1
00:00:12,814 --> 00:00:15,734
Speaker 1: Hi, I'm Claire Murphy. This is Mumma MEA's twice daily

2
00:00:15,774 --> 00:00:19,174
Speaker 1: news podcast, The Quickie. It's coming at us, at work

3
00:00:19,294 --> 00:00:21,774
Speaker 1: and in our private lives. We're being told use it

4
00:00:21,854 --> 00:00:24,734
Speaker 1: now or get left behind. We're being told it's the future,

5
00:00:24,814 --> 00:00:27,974
Speaker 1: so get on board. But what AI tools can we

6
00:00:28,094 --> 00:00:29,054
Speaker 1: actually trust?

7
00:00:29,694 --> 00:00:29,934
Speaker 2: From?

8
00:00:30,014 --> 00:00:33,494
Speaker 1: AI assistance like chat, GBT, Gemini and deep seek to

9
00:00:33,654 --> 00:00:38,934
Speaker 1: video generators, note takers, research tools and search engines, email organizers,

10
00:00:38,974 --> 00:00:41,774
Speaker 1: and ones that take topics and turn them into a podcast.

11
00:00:42,214 --> 00:00:45,894
Speaker 1: AI is infiltrating our daily work and home lives. So

12
00:00:45,934 --> 00:00:48,454
Speaker 1: today we're going to find out which AI we can

13
00:00:48,534 --> 00:00:51,974
Speaker 1: safely use and why one in particular has been deemed

14
00:00:52,014 --> 00:00:55,814
Speaker 1: too dangerous for austraining government devices. But first, here's the

15
00:00:55,854 --> 00:00:59,734
Speaker 1: latest news headlines. Wednesday, February twelve. The jury in Matilda's

16
00:00:59,734 --> 00:01:03,374
Speaker 1: captain Sam Kerr's racial harassment trial has found her not guilty.

17
00:01:03,814 --> 00:01:06,174
Speaker 1: Kerr walked from court overnight with a smile on her

18
00:01:06,214 --> 00:01:10,174
Speaker 1: face after facing prosecutors whod accused her of course distress

19
00:01:10,254 --> 00:01:13,214
Speaker 1: to a police officer who she called stupid and white

20
00:01:13,374 --> 00:01:15,854
Speaker 1: on a night out where a taxi driver had refused

21
00:01:15,854 --> 00:01:18,814
Speaker 1: to let her and fiancee Christi meus out and instead

22
00:01:18,894 --> 00:01:21,054
Speaker 1: took them to a police station after one of them

23
00:01:21,054 --> 00:01:24,374
Speaker 1: had thrown up. The pair believed they'd been kidnapped. Kerr

24
00:01:24,414 --> 00:01:27,734
Speaker 1: explained to the court how her behavior and obvious intoxication

25
00:01:27,974 --> 00:01:30,174
Speaker 1: wasn't a great look, but that she believed she was

26
00:01:30,214 --> 00:01:32,494
Speaker 1: being treated differently due to the color of her skin.

27
00:01:32,894 --> 00:01:35,454
Speaker 1: The officer, not believing she wasn't the one who broke

28
00:01:35,494 --> 00:01:37,654
Speaker 1: the back window of the vehicle for them to escape.

29
00:01:37,974 --> 00:01:40,734
Speaker 1: A jury returned the not guilty verdict after four hours

30
00:01:40,734 --> 00:01:43,614
Speaker 1: of deliberation. Kerr gave her lawyer a thumbs up as

31
00:01:43,614 --> 00:01:46,414
Speaker 1: her partner, Christy Mewers, burst into tears in the gallery.

32
00:01:46,854 --> 00:01:48,974
Speaker 1: Getting a home loan is about to become easier for

33
00:01:48,974 --> 00:01:52,414
Speaker 1: those saddled with student debt after financial regulators promised to

34
00:01:52,494 --> 00:01:55,894
Speaker 1: update their guidelines on lending restrictions at the request of

35
00:01:55,934 --> 00:01:59,254
Speaker 1: Treasurer Jim Chalmers, ASIK and APRA have agreed to clarify

36
00:01:59,334 --> 00:02:03,094
Speaker 1: their guidance to lenders, which will include reducing serviceability and

37
00:02:03,134 --> 00:02:06,974
Speaker 1: reporting requirements for hexteads, meaning banks can exclude the debt

