1
00:00:10,614 --> 00:00:15,294
Speaker 1: You're listening to a Muma Mia podcast. Mamma Mia acknowledges

2
00:00:15,334 --> 00:00:18,174
Speaker 1: the traditional owners of land and waters that this podcast

3
00:00:18,214 --> 00:00:19,054
Speaker 1: is recorded on.

4
00:00:21,974 --> 00:00:22,174
Speaker 2: Hi.

5
00:00:22,294 --> 00:00:25,454
Speaker 1: I'm Claire Murphy. This is Mumma MIA's daily news podcast,

6
00:00:25,494 --> 00:00:28,574
Speaker 1: The Quickie. Have you come across videos like this in

7
00:00:28,614 --> 00:00:29,574
Speaker 1: your scrolling of late?

8
00:00:29,894 --> 00:00:32,774
Speaker 3: Manner is actually updating its privacy policy. I wonder why

9
00:00:32,814 --> 00:00:34,774
Speaker 3: that is, Probably because on June twenty six they are

10
00:00:34,774 --> 00:00:36,974
Speaker 3: going to be using your public photos and your public

11
00:00:36,974 --> 00:00:38,894
Speaker 3: fosts to train its AI. And if you don't want

12
00:00:38,894 --> 00:00:41,374
Speaker 3: that to happen, which I personally don't, you can opt out,

13
00:00:41,414 --> 00:00:43,414
Speaker 3: and they're making it a little difficult to do so,

14
00:00:43,414 --> 00:00:44,654
Speaker 3: so I'm going to show you how to do that.

15
00:00:45,454 --> 00:00:48,294
Speaker 1: Yep, Meta are less than ten days away from using

16
00:00:48,334 --> 00:00:52,294
Speaker 1: your public Facebook and Instagram posts to train its AI tools,

17
00:00:52,494 --> 00:00:55,294
Speaker 1: and there are plenty of digital activists who are very

18
00:00:55,374 --> 00:00:57,934
Speaker 1: much against it. Before we find out what everyone is

19
00:00:57,974 --> 00:01:00,934
Speaker 1: so concerned about, here's the latest from the Quikie newsroom. Monday,

20
00:01:00,974 --> 00:01:05,534
Speaker 1: June seventeen, the Israeli military will hold daily tactical pauses

21
00:01:05,574 --> 00:01:08,294
Speaker 1: in their military activity in Gaza to allow aid to

22
00:01:08,374 --> 00:01:11,414
Speaker 1: flow into the enclave. The IDF said that fighting in

23
00:01:11,454 --> 00:01:14,694
Speaker 1: the southern city of Rafah would continue as they target

24
00:01:14,774 --> 00:01:18,094
Speaker 1: the remaining Hamas brigades, but would pause from five am

25
00:01:18,214 --> 00:01:21,054
Speaker 1: to four pm local time daily along the road that

26
00:01:21,174 --> 00:01:24,374
Speaker 1: leads to the Koremshalom crossing to allow aid trucks to

27
00:01:24,534 --> 00:01:28,414
Speaker 1: enter the region, where international aid organizations have continually been

28
00:01:28,454 --> 00:01:32,934
Speaker 1: warning of a growing humanitarian crisis. Despite growing international pressure

29
00:01:32,974 --> 00:01:35,894
Speaker 1: for a ceasefire, an agreement to halt the fighting still

30
00:01:35,934 --> 00:01:38,814
Speaker 1: appears to be far off. More than eight months after

31
00:01:38,854 --> 00:01:42,054
Speaker 1: the start of the war in Gaza, economists are broadly

32
00:01:42,094 --> 00:01:44,894
Speaker 1: expecting rates to remain on hold as members of the

33
00:01:44,894 --> 00:01:47,814
Speaker 1: Reserve Bank of Australia assemble for the June interst rate

34
00:01:47,894 --> 00:01:51,294
Speaker 1: meeting today. While inflation is well down from its peak,

35
00:01:51,574 --> 00:01:55,694
Speaker 1: inflation accelerated in April, suggesting stretched borrowers will have to

36
00:01:55,734 --> 00:01:59,294
Speaker 1: wait a little longer for those interest rate cuts. Economic

37
00:01:59,334 --> 00:02:01,414
Speaker 1: teams at all four of the big banks are expecting

38
00:02:01,494 --> 00:02:04,094
Speaker 1: rates to remain on hold at four point three five percent,

39
00:02:04,414 --> 00:02:07,494
Speaker 1: predicting the cuts to come in February twenty twenty five.

