to see that lag behind,

422
00:26:05,680 --> 00:26:07,840
Speaker 1: just like we do with technology in general. We tend

423
00:26:07,840 --> 00:26:13,600
Speaker 1: to see technology innovations far out distancing our ability to

424
00:26:13,600 --> 00:26:18,040
Speaker 1: to incorporate that into our our legal you know, kind

425
00:26:18,080 --> 00:26:23,879
Speaker 1: of massive infrastructure. Understandably so obviously that that system is

426
00:26:23,880 --> 00:26:27,320
Speaker 1: going to move much more slowly than technological innovation, but

427
00:26:27,400 --> 00:26:31,399
Speaker 1: it does create these pain points, whether it's uh, you know,

428
00:26:31,480 --> 00:26:34,040
Speaker 1: like autonomous cars. You know, you have different states in

429
00:26:34,080 --> 00:26:37,280
Speaker 1: the United States that will allow for some degree of

430
00:26:37,320 --> 00:26:41,439
Speaker 1: autonomous car testing. Meanwhile, you've got companies like Tesla that

431
00:26:41,480 --> 00:26:45,520
Speaker 1: are rolling out vehicles that have autopilot feature, which, to

432
00:26:45,640 --> 00:26:48,640
Speaker 1: the company's credit, they say, is not meant to be

433
00:26:48,720 --> 00:26:52,280
Speaker 1: taken as an autonomous system. But tell that to all

434
00:26:52,280 --> 00:26:55,680
Speaker 1: the people going down the highway who are leaning back

435
00:26:55,680 --> 00:26:58,399
Speaker 1: in their cars with their hands off the wheel. Um,

436
00:26:58,520 --> 00:27:01,119
Speaker 1: you could argue that that false of the responsibility of

437
00:27:01,119 --> 00:27:03,400
Speaker 1: the individuals. But if enough individuals are doing it, you've

438
00:27:03,400 --> 00:27:05,720
Speaker 1: got to start asking what's the value of actually having

439
00:27:05,720 --> 00:27:10,640
Speaker 1: the system in place? Um, we're seeing that as being

440
00:27:10,640 --> 00:27:13,280
Speaker 1: a kind of a disparity as well. Right, We're seeing

441
00:27:13,320 --> 00:27:17,520
Speaker 1: this this gap between what we're capable of and what

442
00:27:17,560 --> 00:27:19,439
Speaker 1: we should be doing, or at least what you know,

443
00:27:20,320 --> 00:27:23,800
Speaker 1: what are legal system says we should be doing. And meanwhile,

444
00:27:24,720 --> 00:27:27,840
Speaker 1: in contrast to that, and and Kara, you mentioned the

445
00:27:27,840 --> 00:27:30,160
Speaker 1: EU a couple of times in our last podcast. Over

446
00:27:30,160 --> 00:27:35,600
Speaker 1: in the European Union, you have committees dedicated to thinking

447
00:27:35,640 --> 00:27:39,080
Speaker 1: about these sort of things and starting to propose potential

448
00:27:40,040 --> 00:27:46,119
Speaker 1: UH strategies or even presenting different options for strategies for

449
00:27:46,240 --> 00:27:50,000
Speaker 1: dealing with AI. Even beyond these these cases that I'm

450
00:27:50,040 --> 00:27:52,040
Speaker 1: talking about, like they're they're going to the point where

451
00:27:53,160 --> 00:27:55,240
Speaker 1: I remember reporting on this a couple of years ago

452
00:27:55,280 --> 00:27:58,160
Speaker 1: for Forward Thinking, a series I used to do where

453
00:27:58,320 --> 00:28:00,879
Speaker 1: the EU was proposed a committee in the EU, or

454
00:28:00,920 --> 00:28:06,320
Speaker 1: rather was proposing the idea of granting personhood two artificially

455
00:28:06,400 --> 00:28:10,960
Speaker 1: intelligent systems, And on the face of it, that sounds

456
00:28:11,000 --> 00:28:14,240
Speaker 1: absurd to a lot of people, the idea of granting

457
00:28:14,359 --> 00:28:17,440
Speaker 1: a non human the concept of personhood, despite the fact

458
00:28:17,480 --> 00:28:20,440
Speaker 1: that in the United States we have corporations which are

