1
00:00:02,720 --> 00:00:07,200
Speaker 1: Bloomberg Audio Studios, podcasts, radio news.

2
00:00:08,480 --> 00:00:11,160
Speaker 2: If you were in Shenzhen in early March, you might

3
00:00:11,160 --> 00:00:14,520
Speaker 2: have caught an unusual sight outside the headquarters of one

4
00:00:14,520 --> 00:00:15,960
Speaker 2: of China's biggest tech.

5
00:00:15,800 --> 00:00:20,079
Speaker 3: Companies, Chanjo more say the open Claw.

6
00:00:20,960 --> 00:00:24,880
Speaker 2: Crowds of people huddled around mixshift tables in an open courtyard.

7
00:00:25,400 --> 00:00:28,840
Speaker 2: Everyone had a laptop in hand. People were shoulder to

8
00:00:28,920 --> 00:00:33,800
Speaker 2: shoulder typing, tinkering, troubleshooting. It looked like an outdoor hackathon

9
00:00:33,960 --> 00:00:35,800
Speaker 2: had broken out in the middle of the city.

10
00:00:36,320 --> 00:00:40,880
Speaker 4: There were hundreds of people from retire's elderlies to school

11
00:00:40,920 --> 00:00:43,960
Speaker 4: kids and of course Yawn office workers.

12
00:00:43,960 --> 00:00:47,599
Speaker 2: LUs Ding covers China Tech for Bloomberg out of Hong Kong.

13
00:00:47,800 --> 00:00:50,559
Speaker 4: And they were all there to get tension engineers to

14
00:00:50,600 --> 00:00:52,920
Speaker 4: install open Claw on their computer.

15
00:00:53,680 --> 00:00:55,360
Speaker 3: Loose what is open claw.

16
00:00:56,080 --> 00:00:59,639
Speaker 4: It's a very capable AI agentic tool. It's much more

17
00:00:59,680 --> 00:01:03,040
Speaker 4: capable than AI chatbots. It can't help you to do

18
00:01:03,160 --> 00:01:06,200
Speaker 4: your work on your computer like swarming files, edit videos,

19
00:01:06,520 --> 00:01:07,520
Speaker 4: making purchases.

20
00:01:08,120 --> 00:01:11,319
Speaker 2: Tools like open Claw aren't something you can just download

21
00:01:11,360 --> 00:01:12,240
Speaker 2: from an app store.

22
00:01:12,760 --> 00:01:15,480
Speaker 3: They require technical setup and a bit of know how.

23
00:01:15,800 --> 00:01:19,560
Speaker 2: That's why in China people are organizing gatherings like the

24
00:01:19,600 --> 00:01:22,840
Speaker 2: one at ten Cents headquarters to help potential users get

25
00:01:22,880 --> 00:01:27,080
Speaker 2: open Claw up and running. Nearly a thousand people showed

26
00:01:27,160 --> 00:01:29,920
Speaker 2: up to that particular meetup, one of many across the

27
00:01:29,959 --> 00:01:31,360
Speaker 2: country this past month.

28
00:01:33,200 --> 00:01:35,600
Speaker 5: I think it really is a vouch for the amount

29
00:01:35,600 --> 00:01:37,880
Speaker 5: of AI enthusiasm that we've seen in China that we

30
00:01:38,240 --> 00:01:39,760
Speaker 5: really haven't seen in other places.

31
00:01:40,280 --> 00:01:44,160
Speaker 2: Catherine Thorbeck is a Bloomberg Opinion tech columnist based in Tokyo.

32
00:01:44,720 --> 00:01:47,400
Speaker 5: When you have local governments coming out and signaling support

33
00:01:47,600 --> 00:01:49,920
Speaker 5: and sort of the big tech players and the small

34
00:01:49,960 --> 00:01:53,080
Speaker 5: startups and the model builders all jumping on this bandwagon

35
00:01:53,080 --> 00:01:54,880
Speaker 5: at once, I think it really created this sort of

36
00:01:55,280 --> 00:01:58,120
Speaker 5: title wave of you know, it just shows how fast

37
00:01:58,240 --> 00:02:00,960
Speaker 5: technology trends can really spread China.

38
00:02:01,240 --> 00:02:04,520
Speaker 2: China has the most open claw users in the world,

39
00:02:04,960 --> 00:02:08,400
Speaker 2: almost double the activity in the US. That's according to

40
00:02:08,440 --> 00:02:12,880
Speaker 2: cybersecurity firm Security Scorecard. But as the hype surrounding open

41
00:02:12,919 --> 00:02:17,320
Speaker 2: Class surged, so has the scrutiny. The initial enthusiasm for

42
00:02:17,360 --> 00:02:21,240
Speaker 2: the technology is now being met with caution, as users

43
00:02:21,280 --> 00:02:24,520
Speaker 2: in China's government way whether the benefits of the tool

44
00:02:25,080 --> 00:02:28,000
Speaker 2: justify the risk of such deep access.