38
00:02:06,974 --> 00:02:08,894
Speaker 1: if they expect the borrower to be able to pay

39
00:02:08,934 --> 00:02:11,214
Speaker 1: it off in the New Ark future, and they recognize

40
00:02:11,254 --> 00:02:13,814
Speaker 1: that the size of a person's HEX repayment depends on

41
00:02:13,814 --> 00:02:16,774
Speaker 1: their income. Doctor Chalmers also said they're working to make

42
00:02:16,814 --> 00:02:19,374
Speaker 1: sure those with HELP debts are also treated fairly when

43
00:02:19,374 --> 00:02:21,774
Speaker 1: they want to buy a house. The banks have welcomed

44
00:02:21,774 --> 00:02:24,774
Speaker 1: the proposed changes, saying they were unsure how to interpret

45
00:02:24,774 --> 00:02:27,854
Speaker 1: the existing requirements that were essentially holding them back from

46
00:02:27,934 --> 00:02:30,854
Speaker 1: lending to some people with HEX debts, saying because HEX

47
00:02:30,974 --> 00:02:34,454
Speaker 1: is paid pre tax and the bank assesses serviceability post tax,

48
00:02:34,694 --> 00:02:37,814
Speaker 1: it was essentially a form of double counting. The Australian

49
00:02:37,814 --> 00:02:40,694
Speaker 1: Government may end up owning Regional Express Airlines if a

50
00:02:40,734 --> 00:02:44,374
Speaker 1: suitable buyer can't be found. REX administrators have created a

51
00:02:44,414 --> 00:02:47,374
Speaker 1: short list of potential bidders for the airline, saying they're

52
00:02:47,374 --> 00:02:49,854
Speaker 1: committed to finding an owner who will continue to provide

53
00:02:49,854 --> 00:02:52,774
Speaker 1: an ongoing and reasonable level of service to the regions,

54
00:02:53,054 --> 00:02:56,774
Speaker 1: saying that commitment also includes value for money and good governance.

55
00:02:57,054 --> 00:02:59,574
Speaker 1: If those bidders fall through, though, the government will work

56
00:02:59,614 --> 00:03:02,414
Speaker 1: on a contingency plan, including one that will see the

57
00:03:02,454 --> 00:03:06,414
Speaker 1: Commonwealth acquire the airline itself. If REX becomes nationalized, it'll

58
00:03:06,454 --> 00:03:08,694
Speaker 1: be the first time an Australian federal government has owned

59
00:03:08,734 --> 00:03:13,214
Speaker 1: an airline. Years after Quantus was privatized in nineteen ninety five,

60
00:03:13,814 --> 00:03:16,974
Speaker 1: new research has found that ozzie teens aren't just lazy

61
00:03:17,014 --> 00:03:19,814
Speaker 1: wanting to sleep in every day, their internal body clocks

62
00:03:19,814 --> 00:03:23,294
Speaker 1: are actually changing. The research, conducted by the Royal Children's

63
00:03:23,374 --> 00:03:26,774
Speaker 1: Hospital in Melbourne showed a teenager's body clock or diona

64
00:03:26,854 --> 00:03:29,814
Speaker 1: rhythm shifts during this time of life. The hormones more

65
00:03:29,854 --> 00:03:33,134
Speaker 1: stimulated into the evening, wiring you to stay up later,

66
00:03:33,494 --> 00:03:35,934
Speaker 1: but teens still require a lot of sleep as they're

67
00:03:35,974 --> 00:03:38,574
Speaker 1: growing and developing in their brains are changing, so they

68
00:03:38,654 --> 00:03:41,294
Speaker 1: end up needing that sleep in despite staying up later.

69
00:03:41,734 --> 00:03:45,694
Speaker 1: Doctors doing the research also noted that things like inconsistent bedtimes,

70
00:03:45,774 --> 00:03:48,414
Speaker 1: screen use, and caffeine are not helping teens get the

71
00:03:48,454 --> 00:03:51,214
Speaker 1: rest they need, with forty two percent of children having

72
00:03:51,254 --> 00:03:54,894
Speaker 1: a problem with their sleep patterns. That's your latest news headlines. Next,

73
00:03:54,974 --> 00:03:58,214
Speaker 1: how much do we tell our new AI assistance about ourselves?