459
00:28:20,480 --> 00:28:25,920
Speaker 1: exactly that. But yeah, like, hey, hey, wait, we corporations

460
00:28:25,920 --> 00:28:30,159
Speaker 1: can be people too, and uh can give money as

461
00:28:30,160 --> 00:28:33,000
Speaker 1: long as they can give money to politicians, absolutely, so

462
00:28:33,119 --> 00:28:36,560
Speaker 1: why can't robots. But the point that the EU committee

463
00:28:36,560 --> 00:28:40,320
Speaker 1: was making was not that robots have feelings and we

464
00:28:40,320 --> 00:28:43,800
Speaker 1: should really be considerate of them, but rather there needs

465
00:28:43,840 --> 00:28:48,959
Speaker 1: to be some way to start to establish concepts like accountability.

466
00:28:49,640 --> 00:28:53,560
Speaker 1: In the case where some form of AI constructor or

467
00:28:53,720 --> 00:28:58,840
Speaker 1: robot causes harm in the form of damages or injury,

468
00:28:59,360 --> 00:29:01,680
Speaker 1: how do you are to determine who's at fault. We

469
00:29:01,760 --> 00:29:03,880
Speaker 1: touched on this a little bit in the last episode,

470
00:29:04,160 --> 00:29:07,160
Speaker 1: so to me, it's fascinating that those are discussions that

471
00:29:07,160 --> 00:29:11,920
Speaker 1: are popping up, like very serious discussions in the EU.

472
00:29:12,160 --> 00:29:15,400
Speaker 1: And Oz I think when we met briefly in New

473
00:29:15,480 --> 00:29:19,360
Speaker 1: York a couple of weeks ago, we talked about the

474
00:29:19,360 --> 00:29:21,760
Speaker 1: fact that this is an area of the world not

475
00:29:21,920 --> 00:29:27,880
Speaker 1: known for making incredible advances in AI technology. It's not

476
00:29:27,960 --> 00:29:30,480
Speaker 1: like you look at Europe as saying this is where

477
00:29:30,480 --> 00:29:33,520
Speaker 1: the hot bed for AI development is, but it is

478
00:29:33,560 --> 00:29:39,120
Speaker 1: a region that's dedicating a lot of consideration to the

479
00:29:39,240 --> 00:29:43,960
Speaker 1: implications of AI in day to day lives. Yeah. Absolutely,

480
00:29:44,000 --> 00:29:46,720
Speaker 1: I mean, the EU has been out in front on

481
00:29:46,880 --> 00:29:51,400
Speaker 1: thinking about technology and AI. Um Obviously, g d PR

482
00:29:51,560 --> 00:29:54,960
Speaker 1: was a big act on regulating data and informing people

483
00:29:55,000 --> 00:29:58,880
Speaker 1: about how people's data is being used passed in the EU,

484
00:29:59,000 --> 00:30:02,040
Speaker 1: which triggered a whole series of conversations in the US

485
00:30:02,080 --> 00:30:06,200
Speaker 1: about data regulation which will probably see they're starting now,

486
00:30:06,280 --> 00:30:08,040
Speaker 1: we may see them come into law in the next

487
00:30:08,120 --> 00:30:10,840
Speaker 1: two or three or four years. So the EU, good

488
00:30:10,840 --> 00:30:15,960
Speaker 1: old Europe, despite its stifling effects on innovation I think

489
00:30:16,000 --> 00:30:20,040
Speaker 1: and on technological innovation, is an innovator in terms of regulation,

490
00:30:20,040 --> 00:30:22,719
Speaker 1: which probably sounds like a contradiction in terms and this

491
00:30:22,760 --> 00:30:24,720
Speaker 1: is an area that Karas looked into really closely and

492
00:30:24,760 --> 00:30:30,080
Speaker 1: has some I think, very interesting insights about. Yeah. Yeah, well,

493
00:30:30,120 --> 00:30:32,320
Speaker 1: I mean, as I was saying, to you earlier actually

494
00:30:32,320 --> 00:30:34,760
Speaker 1: on the form I guess it was the former the

495
00:30:34,800 --> 00:30:38,640
Speaker 1: episode before this. Um, yeah, you know, I think it's interesting.