45
00:02:28,480 --> 00:02:31,840
Speaker 5: When you give these tools access to your personal data,

46
00:02:31,960 --> 00:02:35,000
Speaker 5: your personal files, your computer, and then you also give

47
00:02:35,000 --> 00:02:37,600
Speaker 5: them permission to take action. The amount of things that

48
00:02:37,600 --> 00:02:40,560
Speaker 5: can go wrong really just compounds. So there have been

49
00:02:40,600 --> 00:02:43,160
Speaker 5: reports of these tools sort of going rogue and you know,

50
00:02:43,200 --> 00:02:45,480
Speaker 5: deleting all of your emails or deleting all of sort

51
00:02:45,480 --> 00:02:50,840
Speaker 5: of your personal files.

52
00:02:52,840 --> 00:02:56,320
Speaker 3: This is the Big Take Asia from Bloomberg News. I'm Wanha.

53
00:02:56,840 --> 00:02:59,239
Speaker 2: Every week we take you inside some of the world's

54
00:02:59,240 --> 00:03:03,160
Speaker 2: biggest and most powerful economies and the markets, tycoons and

55
00:03:03,280 --> 00:03:05,000
Speaker 2: businesses that drive this.

56
00:03:05,040 --> 00:03:06,040
Speaker 3: Ever shifting region.

57
00:03:06,400 --> 00:03:10,240
Speaker 2: Today in the show, China's rushed toward open Claw. What

58
00:03:10,320 --> 00:03:13,320
Speaker 2: the technology does, why it took off so fast in

59
00:03:13,360 --> 00:03:15,919
Speaker 2: the country, and what it tells us about the risks

60
00:03:15,960 --> 00:03:26,880
Speaker 2: and realities of AI's next base. When open Claw launched

61
00:03:26,880 --> 00:03:29,639
Speaker 2: in November of last year, it took off right away.

62
00:03:30,320 --> 00:03:34,040
Speaker 2: Part of that success was the product itself. Open Clau's

63
00:03:34,080 --> 00:03:38,160
Speaker 2: founder describes it as AI that actually does things. One

64
00:03:38,200 --> 00:03:41,360
Speaker 2: of the tech world's biggest names, Jensen Huang of Nvidia,

65
00:03:41,440 --> 00:03:45,480
Speaker 2: called it the single most important release of software probably ever.

66
00:03:46,400 --> 00:03:50,280
Speaker 2: The other part of that success the popularity of the lobster.

67
00:03:53,120 --> 00:03:58,040
Speaker 4: The logo open claw looks like a very cute cartoonish lobster.

68
00:03:58,320 --> 00:04:01,280
Speaker 4: It looks like Aminion from the Menims movie, but then

69
00:04:01,320 --> 00:04:05,800
Speaker 4: it's read and it has claws. Also spicy lobster. It's

70
00:04:06,040 --> 00:04:08,320
Speaker 4: very favorite dish in China. Part of the fan fair

71
00:04:08,400 --> 00:04:09,240
Speaker 4: is also because of that.

72
00:04:09,880 --> 00:04:12,800
Speaker 2: Spend five minutes at an open claw event and you'll

73
00:04:12,800 --> 00:04:16,919
Speaker 2: see lobster hats, lobster plushies, and you'll hear a phrase

74
00:04:17,000 --> 00:04:22,040
Speaker 2: that's become popular in China, yng loncha, which roughly translates

75
00:04:22,080 --> 00:04:26,920
Speaker 2: to raising lobsters loose? What is raising lobsters?

76
00:04:27,640 --> 00:04:31,080
Speaker 4: Raising means that you first install it and then you

77
00:04:31,160 --> 00:04:33,600
Speaker 4: spend time with it, and of course spend money on it,

78
00:04:34,080 --> 00:04:37,400
Speaker 4: and then open clall learns what you like and dislike. Say,

79
00:04:37,400 --> 00:04:40,160
Speaker 4: it learns different skills, and it's like raising a pat,

80
00:04:40,440 --> 00:04:43,160
Speaker 4: only the pet is a virtual AI assistant that can

81
00:04:43,320 --> 00:04:43,920
Speaker 4: work for you.

82
00:04:44,920 --> 00:04:49,239
Speaker 2: A virtual AI assistant also known as an AI agent.