74
00:03:58,454 --> 00:04:00,174
Speaker 1: And what do we need to know before we start

75
00:04:00,174 --> 00:04:08,214
Speaker 1: relying on them for everything we do? Are you using

76
00:04:08,294 --> 00:04:12,454
Speaker 1: AI at work? Yet? Maybe you're using it for personal reasons. We,

77
00:04:12,574 --> 00:04:14,974
Speaker 1: for example, here at the Quickie, use perplexity to help

78
00:04:15,054 --> 00:04:18,454
Speaker 1: us with research and chat GPT to brainstorm creative ideas.

79
00:04:18,814 --> 00:04:21,134
Speaker 1: Some of our colleagues use notebook air lam to turn

80
00:04:21,214 --> 00:04:24,774
Speaker 1: long text into fifteen minute podcast to better absorb the information.

81
00:04:25,374 --> 00:04:28,214
Speaker 1: Friends use chat GPT to make their grocery lists or

82
00:04:28,254 --> 00:04:31,494
Speaker 1: dinner plans for the week. Safe to say AI has

83
00:04:31,694 --> 00:04:34,374
Speaker 1: well and truly arrived in our lives since really hitting

84
00:04:34,374 --> 00:04:37,374
Speaker 1: its straps in twenty twenty three twenty four, But can

85
00:04:37,414 --> 00:04:40,414
Speaker 1: we use these without potentially putting our data at risk?

86
00:04:41,134 --> 00:04:45,254
Speaker 1: The Australian government recently banned Chinese AI chatbot deep Seek

87
00:04:45,254 --> 00:04:49,414
Speaker 1: from all government devices and systems, citing significant national security

88
00:04:49,454 --> 00:04:52,934
Speaker 1: concerns and data privacy risks. The ban, which took effect

89
00:04:52,934 --> 00:04:56,894
Speaker 1: earlier this month, requires all government entities to immediately remove

90
00:04:56,934 --> 00:05:00,574
Speaker 1: deep seak products and prevent their installation on official devices.

91
00:05:00,934 --> 00:05:03,854
Speaker 1: Home Affairs Minister Tony Burke said that deep Seek poses

92
00:05:03,894 --> 00:05:08,694
Speaker 1: an unacceptable risk to government technology. Following the federal directive,

93
00:05:08,814 --> 00:05:13,254
Speaker 1: most Australian states have also implemented similar bands. Queensland, Wa,

94
00:05:13,334 --> 00:05:16,374
Speaker 1: the Act and the Northern Territory have now joined South

95
00:05:16,414 --> 00:05:19,774
Speaker 1: Australia and New South Wales in prohibiting deep Seek's use

96
00:05:19,814 --> 00:05:24,054
Speaker 1: on government devices. Tasmania is still considering its position based

97
00:05:24,054 --> 00:05:28,094
Speaker 1: on security advice. The major concerns about deep Seek is

98
00:05:28,134 --> 00:05:31,974
Speaker 1: its extensive data collection practices and the potential exposure of

99
00:05:32,014 --> 00:05:36,654
Speaker 1: sensitive information to foreign government control. Cybersecurity experts have worn

100
00:05:36,694 --> 00:05:40,014
Speaker 1: that deep Sea collects data like keystroke patterns that can

101
00:05:40,054 --> 00:05:44,974
Speaker 1: identify individuals and potentially match work searches with leisure time activities.

102
00:05:45,294 --> 00:05:48,254
Speaker 1: Now this level of data collection, combined with Chinese data

103
00:05:48,254 --> 00:05:51,734
Speaker 1: sovereignty laws falling under Chinese government jurisdiction, meaning it can

104
00:05:51,774 --> 00:05:55,614
Speaker 1: be accessed by Chinese authorities at any time, has raised

105
00:05:55,654 --> 00:05:59,454
Speaker 1: significant red flags for national security officials. There are also

106
00:05:59,494 --> 00:06:03,094
Speaker 1: concerns about deep Seak's ability to protect its data, suffering

107
00:06:03,134 --> 00:06:05,614
Speaker 1: a serious breach at the end of January that saw

108
00:06:05,654 --> 00:06:10,694
Speaker 1: more than one million sensitive records exposed. Though is different

109
00:06:10,734 --> 00:06:14,414
Speaker 1: from other AI tools like Chat, GBT Perplexity in Gemini,

110
00:06:14,774 --> 00:06:17,974
Speaker 1: its open source and uses fewer resources than its rivals.