496
00:30:40,200 --> 00:30:45,320
Speaker 1: The EUSE approach to me up very recently, you know,

497
00:30:45,440 --> 00:30:49,080
Speaker 1: was to basically collect a group of fifty two experts

498
00:30:49,240 --> 00:30:55,040
Speaker 1: and you know, create these seven core requirements for artificial intelligence. Now,

499
00:30:55,520 --> 00:30:56,880
Speaker 1: I mean I can name them. You know. One is

500
00:30:56,960 --> 00:31:00,440
Speaker 1: human agency and oversight. The other is technical robut us nous.

501
00:31:00,480 --> 00:31:03,800
Speaker 1: The third is privacy and data governance. The fourth is transparency.

502
00:31:04,160 --> 00:31:07,440
Speaker 1: The fifth is AI that AI systems should be sustainable,

503
00:31:07,560 --> 00:31:11,880
Speaker 1: um whatever that means. UM, AI systems should be auditable.

504
00:31:11,920 --> 00:31:13,920
Speaker 1: As we were talking about with you know, black the

505
00:31:13,960 --> 00:31:17,280
Speaker 1: black box problem, you know, AI should also be available

506
00:31:17,320 --> 00:31:19,040
Speaker 1: to all. You know, this is what we talk about

507
00:31:19,040 --> 00:31:24,320
Speaker 1: in terms of bias, gender bias, um, you know, racial bias,

508
00:31:24,360 --> 00:31:28,120
Speaker 1: all of those things. But I don't I have a

509
00:31:28,160 --> 00:31:30,080
Speaker 1: bit of an issue with it just in that it's

510
00:31:31,040 --> 00:31:34,640
Speaker 1: it doesn't seem to have much action involved in it.

511
00:31:34,680 --> 00:31:37,040
Speaker 1: There aren't many action items. I think I think it

512
00:31:37,200 --> 00:31:42,680
Speaker 1: is important, um for governments to to set standards you know,

513
00:31:42,800 --> 00:31:45,720
Speaker 1: as a as a sort of first step, um, you know,

514
00:31:45,800 --> 00:31:49,760
Speaker 1: and I think it also primes, you know, sort of

515
00:31:49,800 --> 00:31:53,760
Speaker 1: average citizens to be aware of misuse, right, because when

516
00:31:53,760 --> 00:31:56,360
Speaker 1: you set when you set up requirements, it means that

517
00:31:56,400 --> 00:31:59,840
Speaker 1: these are things that can be misused and be misused easily.

518
00:32:00,040 --> 00:32:01,880
Speaker 1: And I will say, you know, whenever I go home,

519
00:32:02,040 --> 00:32:05,200
Speaker 1: every time you go onto website on your phone, notifications

520
00:32:05,280 --> 00:32:08,760
Speaker 1: saying are you willing to let this website put cookies?

521
00:32:08,960 --> 00:32:11,120
Speaker 1: Are you willing to share your data? Are you willing

522
00:32:11,120 --> 00:32:14,440
Speaker 1: to accept targeted ads? And that the net effect socially

523
00:32:14,560 --> 00:32:17,080
Speaker 1: of that reminder that your data is valuable and you

524
00:32:17,120 --> 00:32:20,560
Speaker 1: have a choice every single day fifty times that that

525
00:32:20,760 --> 00:32:24,760
Speaker 1: inevitably caused the shifting consciousness and the shifting citizenship. So

526
00:32:25,080 --> 00:32:27,400
Speaker 1: I do think it's easy to say this regulation is toothless,

527
00:32:27,400 --> 00:32:30,240
Speaker 1: but I also think it can, you know, make people think, oh,

528
00:32:30,240 --> 00:32:31,920
Speaker 1: it's a huge culture. I mean when you just think,

529
00:32:31,960 --> 00:32:34,080
Speaker 1: I mean, think about it. Think about cigarettes, right, I mean,

530
00:32:34,080 --> 00:32:37,560
Speaker 1: it's just not you just don't smoke inside right anyway,

531
00:32:37,840 --> 00:32:41,080
Speaker 1: illegally you can't. But I'm saying, you know, just the

532
00:32:41,160 --> 00:32:45,680
Speaker 1: shift in public perception about smoking, the shift in public respection,

533
00:32:45,800 --> 00:32:49,160
Speaker 1: um perspective about sugar, for example, at least in the

534
00:32:49,240 --> 00:32:53,040
Speaker 1: United States. You know, those were all things that at

535
00:32:53,080 --> 00:32:55,280
Speaker 1: one point, we're not really discussed or talked about, and

536
00:32:55,280 --> 00:32:57,040
Speaker 1: you know, everyone kind of smoke cigarettes and that's what

537
00:32:57,040 --> 00:32:59,000
Speaker 1: people did and you know, then they got lung cancer.