83
00:04:49,920 --> 00:04:53,120
Speaker 2: These agents are based on large language models from open

84
00:04:53,160 --> 00:04:56,200
Speaker 2: ai and Anthropic, but you don't just chat with them.

85
00:04:56,320 --> 00:04:59,240
Speaker 2: You can figure them to do things for you. Come

86
00:04:59,279 --> 00:05:02,799
Speaker 2: through emails, book flights, and appointments. They can even trade

87
00:05:02,880 --> 00:05:06,159
Speaker 2: crypto or negotiate a deal for car for you, depending

88
00:05:06,160 --> 00:05:09,839
Speaker 2: on how much access you granted. And while Openclaw is

89
00:05:10,000 --> 00:05:14,040
Speaker 2: hugely popular in China, it isn't a Chinese product. It

90
00:05:14,120 --> 00:05:18,320
Speaker 2: was created by Peter Steinberger, a developer from Austria. But

91
00:05:18,440 --> 00:05:21,200
Speaker 2: since the AI tool is open source, it's free for

92
00:05:21,320 --> 00:05:25,880
Speaker 2: anyone to use or modify. Earlier this year, Chinese developers

93
00:05:25,920 --> 00:05:28,440
Speaker 2: started tinkering with it and it took on a life

94
00:05:28,440 --> 00:05:30,680
Speaker 2: of its own in China.

95
00:05:30,960 --> 00:05:35,359
Speaker 4: Since after China came back from the Lunar New Year holiday,

96
00:05:35,960 --> 00:05:39,840
Speaker 4: there were more discussions about it on social media and

97
00:05:39,880 --> 00:05:43,080
Speaker 4: then a lot of people wanted to try it. It's

98
00:05:43,080 --> 00:05:47,680
Speaker 4: a very new tool that promises a lot of things

99
00:05:47,720 --> 00:05:50,200
Speaker 4: that the previous AI tools cannot do.

100
00:05:51,000 --> 00:05:54,240
Speaker 2: In China, AI is very much a part of daily

101
00:05:54,360 --> 00:05:58,479
Speaker 2: life right now. The country's most popular AI app, Gobao,

102
00:05:58,680 --> 00:06:02,719
Speaker 2: has roughly two hundred twenty seven million active users every month.

103
00:06:03,240 --> 00:06:05,080
Speaker 2: That's almost double that of deep Sicks.

104
00:06:05,720 --> 00:06:07,039
Speaker 3: But even in AI.

105
00:06:06,839 --> 00:06:12,040
Speaker 2: Hyped China, the popularity of open Class stands out.

106
00:06:12,480 --> 00:06:15,360
Speaker 4: We live in a very competitive society in China, and

107
00:06:15,960 --> 00:06:18,400
Speaker 4: starting from kids, we were told that you need to

108
00:06:18,480 --> 00:06:21,080
Speaker 4: always be at the front. You need to do your

109
00:06:21,120 --> 00:06:24,400
Speaker 4: best all the time, otherwise you will leg behind. And

110
00:06:24,440 --> 00:06:27,600
Speaker 4: then with there's something new and then people's reactions are

111
00:06:27,680 --> 00:06:30,640
Speaker 4: usually just to jump on it before everybody else and

112
00:06:30,680 --> 00:06:33,680
Speaker 4: not be late. For example, I talked to this computer

113
00:06:33,760 --> 00:06:39,000
Speaker 4: engineer in Shinjin who after starting using AI agents, he's

114
00:06:39,040 --> 00:06:42,080
Speaker 4: been using it all the time, and he was telling

115
00:06:42,120 --> 00:06:44,400
Speaker 4: me that he couldn't even go to sleep without giving

116
00:06:44,480 --> 00:06:47,520
Speaker 4: some task for his AI to do. And then one

117
00:06:47,560 --> 00:06:50,799
Speaker 4: of our colleagues in China, she talked to this office

118
00:06:50,839 --> 00:06:54,760
Speaker 4: worker whose boss told her to use AI otherwise you

119
00:06:54,800 --> 00:06:58,800
Speaker 4: get fired. Really, the bosses are anxious as well, because

120
00:06:58,839 --> 00:07:01,800
Speaker 4: they also want to have their employees to be very

121
00:07:01,839 --> 00:07:04,360
Speaker 4: productive with AI and then to make more money.

122
00:07:04,960 --> 00:07:08,279
Speaker 2: Part of that anxiety comes from pressure at the top.