111
00:06:18,334 --> 00:06:21,134
Speaker 1: It was also made much cheaper than the competition, putting

112
00:06:21,134 --> 00:06:23,294
Speaker 1: the industry into a spin when it was revealed that

113
00:06:23,374 --> 00:06:25,774
Speaker 1: you can in fact power AI without the need for

114
00:06:25,894 --> 00:06:31,334
Speaker 1: overly expensive hardware. The Chinese government has strongly criticized Australia's decision,

115
00:06:31,574 --> 00:06:36,134
Speaker 1: characterizing it as the politicization of economic trade and technological issues,

116
00:06:36,414 --> 00:06:39,414
Speaker 1: and Beijing maintains that it has never and will never

117
00:06:39,814 --> 00:06:44,254
Speaker 1: require companies to illegally collect or store data. Chinese State

118
00:06:44,294 --> 00:06:48,694
Speaker 1: media has also accused Australia of ideological discrimination, suggesting that

119
00:06:48,734 --> 00:06:52,494
Speaker 1: if security concerns were genuine, similar restrictions would apply to

120
00:06:52,614 --> 00:06:56,854
Speaker 1: US based AI companies too. In an interesting twist, it's

121
00:06:56,854 --> 00:06:59,694
Speaker 1: emerged that one of the pivotal contributors to creating Deep

122
00:06:59,694 --> 00:07:03,134
Speaker 1: Seek studied computer science right here in Australia, and this

123
00:07:03,134 --> 00:07:06,494
Speaker 1: connection highlights the global nature of AI development and the

124
00:07:06,574 --> 00:07:12,734
Speaker 1: complex relationships between technology, education, national security. The Deep Sea controversy,

125
00:07:12,734 --> 00:07:16,694
Speaker 1: though racist questions about the trustworthiness of all AI platforms

126
00:07:16,694 --> 00:07:20,254
Speaker 1: in general. Well popular tools like chat, GBT, perplexity in

127
00:07:20,294 --> 00:07:23,574
Speaker 1: GEM and I continue to operate freely. Experts warn about

128
00:07:23,574 --> 00:07:28,174
Speaker 1: sharing personal information with any AI system. Doctor Nickiswaney is

129
00:07:28,174 --> 00:07:31,294
Speaker 1: the founder and director of AI her Way Nikki. There

130
00:07:31,294 --> 00:07:34,134
Speaker 1: are so many tools we can access, now, who do

131
00:07:34,214 --> 00:07:37,614
Speaker 1: we trust? Are there any that are more trustworthy than

132
00:07:37,614 --> 00:07:37,854
Speaker 1: the other?

133
00:07:39,134 --> 00:07:43,294
Speaker 2: Oh? I don't know whether to answer this personally or professionally. Yeah,

134
00:07:43,374 --> 00:07:46,094
Speaker 2: but I think when we got to analyze what products

135
00:07:46,134 --> 00:07:48,374
Speaker 2: we're using it and why. I guess the first thing

136
00:07:48,374 --> 00:07:52,334
Speaker 2: to remind people is that all of these are commodified products, right,

137
00:07:52,374 --> 00:07:54,454
Speaker 2: so I often liken it too. When you go into

138
00:07:54,494 --> 00:07:57,254
Speaker 2: a casino, they put right lights everywhere, there's lots of

139
00:07:57,254 --> 00:08:00,774
Speaker 2: flashing colors, there's lots of distractions in entertainment. There's no windows,

140
00:08:00,814 --> 00:08:02,374
Speaker 2: so you can't tell how long you've been in there,

141
00:08:02,654 --> 00:08:05,174
Speaker 2: and the whole point is to kind of trap you

142
00:08:05,214 --> 00:08:06,774
Speaker 2: in there for as long as possible, because the longer

143
00:08:06,774 --> 00:08:08,494
Speaker 2: that you're in there, the more money you're likely to

144
00:08:08,494 --> 00:08:11,374
Speaker 2: spend and hand over to business, and that's good for business.