538
00:32:59,080 --> 00:33:02,640
Speaker 1: But you know, I think the there is as long

539
00:33:02,680 --> 00:33:07,480
Speaker 1: as there is a public discourse about privacy, people will

540
00:33:07,520 --> 00:33:10,320
Speaker 1: begin to care more and more about privacy. So I think,

541
00:33:10,600 --> 00:33:15,200
Speaker 1: you know, the EU presenting um these seven requirements is

542
00:33:15,520 --> 00:33:19,240
Speaker 1: a step in the right direction. Absolutely, and if if anything,

543
00:33:19,600 --> 00:33:23,640
Speaker 1: will just make people think about these seven tenants when

544
00:33:23,680 --> 00:33:26,120
Speaker 1: they're going about their sort of everyday lives, and I

545
00:33:26,120 --> 00:33:30,240
Speaker 1: think we'll we'll make companies, um, you know, focus on

546
00:33:30,480 --> 00:33:35,200
Speaker 1: building these requirements into whatever they are developing. You know,

547
00:33:35,680 --> 00:33:39,640
Speaker 1: I think it's important for companies that are developing new

548
00:33:39,680 --> 00:33:42,800
Speaker 1: technologies to think about bias, especially when the people who

549
00:33:42,800 --> 00:33:47,800
Speaker 1: are developing the technologies might be quite homogeneous. So I

550
00:33:47,800 --> 00:33:49,560
Speaker 1: think it's worth saying that, you know, we we tend

551
00:33:49,600 --> 00:33:52,320
Speaker 1: to see the new world order in terms of America

552
00:33:52,440 --> 00:33:55,200
Speaker 1: versus China. You know, not to wave the flag for Europe,

553
00:33:55,200 --> 00:33:59,040
Speaker 1: which unfortunately my my country may longer no longer be

554
00:33:59,160 --> 00:34:01,960
Speaker 1: part of. But it's a it's a huge it's a

555
00:34:02,040 --> 00:34:06,080
Speaker 1: huge group of nations with tremendous purchasing power, and it's

556
00:34:06,200 --> 00:34:10,600
Speaker 1: genuinely significant market for US technology companies. And so you know,

557
00:34:10,760 --> 00:34:13,920
Speaker 1: the effects of this may seem far off and slightly irrelevant,

558
00:34:13,960 --> 00:34:16,880
Speaker 1: but you know, these finds that the EU is already

559
00:34:16,880 --> 00:34:20,120
Speaker 1: slapping Facebook and Google with they're not necessarily materials to

560
00:34:20,200 --> 00:34:23,600
Speaker 1: business yet. But you know, Europe as a voice for

561
00:34:23,880 --> 00:34:28,279
Speaker 1: regulation is an important one because you know, actually these

562
00:34:28,320 --> 00:34:32,160
Speaker 1: new technologies, AI technologies, they tend to affect the poorest

563
00:34:32,200 --> 00:34:35,160
Speaker 1: in society and the most vulnerable in society in the

564
00:34:35,200 --> 00:34:40,840
Speaker 1: most negative ways. So algorithmic discrimination, the replacement of you know,

565
00:34:41,320 --> 00:34:45,400
Speaker 1: of low education, shall we say, jobs like driving or packaging.