40
00:02:08,214 --> 00:02:11,454
Speaker 1: As Prime Minister Anthony Babernesi prepares to meet with China's

41
00:02:11,494 --> 00:02:15,494
Speaker 1: Premier Le Chiang today, supporters of writer Yangheng Juna calling

42
00:02:15,534 --> 00:02:18,574
Speaker 1: on the PM to raise his case. Lee is making

43
00:02:18,574 --> 00:02:22,534
Speaker 1: his first visit to Australia in seven years, Albanesi promising

44
00:02:22,614 --> 00:02:25,654
Speaker 1: to raise the case with China's second highest strength official.

45
00:02:26,014 --> 00:02:29,294
Speaker 1: After a Beijing court handed the Australian citizen a suspended

46
00:02:29,334 --> 00:02:33,414
Speaker 1: death sentence back in February, accusing the writer of espionage.

47
00:02:33,574 --> 00:02:37,414
Speaker 1: Yang has denied the charges. Beijing's High People's Court reviewed

48
00:02:37,454 --> 00:02:40,894
Speaker 1: and upheld the lower court sentence, in forming Australian diplomats

49
00:02:40,894 --> 00:02:44,414
Speaker 1: on May thirty, denying them a consular visit and moving

50
00:02:44,494 --> 00:02:47,054
Speaker 1: Yang to a permanent prison after five years in a

51
00:02:47,094 --> 00:02:51,734
Speaker 1: state security detention center. Prince William's three children have written

52
00:02:51,774 --> 00:02:55,134
Speaker 1: their dad a Father's Day message, which his wife, Princess Catherine,

53
00:02:55,134 --> 00:02:57,654
Speaker 1: has posted alongside a photo she took of them at

54
00:02:57,654 --> 00:03:00,854
Speaker 1: the beach. The message simply says we love you Papa,

55
00:03:00,894 --> 00:03:04,334
Speaker 1: Happy Father's Day, GC and L. The photo is of

56
00:03:04,334 --> 00:03:06,614
Speaker 1: William and the three kids their backs to the camera

57
00:03:06,974 --> 00:03:09,734
Speaker 1: or with their arms around each other looking out to see.

58
00:03:09,814 --> 00:03:12,654
Speaker 1: William also posted a throwback shot of him playing ball

59
00:03:12,694 --> 00:03:15,454
Speaker 1: with his dad as a toddler, writing Happy Father's Day,

60
00:03:15,494 --> 00:03:18,854
Speaker 1: part w The photos release came a day after the

61
00:03:18,854 --> 00:03:21,694
Speaker 1: Princess of Wales's return to royal duties for the first

62
00:03:21,694 --> 00:03:25,614
Speaker 1: time since revealing she'd been diagnosed with cancer, joining King

63
00:03:25,694 --> 00:03:28,414
Speaker 1: Charles on the balcony of Buckingham Palace after watching the

64
00:03:28,414 --> 00:03:31,414
Speaker 1: military parade to celebrate the monarch's official birthday.

65
00:03:31,814 --> 00:03:33,134
Speaker 2: That's what's happening in the world today.

66
00:03:33,174 --> 00:03:37,374
Speaker 1: Next, Meta has let us know that our publicly posted photos, videos,

67
00:03:37,374 --> 00:03:40,174
Speaker 1: and more are about to be used to train their AI.

68
00:03:40,694 --> 00:03:43,414
Speaker 1: But what does that mean for us and why digital

69
00:03:43,454 --> 00:03:56,734
Speaker 1: rights activists fighting it. Recently, Gratzier sent us an email

70
00:03:56,774 --> 00:03:58,614
Speaker 1: with a link to a story asking us to look

71
00:03:58,614 --> 00:03:59,614
Speaker 1: into it a bit more.