123
00:07:08,839 --> 00:07:11,440
Speaker 2: In China, the push for AI is coming from the

124
00:07:11,560 --> 00:07:15,200
Speaker 2: highest levels of the government, and AI even made its

125
00:07:15,240 --> 00:07:18,320
Speaker 2: way into President Siegenping's New Year address this year.

126
00:07:19,040 --> 00:07:24,840
Speaker 1: Curzy s many large AI models have been competing in

127
00:07:24,880 --> 00:07:28,040
Speaker 1: a race to the top, and breakthroughs have been achieved

128
00:07:28,160 --> 00:07:32,080
Speaker 1: in the research and development of our own chips. All

129
00:07:32,160 --> 00:07:35,480
Speaker 1: this has turned China into one of the economies with

130
00:07:35,600 --> 00:07:38,720
Speaker 1: the fastest growing innovation capability.

131
00:07:39,640 --> 00:07:44,760
Speaker 2: Bloomberg Opinions Katherine Thorbeck says the government is aggressively driving

132
00:07:44,920 --> 00:07:47,120
Speaker 2: this AI push.

133
00:07:46,440 --> 00:07:48,520
Speaker 5: The government really wants to be a world leader when

134
00:07:48,520 --> 00:07:50,800
Speaker 5: it comes to AI technology, and when you look at

135
00:07:50,840 --> 00:07:53,960
Speaker 5: the AI plus plan that they rolled out last year,

136
00:07:53,960 --> 00:07:56,480
Speaker 5: they really are trying to use AI to solve a

137
00:07:56,480 --> 00:07:59,440
Speaker 5: lot of sort of societal problems, whether that's boosting consumption,

138
00:07:59,560 --> 00:08:01,920
Speaker 5: which it's not totally clear how AI will do that,

139
00:08:02,160 --> 00:08:04,400
Speaker 5: but they definitely are trying to integrate AI into as

140
00:08:04,440 --> 00:08:07,520
Speaker 5: many facets of society as possible. And then another reason

141
00:08:07,560 --> 00:08:09,800
Speaker 5: I think that it's moving so fast in China is

142
00:08:09,840 --> 00:08:12,880
Speaker 5: because of the open source embrace. All of the big

143
00:08:13,000 --> 00:08:17,280
Speaker 5: Chinese AI model builders are really going full steam ahead

144
00:08:17,280 --> 00:08:20,600
Speaker 5: on open source, which means they're putting out their new models,

145
00:08:20,640 --> 00:08:22,600
Speaker 5: they're putting out the weights for them, they're free for

146
00:08:22,640 --> 00:08:25,360
Speaker 5: people to use and download. And I think as a

147
00:08:25,360 --> 00:08:28,880
Speaker 5: result of that, this whole ecosystem of developers and tinkers

148
00:08:28,960 --> 00:08:30,760
Speaker 5: can all sort of look at the code and move

149
00:08:30,800 --> 00:08:32,200
Speaker 5: things along a lot faster.

150
00:08:33,640 --> 00:08:37,880
Speaker 2: In other words, when AI becomes a political priority, tech

151
00:08:37,920 --> 00:08:41,920
Speaker 2: trends like open claw can move quickly. Several local governments

152
00:08:41,960 --> 00:08:44,920
Speaker 2: have already backed that momentum by offering more than a

153
00:08:45,000 --> 00:08:48,400
Speaker 2: million dollars in subsidies to support open claw and similar

154
00:08:48,440 --> 00:08:52,600
Speaker 2: open source projects, and the country's tech giants have followed suit.

155
00:08:53,080 --> 00:08:55,720
Speaker 2: Ten Cent and buy Do launch their own versions of

156
00:08:55,760 --> 00:08:58,480
Speaker 2: open claw in March, using them as a way to

157
00:08:58,559 --> 00:09:02,120
Speaker 2: pull people onto their page platforms. Loose tried some of

158
00:09:02,160 --> 00:09:03,239
Speaker 2: them out for herself.

159
00:09:04,240 --> 00:09:07,400
Speaker 4: Now there are a lot of cloud based offers that

160
00:09:07,480 --> 00:09:10,800
Speaker 4: you can raise the observer on the cloud so it

161
00:09:10,840 --> 00:09:15,920
Speaker 4: doesn't access your local desktop and local files. I chose

162
00:09:16,000 --> 00:09:18,959
Speaker 4: two platforms. One is from ali Baba and it was

163
00:09:19,000 --> 00:09:22,640
Speaker 4: a free service. Another two that's under by Dance. It's

164
00:09:22,760 --> 00:09:28,120
Speaker 4: TikTok's parent company. I paid ninety nine per month and

165
00:09:28,200 --> 00:09:31,680
Speaker 4: that's about fourteen to fifteen dollars. Once you paid for

166
00:09:31,760 --> 00:09:34,960
Speaker 4: a month, you have a certain amount of tokens or

167
00:09:35,000 --> 00:09:38,480
Speaker 4: credits that you can use. If you, for example, wanted

168
00:09:38,480 --> 00:09:41,480
Speaker 4: to start your email, you need to granted the access

169
00:09:41,480 --> 00:09:44,079
Speaker 4: to your email and then it will do it.