145
00:08:11,694 --> 00:08:15,054
Speaker 2: So when we use AI models, as yet, we don't

146
00:08:15,134 --> 00:08:21,454
Speaker 2: have truly free, truly open source, sovereign options. So what's

147
00:08:21,454 --> 00:08:24,414
Speaker 2: important to remember is any model that we're using, it's

148
00:08:24,414 --> 00:08:26,814
Speaker 2: a product, and they'd like us to use their product

149
00:08:27,054 --> 00:08:29,094
Speaker 2: for as long as possible because we either paying with

150
00:08:29,134 --> 00:08:31,934
Speaker 2: money or we're paying with data, and both those things

151
00:08:32,014 --> 00:08:34,774
Speaker 2: are a real gold mine to any business that controls

152
00:08:34,774 --> 00:08:40,054
Speaker 2: these AI models. That being said, we do have global

153
00:08:40,054 --> 00:08:43,974
Speaker 2: security concerns over something like deep Seek, partially because of

154
00:08:43,974 --> 00:08:46,734
Speaker 2: who owns it, So we obviously have some sort of

155
00:08:46,814 --> 00:08:49,574
Speaker 2: global tensions with security in China, we saw that through

156
00:08:49,574 --> 00:08:52,454
Speaker 2: the kind of TikTok debate, and Deepseek is owned by

157
00:08:52,454 --> 00:08:56,094
Speaker 2: the people who make TikTok as well, then we also

158
00:08:56,294 --> 00:08:58,734
Speaker 2: have the sheer design of deep Seak. So deep Seek

159
00:08:58,854 --> 00:09:02,054
Speaker 2: is made on slightly lower quality chips and with a

160
00:09:02,054 --> 00:09:04,894
Speaker 2: little bit less reasoning power, and just through the virtue

161
00:09:04,894 --> 00:09:07,134
Speaker 2: of how it's been designed, it is a little bit

162
00:09:07,174 --> 00:09:10,574
Speaker 2: more prone to data leakage. People should have awareness of

163
00:09:10,654 --> 00:09:13,574
Speaker 2: which model they're using. You should be using models where

164
00:09:13,614 --> 00:09:15,734
Speaker 2: you control whether your data is being shared with the

165
00:09:15,734 --> 00:09:19,054
Speaker 2: people who own that AI model. And then beyond that,

166
00:09:19,614 --> 00:09:22,614
Speaker 2: the US owned models and UK owned models probably are

167
00:09:22,694 --> 00:09:26,014
Speaker 2: a bit safer. But I don't want to make people

168
00:09:26,054 --> 00:09:30,094
Speaker 2: feel like it's downright safe, because, like I said, they

169
00:09:30,174 --> 00:09:32,774
Speaker 2: want your data and that's something that you've got to

170
00:09:32,774 --> 00:09:34,374
Speaker 2: think about, regardless of who you're using.

171
00:09:34,734 --> 00:09:36,534
Speaker 1: You kind of brought up there. We are concerned about

172
00:09:36,574 --> 00:09:38,694
Speaker 1: deep Seek for several reasons, but one of those is

173
00:09:38,734 --> 00:09:41,134
Speaker 1: the fact that it is owned by a Chinese company

174
00:09:41,254 --> 00:09:44,134
Speaker 1: who are directly linked to the Chinese government. Why are

175
00:09:44,174 --> 00:09:48,534
Speaker 1: we scared of China having our data but not scared

176
00:09:48,574 --> 00:09:53,134
Speaker 1: of oligarchs like Mark Zuckerberg having our data? For example?

177
00:09:53,174 --> 00:09:56,454
Speaker 1: Like what's the difference between Mark Zuckerberg having it or

178
00:09:56,494 --> 00:09:57,654
Speaker 1: the Chinese government having it.

179
00:09:58,134 --> 00:10:00,734
Speaker 2: Yeah, there's a little bit of social conditioning, I think,

180
00:10:00,774 --> 00:10:03,734
Speaker 2: so we just inherently have a little bit more kind

181
00:10:03,734 --> 00:10:07,934
Speaker 2: of allegiance with states like the US and UK. Whether

182
00:10:08,014 --> 00:10:11,014
Speaker 2: or not that's completely correctly placed, I think is more

183
00:10:11,054 --> 00:10:13,134
Speaker 2: of a kind of individual choice. There's just a bit

184
00:10:13,134 --> 00:10:16,214
Speaker 2: more kind of mystery behind what happens to data inside

185
00:10:16,294 --> 00:10:19,694
Speaker 2: of China. There's a little bit more ambiguity between how

186
00:10:19,734 --> 00:10:21,854
Speaker 2: much control the government have, how much access they have,