72
00:03:59,654 --> 00:04:02,334
Speaker 2: For her. The story was about Meta.

73
00:04:02,414 --> 00:04:06,774
Speaker 4: Meta, the company behind social media juggernaut's Facebook and Instagram.

74
00:04:06,374 --> 00:04:08,294
Speaker 2: Are coming for your data.

75
00:04:08,414 --> 00:04:11,454
Speaker 1: They wanted to train the artificial intelligence and if you

76
00:04:11,494 --> 00:04:12,774
Speaker 1: don't agree, well then too bad.

77
00:04:12,974 --> 00:04:14,374
Speaker 2: You can't say no yes.

78
00:04:14,414 --> 00:04:19,054
Speaker 1: From June twenty six, Meta, the parent company of Facebook, Messenger, WhatsApp,

79
00:04:19,094 --> 00:04:22,654
Speaker 1: and Instagram, are updating their privacy policy. You might have

80
00:04:22,694 --> 00:04:26,254
Speaker 1: seen it already, and maybe, like me, you completely ignored it.

81
00:04:26,734 --> 00:04:29,974
Speaker 1: A message in your Facebook or Instagram notifications that read

82
00:04:30,054 --> 00:04:30,534
Speaker 1: like this.

83
00:04:30,854 --> 00:04:35,254
Speaker 5: We're planning UAI features for you. Learn how we use

84
00:04:35,334 --> 00:04:36,254
Speaker 5: your information.

85
00:04:37,054 --> 00:04:39,614
Speaker 1: From June twenty six, the company will be allowed to

86
00:04:39,694 --> 00:04:44,094
Speaker 1: use the photos, videos, captions, messages, stories and posts of

87
00:04:44,254 --> 00:04:46,934
Speaker 1: anyone over the age of eighteen and from as far

88
00:04:46,974 --> 00:04:50,454
Speaker 1: back as two thousand and seven to train their AI systems.

89
00:04:51,054 --> 00:04:53,934
Speaker 1: Whether your account is public or private, Meta will be

90
00:04:54,054 --> 00:04:58,014
Speaker 1: scraping it. Only your dms are safe and sorry, my

91
00:04:58,134 --> 00:05:01,094
Speaker 1: fellow Wazzies. But unlike our friends in the EU or

92
00:05:01,134 --> 00:05:04,014
Speaker 1: in some US states, whether are laws giving people the

93
00:05:04,014 --> 00:05:07,374
Speaker 1: option to opt out of how Meta uses their data. Here,

94
00:05:07,694 --> 00:05:10,934
Speaker 1: we have no such laws, so our personal info is

95
00:05:11,054 --> 00:05:14,134
Speaker 1: just sitting there free for Meta to AI away at

96
00:05:14,174 --> 00:05:19,934
Speaker 1: it as it likes. We don't know exactly what this

97
00:05:20,054 --> 00:05:23,614
Speaker 1: AI training will be eventually used for, but the company

98
00:05:23,654 --> 00:05:27,174
Speaker 1: has already been shifting the weight integrates AI into your

99
00:05:27,214 --> 00:05:28,614
Speaker 1: social media experience.

100
00:05:29,014 --> 00:05:32,174
Speaker 6: We are releasing the new version of metai, our assistant

101
00:05:32,254 --> 00:05:35,214
Speaker 6: that you can ask any question across our apps and glasses,

102
00:05:35,694 --> 00:05:38,254
Speaker 6: and our goal is to build the world's leading AI

103
00:05:38,574 --> 00:05:40,374
Speaker 6: and make it available to everyone.

104
00:05:40,934 --> 00:05:44,134
Speaker 1: The Meta AI chatbots are in the Facebook, Messenger and

105
00:05:44,214 --> 00:05:47,814
Speaker 1: WhatsApp services. Now users able to ask it anything from

106
00:05:47,934 --> 00:05:51,134
Speaker 1: helping plan a holiday to settling an argument between friends.