566
00:34:45,640 --> 00:34:49,160
Speaker 1: You know, these are things that are being experienced at

567
00:34:49,200 --> 00:34:51,239
Speaker 1: the hard edge by people who don't have much of

568
00:34:51,239 --> 00:34:53,640
Speaker 1: a voice in the political system. And so the fact

569
00:34:53,640 --> 00:34:56,360
Speaker 1: that the EU is is taking up the mantle, even

570
00:34:56,400 --> 00:34:59,920
Speaker 1: though it comes with hypocrisy, like using facial recognition at

571
00:35:00,040 --> 00:35:03,040
Speaker 1: the ports of entry, even though you know that the

572
00:35:03,200 --> 00:35:05,360
Speaker 1: level of the finds that can be applied to companies

573
00:35:05,400 --> 00:35:07,839
Speaker 1: like Google and Facebook can never hurt their bottom line,

574
00:35:08,239 --> 00:35:11,319
Speaker 1: I would say, is an important and valuable thing that

575
00:35:11,400 --> 00:35:13,759
Speaker 1: it's happening. And so you know, I think the EU

576
00:35:14,120 --> 00:35:17,759
Speaker 1: can offer some point of reference for how we may

577
00:35:17,760 --> 00:35:20,560
Speaker 1: think about regulating AI and technology in the US in

578
00:35:20,600 --> 00:35:25,120
Speaker 1: the future. And I think that you know, you've heard

579
00:35:25,360 --> 00:35:30,200
Speaker 1: futurists say we need to have some very serious conversations

580
00:35:30,800 --> 00:35:34,920
Speaker 1: about AI and the ethics of AI and how we

581
00:35:34,960 --> 00:35:38,440
Speaker 1: can make certain we're being responsible custodians of a I

582
00:35:39,200 --> 00:35:41,319
Speaker 1: and UH. To me, that was one of those things

583
00:35:41,400 --> 00:35:43,560
Speaker 1: where it was like, we need to talk about talking

584
00:35:43,600 --> 00:35:46,239
Speaker 1: about this, like that was the conversation for a very

585
00:35:46,320 --> 00:35:48,560
Speaker 1: long time. We need to talk about talking about it.

586
00:35:48,880 --> 00:35:50,960
Speaker 1: It's like having a meeting to talk about when you're

587
00:35:50,960 --> 00:35:53,040
Speaker 1: gonna have your meeting. Have you ever worked in an

588
00:35:53,080 --> 00:35:56,319
Speaker 1: office before? Oh? Yeah, no, I worked in I worked

589
00:35:56,320 --> 00:35:59,600
Speaker 1: in a college administrative office before. So I've had meeting

590
00:35:59,760 --> 00:36:03,960
Speaker 1: about how we can have fewer meetings and this is

591
00:36:03,960 --> 00:36:09,360
Speaker 1: not the way. Yeah. So, so having the EU actually

592
00:36:09,360 --> 00:36:11,840
Speaker 1: take this step, even if you were to argue like

593
00:36:11,880 --> 00:36:14,120
Speaker 1: this is a very early step and maybe there's not

594
00:36:14,160 --> 00:36:17,480
Speaker 1: a lot of teeth to it yet, it is a

595
00:36:17,520 --> 00:36:20,960
Speaker 1: step as opposed to what I've seen elsewhere where it's

596
00:36:20,960 --> 00:36:23,280
Speaker 1: been talking about taking a step but not even doing

597
00:36:23,320 --> 00:36:27,560
Speaker 1: that much. So I'm encouraged by it because it actually

598
00:36:27,560 --> 00:36:30,399
Speaker 1: moves the conversation forward instead of us saying we need

599
00:36:30,440 --> 00:36:34,000
Speaker 1: to have this conversation, the conversation has started. I don't

600
00:36:34,000 --> 00:36:36,239
Speaker 1: think it's over yet. I think that this is a

601
00:36:36,280 --> 00:36:41,440
Speaker 1: good way to actually force more parties to get involved

602
00:36:41,440 --> 00:36:46,040
Speaker 1: and think about it and maybe even start proactively thinking,

603
00:36:46,040 --> 00:36:48,839
Speaker 1: how can I make certain that the thing we are

604
00:36:48,880 --> 00:36:52,120
Speaker 1: building is actually being built in a responsible way that

605
00:36:52,239 --> 00:36:55,400
Speaker 1: isn't going to that we can mitigate as many unintended

606
00:36:55,400 --> 00:36:58,799
Speaker 1: consequences as possible, knowing that that is impossible to do

607
00:36:59,719 --> 00:37:04,319
Speaker 1: uh entirely, but to really try for it. So to me,