170
00:09:44,559 --> 00:09:46,640
Speaker 3: And what did you ask it to do for you?

171
00:09:47,200 --> 00:09:49,800
Speaker 4: I asked it to send news briefs about the companies

172
00:09:49,840 --> 00:09:52,440
Speaker 4: that I cover as a journalist. It did not do

173
00:09:52,520 --> 00:09:55,200
Speaker 4: a very good job because the news resources that it

174
00:09:55,240 --> 00:09:58,040
Speaker 4: can access is very limited. And now it just asked

175
00:09:58,080 --> 00:10:00,120
Speaker 4: it to send me a cat photo every hour.

176
00:10:00,760 --> 00:10:02,200
Speaker 3: A cat photo every hour.

177
00:10:02,320 --> 00:10:04,600
Speaker 4: Yeah, that's what I need in my life.

178
00:10:06,160 --> 00:10:10,040
Speaker 2: But beneath the novelty, there's a serious difference between open

179
00:10:10,080 --> 00:10:13,280
Speaker 2: claw and traditional AI chat bots like deep seek and

180
00:10:13,360 --> 00:10:17,560
Speaker 2: chat GBT, and that's the amount of access these tools require.

181
00:10:18,400 --> 00:10:20,480
Speaker 4: For open Claw, you need to give it a lot

182
00:10:20,520 --> 00:10:24,040
Speaker 4: of access. For example, it needs to read your files

183
00:10:24,280 --> 00:10:27,360
Speaker 4: to swart your file. It need to access your editing

184
00:10:27,440 --> 00:10:30,480
Speaker 4: software to edit for you. And if you wanted to

185
00:10:30,480 --> 00:10:32,280
Speaker 4: buy things for you, you need to give it to your

186
00:10:32,400 --> 00:10:35,480
Speaker 4: credits information, which is kind of a redfleg.

187
00:10:38,320 --> 00:10:41,440
Speaker 2: Beijing may be all in on the idea of spurring

188
00:10:41,480 --> 00:10:46,120
Speaker 2: innovation with AI, but open claus powerful capabilities are giving

189
00:10:46,160 --> 00:10:48,480
Speaker 2: the government pause after the break.

190
00:10:48,600 --> 00:10:50,640
Speaker 3: Why the sweeping access.

191
00:10:50,240 --> 00:10:54,960
Speaker 2: Granted to AI agents makes Chinese regulators uneasy, and why

192
00:10:55,000 --> 00:11:12,280
Speaker 2: some Chinese users are shying away from it too. While

193
00:11:12,320 --> 00:11:15,440
Speaker 2: open Claw has been embraced by users, tech giants, and

194
00:11:15,520 --> 00:11:19,000
Speaker 2: even some local governments in China, the response from Beijing

195
00:11:19,280 --> 00:11:20,840
Speaker 2: has been more restrained.

196
00:11:22,720 --> 00:11:26,839
Speaker 5: Open Clay open there.

197
00:11:27,920 --> 00:11:30,920
Speaker 2: The Chinese government has issued a security alert on the

198
00:11:31,000 --> 00:11:34,640
Speaker 2: risks of using open Claw, emphasizing that users might be

199
00:11:34,720 --> 00:11:39,040
Speaker 2: susceptible to cyber attacks and data leaks. Bloomberg's loose Ding

200
00:11:39,200 --> 00:11:43,320
Speaker 2: and Bloomberg opinions. Katherine Thorbeck say regulators have allowed the

201
00:11:43,400 --> 00:11:46,280
Speaker 2: tool to spread, but with clear limits.

202
00:11:47,400 --> 00:11:53,720
Speaker 4: The military that's overseeing technology industry has issued warnings about

203
00:11:53,720 --> 00:11:58,120
Speaker 4: the security risk of this tool. State media also posted

204
00:11:58,880 --> 00:12:03,000
Speaker 4: articles about how to use open claw safely. Then they

205
00:12:03,040 --> 00:12:07,160
Speaker 4: were also authority spanning the use of open claw in

206
00:12:07,960 --> 00:12:11,960
Speaker 4: sensitive industries like government agencies, banks.