107
00:05:51,654 --> 00:05:55,094
Speaker 1: But it's not just a text based relationship. These days,

108
00:05:55,534 --> 00:05:58,294
Speaker 1: there are celebs who've signed up to lend their persona

109
00:05:58,414 --> 00:06:01,814
Speaker 1: to some of these bots, including Snoop Dogg and Kendall Jenner.

110
00:06:02,214 --> 00:06:05,134
Speaker 1: And they don't just answer a question. They're to add

111
00:06:05,174 --> 00:06:08,254
Speaker 1: another layer to your search where you can debate sports

112
00:06:08,334 --> 00:06:11,734
Speaker 1: with a Tom Brady style AI chatbot, or get roasted

113
00:06:11,774 --> 00:06:14,694
Speaker 1: by Zack, a Big Brother bot based on famous YouTuber

114
00:06:14,774 --> 00:06:18,654
Speaker 1: mister Beast. It's a two way conversation, not just asking

115
00:06:18,654 --> 00:06:21,254
Speaker 1: a machine a question. All of this is leading us

116
00:06:21,294 --> 00:06:24,774
Speaker 1: closer and closer to the metaverse, a project. Meta founder

117
00:06:24,774 --> 00:06:28,414
Speaker 1: Mark Zuckerberg has been investing tens of billions of dollars

118
00:06:28,454 --> 00:06:32,014
Speaker 1: in creating augmented reality headsets so we never.

119
00:06:31,814 --> 00:06:32,814
Speaker 2: Have to disconnect.

120
00:06:33,014 --> 00:06:34,574
Speaker 6: No, God, please no.

121
00:06:35,134 --> 00:06:35,814
Speaker 2: But what are the.

122
00:06:35,734 --> 00:06:40,934
Speaker 1: Worst case scenarios? Maybe you're an artist who posts images.

123
00:06:40,534 --> 00:06:41,094
Speaker 2: Of your work.

124
00:06:41,334 --> 00:06:45,014
Speaker 1: Will the AI bots use that to create images? Will

125
00:06:45,054 --> 00:06:47,934
Speaker 1: your likeness be used to create AI images of people

126
00:06:47,974 --> 00:06:51,054
Speaker 1: in AI generated movies? Will your voice be used to

127
00:06:51,054 --> 00:06:55,174
Speaker 1: create AI ads? Some digital rights activists have called the

128
00:06:55,254 --> 00:06:58,494
Speaker 1: move an abuse of personal data, and some have filed

129
00:06:58,494 --> 00:07:01,894
Speaker 1: complaints with local authorities across the world trying to stop

130
00:07:01,934 --> 00:07:05,734
Speaker 1: the scrape from happening, but Meta maintains that it's confident

131
00:07:05,854 --> 00:07:09,094
Speaker 1: it complies with privacy laws, in line with how other

132
00:07:09,094 --> 00:07:12,094
Speaker 1: Big ten firms have used their users data to develop

133
00:07:12,174 --> 00:07:18,614
Speaker 1: their own AI experiences. So should we be worried? Toby

134
00:07:18,654 --> 00:07:21,814
Speaker 1: Walsh is a Laureate Fellow and Scientia Professor of Artificial

135
00:07:21,854 --> 00:07:25,134
Speaker 1: Intelligence at the Department of Computer Science and Engineering at

136
00:07:25,134 --> 00:07:28,814
Speaker 1: the University of New South Wales. Toby, is this data

137
00:07:28,814 --> 00:07:32,734
Speaker 1: scripe something you think will actually impact us as individuals

138
00:07:32,774 --> 00:07:35,614
Speaker 1: who use any of the Meta suite of social media platforms.

139
00:07:36,254 --> 00:07:39,134
Speaker 7: They're wanting to train on your data because that allows

140
00:07:39,214 --> 00:07:44,734
Speaker 7: them to deliver a much more individualized, personalized AI Howard service.