608
00:37:04,480 --> 00:37:06,399
Speaker 1: this is this is one of those conversations I could

609
00:37:06,440 --> 00:37:09,040
Speaker 1: have all day long. But I know you guys need

610
00:37:09,080 --> 00:37:10,640
Speaker 1: to get going because there's going to be someone else

611
00:37:10,640 --> 00:37:12,359
Speaker 1: who's going to have to use the studio you guys

612
00:37:12,360 --> 00:37:14,080
Speaker 1: are in, So I'm going to I'm not going to

613
00:37:14,200 --> 00:37:16,480
Speaker 1: have us go all day long on this. Plus, we're

614
00:37:16,480 --> 00:37:19,160
Speaker 1: recording this on a holiday weekend, and I know everybody

615
00:37:19,200 --> 00:37:22,200
Speaker 1: wants to get home, so we're gonna wrap it up.

616
00:37:22,239 --> 00:37:25,160
Speaker 1: But I want to thank you guys so much for

617
00:37:25,239 --> 00:37:30,399
Speaker 1: agreeing to come onto my show. Uh it's a fascinating conversation,

618
00:37:30,600 --> 00:37:34,239
Speaker 1: and you guys obviously have some great perspectives on this.

619
00:37:34,760 --> 00:37:39,200
Speaker 1: And again to my listeners out there, if you haven't

620
00:37:39,200 --> 00:37:42,800
Speaker 1: subscribed to Sleepwalkers, go check it out. It's a really

621
00:37:42,880 --> 00:37:47,399
Speaker 1: well done show. I've been very impressed as someone who

622
00:37:47,400 --> 00:37:51,080
Speaker 1: has a solo show. Most of the time I listened

623
00:37:51,080 --> 00:37:54,200
Speaker 1: to it, I just think, wow, that's that's so awesome,

624
00:37:54,800 --> 00:37:57,400
Speaker 1: that's so great. I wish I were on that show

625
00:37:57,440 --> 00:38:00,400
Speaker 1: every now and then. So hint, hint, if you vernee me,

626
00:38:00,440 --> 00:38:03,239
Speaker 1: you just we love it. We would absolutely love it.

627
00:38:03,320 --> 00:38:06,240
Speaker 1: Let's talk about that. We had a great time today.

628
00:38:06,520 --> 00:38:10,719
Speaker 1: Thank you, Thank you guys, and you guys, if you

629
00:38:10,760 --> 00:38:13,200
Speaker 1: want to get in touch with me, drop me a line,

630
00:38:13,280 --> 00:38:16,200
Speaker 1: say hey, you need to talk about this other topic

631
00:38:16,360 --> 00:38:18,719
Speaker 1: or those guys were great. Have them back on the

632
00:38:18,760 --> 00:38:21,319
Speaker 1: show as soon as possible. Send me an email. It's

633
00:38:21,400 --> 00:38:24,640
Speaker 1: tech Stuff at how stuff works dot com. Pop on

634
00:38:24,719 --> 00:38:27,279
Speaker 1: over to the website that's tech Stuff podcast dot com.

635
00:38:27,320 --> 00:38:30,799
Speaker 1: You'll find the archive for all of my episodes. If

636
00:38:30,840 --> 00:38:33,480
Speaker 1: you are really bored, they're more than a thousand of them,

637
00:38:33,560 --> 00:38:36,360
Speaker 1: so have at it. And then you can pop on

638
00:38:36,400 --> 00:38:38,160
Speaker 1: over to the merchandise store and you can finally get

639
00:38:38,160 --> 00:38:41,160
Speaker 1: yourself that tech Stuff mug that you've been wanting all

640
00:38:41,160 --> 00:38:44,479
Speaker 1: this time, and I'll talk to you again. Really. Sion

641
00:38:49,239 --> 00:38:51,440
Speaker 1: Text Stuff is a production of I Heart Radio's How

642
00:38:51,520 --> 00:38:54,879
Speaker 1: Stuff Works. For more podcasts from my heart Radio, visit

643
00:38:54,920 --> 00:38:58,040
Speaker 1: the I heart Radio app, Apple Podcasts, or wherever you

644
00:38:58,080 --> 00:39:01,640
Speaker 1: listen to your favorite shows. Eight