207
00:12:12,720 --> 00:12:15,000
Speaker 5: So I think the government is very aware of the risks,

208
00:12:15,040 --> 00:12:16,800
Speaker 5: and I think they're just trying to strike that balance

209
00:12:16,840 --> 00:12:20,280
Speaker 5: between embracing innovation and seeing where this takes them. And

210
00:12:20,640 --> 00:12:23,199
Speaker 5: also they're obviously very concerned about sort of data leaks

211
00:12:23,200 --> 00:12:24,040
Speaker 5: and data privacy.

212
00:12:24,800 --> 00:12:29,559
Speaker 2: That caution centers on one big question, what happens when

213
00:12:29,600 --> 00:12:34,000
Speaker 2: these tools are given broad access to people's data and devices.

214
00:12:34,840 --> 00:12:37,640
Speaker 2: You mentioned the security risks associated with open claw, What

215
00:12:38,000 --> 00:12:40,600
Speaker 2: are those and what can go wrong?

216
00:12:41,200 --> 00:12:43,440
Speaker 5: Oh gosh, So when it comes to the cybersecurity risks

217
00:12:43,440 --> 00:12:46,040
Speaker 5: for open claw, I'm not even sure where to start.

218
00:12:46,080 --> 00:12:49,520
Speaker 5: I think when you give these tools access to your

219
00:12:49,520 --> 00:12:52,960
Speaker 5: personal data, your personal files, your computer, and then you

220
00:12:53,000 --> 00:12:55,679
Speaker 5: also give them permission to take action, the amount of

221
00:12:55,679 --> 00:12:58,720
Speaker 5: things that can go wrong really just compounds. So there

222
00:12:58,760 --> 00:13:01,240
Speaker 5: have been reports of these tools sort of going rogue

223
00:13:01,240 --> 00:13:03,720
Speaker 5: and you know, deleting all of your emails or deleting

224
00:13:03,720 --> 00:13:06,400
Speaker 5: all of sort of your personal files. So that's one

225
00:13:06,440 --> 00:13:08,320
Speaker 5: side of it, and then on the other side, I

226
00:13:08,320 --> 00:13:10,480
Speaker 5: think it also sort of invites all sorts of new

227
00:13:10,520 --> 00:13:14,400
Speaker 5: sort of cybersecurity attackers and cyber attackers. And so the

228
00:13:14,440 --> 00:13:16,400
Speaker 5: way I sort of think about it is, if you remember,

229
00:13:16,559 --> 00:13:19,439
Speaker 5: like the two thousands Internet, when we were all just

230
00:13:19,520 --> 00:13:22,600
Speaker 5: kind of downloading various softwares onto our PC, and none

231
00:13:22,640 --> 00:13:24,920
Speaker 5: of us really knew what we were downloading, but there

232
00:13:24,960 --> 00:13:27,760
Speaker 5: was definitely a curiosity, but we might not have understood

233
00:13:27,760 --> 00:13:30,160
Speaker 5: that we're also downloading all these viruses. So I think

234
00:13:30,160 --> 00:13:32,440
Speaker 5: we're seeing some of that as well with people, because,

235
00:13:32,480 --> 00:13:34,840
Speaker 5: as I mentioned, you need some little bit of a

236
00:13:34,840 --> 00:13:37,880
Speaker 5: tech background to really use open claw and plug ins

237
00:13:37,920 --> 00:13:39,160
Speaker 5: and sort of the things that you can do on

238
00:13:39,200 --> 00:13:40,840
Speaker 5: top of it. So I think people are downloading all

239
00:13:40,920 --> 00:13:43,839
Speaker 5: kinds of things and not totally aware of what they're

240
00:13:43,880 --> 00:13:46,200
Speaker 5: installing and what sort of permissions they're granting.

241
00:13:46,960 --> 00:13:49,040
Speaker 2: By now people in China have had some time to

242
00:13:49,080 --> 00:13:51,520
Speaker 2: play around with open claw, I want to talk about

243
00:13:51,559 --> 00:13:55,320
Speaker 2: how Chinese users are feeling about the tool. Now you know,

244
00:13:55,360 --> 00:13:59,000
Speaker 2: after the initial excitement and frenzy, there have you heard

245
00:13:59,080 --> 00:14:03,040
Speaker 2: about this one lobster freeway saying that's trending on Chinese

246
00:14:03,040 --> 00:14:03,640
Speaker 2: social media?