141
00:07:45,094 --> 00:07:46,934
Speaker 7: So on one level, that sounds like that's good, and

142
00:07:47,054 --> 00:07:49,774
Speaker 7: it will be the tools will be much more responsive,

143
00:07:49,934 --> 00:07:53,854
Speaker 7: much more useful to us. But equally it's also what

144
00:07:53,854 --> 00:07:56,214
Speaker 7: else are they going to do with that information that

145
00:07:56,254 --> 00:07:58,374
Speaker 7: they've found out about us? Are they're going to be

146
00:07:58,814 --> 00:08:01,814
Speaker 7: setting us the latest crypto coin? Are they're going to

147
00:08:01,854 --> 00:08:04,574
Speaker 7: be setting us more cat food? Because we're a cat lover.

148
00:08:05,294 --> 00:08:09,094
Speaker 1: So basically the concern should be that they've taken our

149
00:08:10,654 --> 00:08:13,334
Speaker 1: and they're putting us in sort of an endless feedback

150
00:08:13,374 --> 00:08:18,294
Speaker 1: loop of our own choosing and done that by sort

151
00:08:18,334 --> 00:08:21,014
Speaker 1: of exploiting the things that we've provided to them.

152
00:08:21,054 --> 00:08:23,734
Speaker 4: Am I right, Yes, what they're going to do is

153
00:08:23,774 --> 00:08:26,654
Speaker 4: going to be much more power, much more persuades it

154
00:08:26,934 --> 00:08:28,934
Speaker 4: than what they've been able to do with the social

155
00:08:28,974 --> 00:08:32,054
Speaker 4: media data that they've trained on the past, because they're

156
00:08:32,094 --> 00:08:34,934
Speaker 4: going to be able to generate, for example, synthetic voices

157
00:08:34,974 --> 00:08:35,534
Speaker 4: that sound.

158
00:08:35,374 --> 00:08:37,974
Speaker 7: Like our voices, or synthetic voices that sound like the

159
00:08:38,054 --> 00:08:41,254
Speaker 7: voices of our best friends, videos that include us saying

160
00:08:41,294 --> 00:08:44,494
Speaker 7: things that we never said. We should be outraged they're

161
00:08:44,494 --> 00:08:47,454
Speaker 7: doing it, and there's no way that we can opt

162
00:08:47,494 --> 00:08:48,614
Speaker 7: out of them doing it.

163
00:08:48,694 --> 00:08:50,494
Speaker 4: If you are Ound Australia, and.

164
00:08:50,454 --> 00:08:52,214
Speaker 1: I think a lot of the worry is that AI

165
00:08:52,334 --> 00:08:55,614
Speaker 1: will be trained on using our photos and videos in

166
00:08:55,654 --> 00:09:00,174
Speaker 1: particular to create essentially something that might be made in

167
00:09:00,214 --> 00:09:03,734
Speaker 1: our likeness. Could they use our faces, for example, to

168
00:09:03,814 --> 00:09:07,894
Speaker 1: create an AI persona that could be used in a

169
00:09:07,934 --> 00:09:11,534
Speaker 1: movie for example, Or could they use our voices to

170
00:09:11,574 --> 00:09:15,414
Speaker 1: create AI advertising? Could they use our personal data for

171
00:09:15,454 --> 00:09:16,014
Speaker 1: things like that?

172
00:09:16,454 --> 00:09:19,014
Speaker 7: It's technically possible that they could do that. Your best

173
00:09:19,054 --> 00:09:21,854
Speaker 7: friend could be trying to sell you some product in

174
00:09:21,894 --> 00:09:24,814
Speaker 7: the future because they have that data from placebook. They

175
00:09:24,894 --> 00:09:27,094
Speaker 7: know who your friend is, and they also know what

176
00:09:27,214 --> 00:09:30,214
Speaker 7: your best friend sounds like. Are need a few seconds

177
00:09:30,334 --> 00:09:33,454
Speaker 7: of someone's voice to be able to clone their voice? Equally,

178
00:09:33,454 --> 00:09:35,094
Speaker 7: it might not just be their voice, it might be

179
00:09:35,134 --> 00:09:37,534
Speaker 7: a video of that person. So yes, I think we

180
00:09:37,574 --> 00:09:40,494
Speaker 7: should be concerned how they do these things. Let me

181
00:09:40,574 --> 00:09:43,854
Speaker 7: not forget. Also, it was exactly those sorts of technologies

182
00:09:43,894 --> 00:09:47,854
Speaker 7: which used disastrously in that school in Melbourne to create

183
00:09:47,854 --> 00:09:49,654
Speaker 7: all those detate news.

184
00:09:50,174 --> 00:09:53,774
Speaker 1: Other platforms already scrape data for this purpose. If you

185
00:09:53,814 --> 00:09:58,214
Speaker 1: look through the privacy policy of Twitter Akax, for example,

186
00:09:58,334 --> 00:10:00,774
Speaker 1: they state very clearly that they're using anything that you

187
00:10:00,894 --> 00:10:05,454
Speaker 1: post to train their ai. Reddit has even sold their

188
00:10:05,534 --> 00:10:09,774
Speaker 1: data to others like Google and open Ai to use

189
00:10:09,854 --> 00:10:13,214
Speaker 1: to train their AI bots. Is this something that's just

190
00:10:13,494 --> 00:10:16,174
Speaker 1: unavoidable now? If we're putting things out into a public

191
00:10:16,174 --> 00:10:18,574
Speaker 1: space freely handing over this data as we have been

192
00:10:18,614 --> 00:10:21,454
Speaker 1: doing for you know, a couple of decades now, is

193
00:10:21,494 --> 00:10:23,454
Speaker 1: this just something we should expect to happen to us.

194
00:10:23,934 --> 00:10:26,174
Speaker 7: No, it's not something that we should expect to happen.

195
00:10:26,294 --> 00:10:29,454
Speaker 7: I think we should be demanding that we have similar

196
00:10:29,534 --> 00:10:33,214
Speaker 7: data protection rights as citizens In other countries, and in Europe,

197
00:10:33,254 --> 00:10:36,614
Speaker 7: for example, they do have much stricter data privacy rights,

198
00:10:36,734 --> 00:10:39,774
Speaker 7: and the government in this country has been slow to

199
00:10:39,934 --> 00:10:43,774
Speaker 7: update the data protection and regulation that we have here.

200
00:10:44,094 --> 00:10:48,214
Speaker 7: For example, in Europe, you could demand that your datado.

201
00:10:48,494 --> 00:10:50,854
Speaker 7: We don't have that right here, which is why we

202
00:10:50,934 --> 00:10:54,134
Speaker 7: can't say to Facebook or any other companies, no, actually

203
00:10:54,134 --> 00:10:56,494
Speaker 7: I want you to forget that data that you've collected

204
00:10:56,534 --> 00:10:56,974
Speaker 7: from there.

205
00:10:57,734 --> 00:11:00,654
Speaker 1: If you were to delete all of your meta suite

206
00:11:00,654 --> 00:11:03,774
Speaker 1: of social media accounts, can they still use that to

207
00:11:03,814 --> 00:11:04,494
Speaker 1: scripe data.

208
00:11:05,014 --> 00:11:08,814
Speaker 7: Unfortunately, they've already taken snapshots of that data, so even

209
00:11:08,814 --> 00:11:12,494
Speaker 7: if you delete it from your current account, they will

210
00:11:12,494 --> 00:11:15,414
Speaker 7: have the records about moving backwards in the past. Again,

211
00:11:15,454 --> 00:11:18,174
Speaker 7: that's why we would need regulation where in places like

212
00:11:18,214 --> 00:11:20,774
Speaker 7: Europe you can actually ask them to delete those records.

213
00:11:21,174 --> 00:11:24,774
Speaker 1: What laws do we need to enact to ensure that

214
00:11:24,814 --> 00:11:27,334
Speaker 1: this doesn't happen to us and it is already happening,

215
00:11:27,374 --> 00:11:29,934
Speaker 1: even just to me personally. The company I work for,

216
00:11:30,014 --> 00:11:33,294
Speaker 1: Mum and Mea, created an AI voice training it using

217
00:11:33,574 --> 00:11:35,974
Speaker 1: my voice and others. They didn't need my permission to

218
00:11:36,014 --> 00:11:38,574
Speaker 1: do so they still don't. What law should we be

219
00:11:38,574 --> 00:11:41,454
Speaker 1: putting in place to make sure that our permission is

220
00:11:41,494 --> 00:11:43,134
Speaker 1: given before this stuff is handed over.