247
00:14:04,240 --> 00:14:06,680
Speaker 5: Yes, this sort of hashtag that you're referring to that

248
00:14:06,720 --> 00:14:10,320
Speaker 5: has taken off now. One lobster three ways refers to

249
00:14:11,040 --> 00:14:14,200
Speaker 5: paying somebody to install open claw for them, and then

250
00:14:14,720 --> 00:14:17,080
Speaker 5: paying somebody else or a lot of times that same

251
00:14:17,080 --> 00:14:20,000
Speaker 5: person to uninstall open claw for them, and then also

252
00:14:20,080 --> 00:14:22,520
Speaker 5: paying them to clean up the mess that open Claw

253
00:14:22,600 --> 00:14:25,320
Speaker 5: might have left behind on their computer, whether that's deleting

254
00:14:25,360 --> 00:14:28,840
Speaker 5: files or swing all kinds of chaos. So I definitely

255
00:14:28,920 --> 00:14:32,760
Speaker 5: think that the initial excitement has moderated a little bit.

256
00:14:35,520 --> 00:14:38,960
Speaker 2: But even as the hype settles, Catherine says, it's clear

257
00:14:39,080 --> 00:14:42,640
Speaker 2: that China's big tech companies are betting on agenic Ai

258
00:14:43,160 --> 00:14:46,400
Speaker 2: to be the next big driver of growth. The country

259
00:14:46,440 --> 00:14:49,240
Speaker 2: is a testing ground for it, and what happens next

260
00:14:49,480 --> 00:14:50,520
Speaker 2: is still worth watching.

261
00:14:51,160 --> 00:14:54,560
Speaker 5: I really look at it as this real world experiment

262
00:14:54,640 --> 00:14:58,040
Speaker 5: where regular people are playing around with agentic tools and

263
00:14:58,080 --> 00:15:00,200
Speaker 5: seeing what it can do for their lives, and it

264
00:15:00,240 --> 00:15:03,400
Speaker 5: can actually boost productivity and sort of what the risks are,

265
00:15:03,440 --> 00:15:06,120
Speaker 5: and that's happening at such a massive scale in China,

266
00:15:06,640 --> 00:15:09,320
Speaker 5: and it's really a double edged sword because the more

267
00:15:09,520 --> 00:15:14,320
Speaker 5: permissions and data and access that you give to AI agents,

268
00:15:14,400 --> 00:15:16,800
Speaker 5: the more they can do, and that's really exciting. But

269
00:15:16,840 --> 00:15:18,600
Speaker 5: then at the same time, the flip side of that

270
00:15:18,880 --> 00:15:21,000
Speaker 5: is the more damage they can also cause. So I

271
00:15:21,040 --> 00:15:25,280
Speaker 5: think making AI agents secure isn't something that either side

272
00:15:25,280 --> 00:15:26,920
Speaker 5: has really figured out yet, and I think it's going

273
00:15:26,960 --> 00:15:29,160
Speaker 5: to take a lot of time. I will say that

274
00:15:29,200 --> 00:15:32,560
Speaker 5: one benefit to open claw is that it is open source,

275
00:15:32,640 --> 00:15:35,640
Speaker 5: so that means that the more people at this scale

276
00:15:35,680 --> 00:15:37,880
Speaker 5: that are all using it at the same time, the

277
00:15:37,880 --> 00:15:40,000
Speaker 5: more it sort of hardens it because people are, you know,

278
00:15:40,040 --> 00:15:43,200
Speaker 5: figuring out bugs, then people are sending patches. I think

279
00:15:43,200 --> 00:15:45,560
Speaker 5: if you look at it at the enterprise level, this

280
00:15:45,680 --> 00:15:49,080
Speaker 5: has the potential to be absolutely enormous, and I think

281
00:15:49,080 --> 00:15:51,600
Speaker 5: in ways that we don't even fully understand, you know.

282
00:15:51,640 --> 00:15:53,560
Speaker 5: I think if you can fully automate the work that

283
00:15:53,960 --> 00:15:56,320
Speaker 5: say an intern does or an entry level worker does,

284
00:15:56,400 --> 00:15:58,280
Speaker 5: I think that can really have profound effects on the

285
00:15:58,320 --> 00:15:58,960
Speaker 5: labor market.

286
00:16:00,280 --> 00:16:03,600
Speaker 2: And in terms of the flip side, where could this

287
00:16:03,720 --> 00:16:04,560
Speaker 2: experiment fail.

288
00:16:05,240 --> 00:16:07,680
Speaker 5: I think in the longer term, you know, as much

289
00:16:07,760 --> 00:16:10,760
Speaker 5: as the tech industry likes to say that AI agents

290
00:16:10,800 --> 00:16:14,200
Speaker 5: are there to empower workers. I don't necessarily buy that.