221
00:11:43,734 --> 00:11:47,414
Speaker 7: We need protection and a number of different ways. I

222
00:11:47,414 --> 00:11:50,214
Speaker 7: think we need to hold the social media platforms more accountable,

223
00:11:50,254 --> 00:11:53,574
Speaker 7: so when they distribute content, we actually hold them to

224
00:11:53,614 --> 00:11:56,214
Speaker 7: the safest of standards. We hold old fashion media like

225
00:11:56,254 --> 00:11:59,294
Speaker 7: the ABC. Some of this content, of the distributing it

226
00:11:59,454 --> 00:12:01,734
Speaker 7: only becomes harmful when it's put in front of lots

227
00:12:01,734 --> 00:12:03,694
Speaker 7: of eyeboards, And so I think you need to hold

228
00:12:03,734 --> 00:12:06,254
Speaker 7: the tech companies more accountable. As an example, the fact

229
00:12:06,294 --> 00:12:10,414
Speaker 7: that you can so easily access and use tools that

230
00:12:10,574 --> 00:12:13,454
Speaker 7: create de fake news, I think again, why do we

231
00:12:13,494 --> 00:12:16,534
Speaker 7: allow people to sell such tools and distribute such tools.

232
00:12:17,014 --> 00:12:19,094
Speaker 7: They only seem to be bringing harm into our lives,

233
00:12:19,094 --> 00:12:21,614
Speaker 7: not benefits. And then finally, I think we need laws

234
00:12:21,654 --> 00:12:24,854
Speaker 7: around our data privacy and data protection that allow us

235
00:12:24,934 --> 00:12:25,974
Speaker 7: a right to be forgotten.

236
00:12:26,574 --> 00:12:29,494
Speaker 1: Soby we keep hearing that phrase social media platform should

237
00:12:29,494 --> 00:12:31,774
Speaker 1: be held more accountable. We hear it over and over

238
00:12:31,814 --> 00:12:34,894
Speaker 1: in many different ways, attached to many different stories, and

239
00:12:34,974 --> 00:12:39,134
Speaker 1: it doesn't seem like they ever truly are held accountable.

240
00:12:39,374 --> 00:12:41,654
Speaker 1: Even when they're in front of courts. They don't tend

241
00:12:41,694 --> 00:12:44,534
Speaker 1: to really end up facing any consequences for the things

242
00:12:44,574 --> 00:12:47,974
Speaker 1: that they have created. How do we actually hold social

243
00:12:48,014 --> 00:12:50,134
Speaker 1: media platforms accountable for this kind of stuff?

244
00:12:50,854 --> 00:12:53,534
Speaker 7: I think it's entirely possible to hold the social media

245
00:12:53,574 --> 00:12:57,054
Speaker 7: platforms more accountable. Regulators actually discover quite a useful tool,

246
00:12:57,054 --> 00:13:00,134
Speaker 7: which is if you try to find them some fractions

247
00:13:00,494 --> 00:13:03,654
Speaker 7: four percent or five percent of their global turnover, then

248
00:13:03,974 --> 00:13:06,214
Speaker 7: that's serious money for them, and they actually do pay attention.

249
00:13:06,494 --> 00:13:08,654
Speaker 7: So I think it is possible to regulate and to

250
00:13:08,734 --> 00:13:11,494
Speaker 7: ensure that they do a better job. We did this,

251
00:13:11,894 --> 00:13:15,574
Speaker 7: for example, after the tragedies in christ Church, that terrorist

252
00:13:15,574 --> 00:13:19,974
Speaker 7: act that happened, we pass laws here in Australia saying

253
00:13:20,134 --> 00:13:23,854
Speaker 7: to the social media platforms, if such content is not

254
00:13:24,014 --> 00:13:26,934
Speaker 7: taken down quickly enough, we will hold the officers of

255
00:13:26,974 --> 00:13:30,374
Speaker 7: those companies criminally liable. Now that's not being perfect, but

256
00:13:30,894 --> 00:13:33,534
Speaker 7: the data I've seen all suggest that those platforms now

257
00:13:33,574 --> 00:13:36,374
Speaker 7: take down that content a lot quicker than they used

258
00:13:36,374 --> 00:13:38,734
Speaker 7: to because of the threat of those laws.

259
00:13:39,334 --> 00:13:41,054
Speaker 1: Will we eventually end up in a bit of an

260
00:13:41,094 --> 00:13:44,414
Speaker 1: AI loop, though, Toby, because I'm seeing personally more and

261
00:13:44,454 --> 00:13:47,534
Speaker 1: more AI generated content on my own social media feeds,

262
00:13:47,814 --> 00:13:52,094
Speaker 1: and if AI is using these platforms to train itself,

263
00:13:52,574 --> 00:13:55,214
Speaker 1: won't it essentially be then training itself eventually?

264
00:13:55,814 --> 00:13:59,534
Speaker 7: That's something actually quite worrying. We may look back at

265
00:13:59,534 --> 00:14:03,174
Speaker 7: twenty twenty four as the golden made there when ninety

266
00:14:03,214 --> 00:14:07,054
Speaker 7: nine percent of the content was high quality human generated content,

267
00:14:07,694 --> 00:14:10,854
Speaker 7: because in the future, maybe that ninety nine percent of

268
00:14:10,854 --> 00:14:14,534
Speaker 7: the content is synthetic, machine generated content. And that's going

269
00:14:14,614 --> 00:14:17,254
Speaker 7: to not be such a good world in which human

270
00:14:17,334 --> 00:14:20,254
Speaker 7: voices are drowned out by the computer voices. And we

271
00:14:20,294 --> 00:14:23,094
Speaker 7: already know that that's going to be quite dangerous. That

272
00:14:23,254 --> 00:14:25,974
Speaker 7: any small biases that exist in the content that we

273
00:14:26,054 --> 00:14:30,174
Speaker 7: have today, every time we train the next generation on

274
00:14:30,214 --> 00:14:33,254
Speaker 7: the outputs of the previous generation, those small bises tend

275
00:14:33,334 --> 00:14:36,694
Speaker 7: to get amplified, and so any racism or sexism or

276
00:14:36,694 --> 00:14:39,214
Speaker 7: agism or whatever it is that we be worried about will

277
00:14:39,254 --> 00:14:42,294
Speaker 7: get amplified. Therefore, we may look back and think, oh, well,

278
00:14:42,334 --> 00:14:44,414
Speaker 7: that was when the Internet was good, when it was human.

279
00:14:48,494 --> 00:14:51,814
Speaker 1: Scrolling through a Reddit thread talking about this decision by Meta,

280
00:14:51,854 --> 00:14:54,894
Speaker 1: and one comment summed up exactly what it means to

281
00:14:54,934 --> 00:14:57,334
Speaker 1: be a user of a social media platform.

282
00:14:57,494 --> 00:15:01,214
Speaker 5: Facebook is not a product for users. It's users are

283
00:15:01,254 --> 00:15:02,934
Speaker 5: the products for Facebook.

284
00:15:03,574 --> 00:15:07,774
Speaker 2: Without us, Meta technically has nothing. But since they do.

285
00:15:07,774 --> 00:15:10,734
Speaker 1: Already have us, they will use what we have freely

286
00:15:10,854 --> 00:15:13,494
Speaker 1: given them, and less laws are put in place to

287
00:15:13,534 --> 00:15:14,294
Speaker 1: stop them.

288
00:15:14,414 --> 00:15:16,774
Speaker 6: In the meantime, Join that AI, I mean, know what

289
00:15:16,814 --> 00:15:17,134
Speaker 6: you think.

290
00:15:18,974 --> 00:15:21,694
Speaker 1: Thanks for tuning in today, friends. The Quickie is produced

291
00:15:21,734 --> 00:15:25,054
Speaker 1: by me Claire Murphy and our executive producer Callie Rintle,

292
00:15:25,174 --> 00:15:27,014
Speaker 1: with audio production by Tom Lyin.