291
00:16:14,320 --> 00:16:17,880
Speaker 5: I think the goal is really to replace workers, and

292
00:16:17,920 --> 00:16:19,640
Speaker 5: so I think in China that's actually going to be

293
00:16:19,640 --> 00:16:22,320
Speaker 5: a really big thing to watch because youth unemployment has

294
00:16:22,400 --> 00:16:25,840
Speaker 5: remained persistently high since the pandemic, and I think if

295
00:16:25,880 --> 00:16:29,840
Speaker 5: you have large swaths of educated and unemployed young people

296
00:16:29,920 --> 00:16:31,640
Speaker 5: that are looking for jobs and that are being put

297
00:16:31,640 --> 00:16:34,080
Speaker 5: out of work by these AI tools, I think that

298
00:16:34,400 --> 00:16:36,680
Speaker 5: doesn't bode well for social stability, and I think that

299
00:16:36,720 --> 00:16:39,200
Speaker 5: will be sort of a big test for the Chinese government.

300
00:16:39,680 --> 00:16:41,560
Speaker 5: I think in the near term we'll have to see

301
00:16:41,920 --> 00:16:44,880
Speaker 5: how much this actually sticks and if people are actually

302
00:16:45,480 --> 00:16:47,560
Speaker 5: able to build trust with AI agents.

303
00:16:47,880 --> 00:16:52,320
Speaker 2: And ultimately, Catherine says, how China handles tools like Openclaw

304
00:16:52,720 --> 00:16:56,080
Speaker 2: could shape the next phase of the global AI race.

305
00:16:57,040 --> 00:17:00,400
Speaker 5: Building the top of the line frontier models is side

306
00:17:00,440 --> 00:17:02,160
Speaker 5: of it, and then the other side of the sort

307
00:17:02,200 --> 00:17:04,840
Speaker 5: of AI race is diffusion and how do you integrate

308
00:17:04,880 --> 00:17:07,680
Speaker 5: AI into as many facets of the economy and as

309
00:17:07,680 --> 00:17:10,360
Speaker 5: many facets of society as possible. And I think we're

310
00:17:10,400 --> 00:17:13,800
Speaker 5: seeing China really pick up on the diffusion race. We're

311
00:17:13,840 --> 00:17:17,680
Speaker 5: really seeing that they're willing to try out this new technology,

312
00:17:17,760 --> 00:17:20,840
Speaker 5: willing to see how it can increase productivity in their

313
00:17:20,840 --> 00:17:24,000
Speaker 5: personal lives and their work lives, which I think actually

314
00:17:24,280 --> 00:17:26,159
Speaker 5: is the race that we should be watching, because I

315
00:17:26,160 --> 00:17:29,000
Speaker 5: think that's where you really see the biggest benefits from

316
00:17:29,040 --> 00:17:33,080
Speaker 5: major tech evolutions in history. I think from that perspective,

317
00:17:33,320 --> 00:17:36,120
Speaker 5: it definitely shows that China is sort of leaping ahead

318
00:17:36,119 --> 00:17:39,320
Speaker 5: when it comes to AI. They're embracing it, they're using it,

319
00:17:39,359 --> 00:17:42,200
Speaker 5: they're experimenting, and I think in doing so, they're also

320
00:17:42,280 --> 00:17:44,880
Speaker 5: sort of making their own models better. You know, they're

321
00:17:44,920 --> 00:17:47,919
Speaker 5: giving real world feedback to their own models, and I

322
00:17:47,920 --> 00:17:50,199
Speaker 5: think that just only improves them. So I think that

323
00:17:50,200 --> 00:17:52,160
Speaker 5: could really tip the scales in the long run when

324
00:17:52,160 --> 00:17:54,440
Speaker 5: it comes to this broader sort of AI race.

325
00:18:03,800 --> 00:18:07,119
Speaker 3: This is The Big Take Asia from Bloomberg News. I'm wanha.

326
00:18:07,440 --> 00:18:10,120
Speaker 2: To get more from The Big Take and unlimited access

327
00:18:10,160 --> 00:18:13,120
Speaker 2: to all of Bloomberg dot Com, subscribe today at Bloomberg

328
00:18:13,160 --> 00:18:16,440
Speaker 2: dot com slash podcast Offer. If you like the episode,

329
00:18:16,600 --> 00:18:19,159
Speaker 2: make sure to subscribe and review The Big Take Asia

330
00:18:19,240 --> 00:18:20,880
Speaker 2: wherever you listen to podcasts.

331
00:18:20,960 --> 00:18:23,720
Speaker 3: It really helps people find the show. Thanks for listening,

332
00:18:23,840 --> 00:18:24,560
Speaker 3: See you next time.