187
00:10:21,934 --> 00:10:24,134
Speaker 2: and also what they choose to do with that, and

188
00:10:24,174 --> 00:10:27,454
Speaker 2: because they're not as aligned on you know, we've seen

189
00:10:27,494 --> 00:10:31,214
Speaker 2: issues with human rights on a global scale, there's equity

190
00:10:31,214 --> 00:10:34,134
Speaker 2: issues and security concerns. So it's also just because it's

191
00:10:34,134 --> 00:10:35,974
Speaker 2: a little bit more of a black box, we have

192
00:10:36,014 --> 00:10:38,774
Speaker 2: a little bit more mystery around how that's data is used,

193
00:10:38,814 --> 00:10:41,654
Speaker 2: and that comes into the kind of security concerns over

194
00:10:42,174 --> 00:10:46,014
Speaker 2: these US models. That being said, the transparency over what

195
00:10:46,214 --> 00:10:50,094
Speaker 2: the US does with data, I think, gosh, the inaugration

196
00:10:50,454 --> 00:10:53,974
Speaker 2: really kind of shine a spotlight on how close these

197
00:10:54,014 --> 00:10:56,494
Speaker 2: tech companies are to the government now in the US,

198
00:10:56,854 --> 00:10:59,774
Speaker 2: and I think that's also worthwhile watching over the next

199
00:10:59,854 --> 00:11:01,854
Speaker 2: kind of six to twelve months of how those power

200
00:11:01,934 --> 00:11:02,814
Speaker 2: dynamics play out.

201
00:11:03,614 --> 00:11:07,174
Speaker 1: Are there any Australian based AI systems that we could

202
00:11:07,214 --> 00:11:10,174
Speaker 1: potentially look at Nikki i'm in i Canvas, for example,

203
00:11:10,174 --> 00:11:12,694
Speaker 1: it's based in Sydney and they have an AI model

204
00:11:12,774 --> 00:11:16,574
Speaker 1: for their graphic design business. Are there other Australian models

205
00:11:16,574 --> 00:11:17,814
Speaker 1: that we might be able to look.

206
00:11:17,694 --> 00:11:20,814
Speaker 2: At not that compete to the same level. So there's

207
00:11:20,814 --> 00:11:22,814
Speaker 2: a few articles and there's definitely a push right now

208
00:11:22,814 --> 00:11:26,654
Speaker 2: that Australia needs an Australian owned AI. I just came

209
00:11:26,694 --> 00:11:29,454
Speaker 2: back from a forum about AI use for developing countries.

210
00:11:29,494 --> 00:11:31,374
Speaker 2: It's the same thing world over. So we have this

211
00:11:31,454 --> 00:11:34,174
Speaker 2: huge issue right now with data sovereignty. We know that

212
00:11:34,294 --> 00:11:38,694
Speaker 2: AI is really important to efficiencies and productivity, and we're

213
00:11:38,734 --> 00:11:40,894
Speaker 2: not just talking about rewriting your email right when we're

214
00:11:40,894 --> 00:11:43,694
Speaker 2: talking about AI to global scale, we're talking about world

215
00:11:43,734 --> 00:11:47,054
Speaker 2: economies and how much power and access to influence we

216
00:11:47,134 --> 00:11:50,014
Speaker 2: all have at the moment. We do have this real

217
00:11:50,094 --> 00:11:52,454
Speaker 2: issue with the fact that there is not a lot

218
00:11:52,454 --> 00:11:55,134
Speaker 2: of autonomy over who owns these AI models and how

219
00:11:55,174 --> 00:11:58,294
Speaker 2: we can make them fit our own purpose, our own culture,

220
00:11:58,374 --> 00:12:01,134
Speaker 2: our own ethics, our own idea of what these models

221
00:12:01,174 --> 00:12:03,654
Speaker 2: should achieve. So it's definitely something we need to address.

222
00:12:03,854 --> 00:12:06,254
Speaker 2: It's definitely something that the government are talking about. It's

223
00:12:06,254 --> 00:12:09,134
Speaker 2: definitely something we've seen with the UK announcements. We have

224
00:12:09,254 --> 00:12:12,814
Speaker 2: nation now throwing money at the problem, so to speak,

225
00:12:13,094 --> 00:12:16,974
Speaker 2: to try and own these models. The thing about deep

226
00:12:17,054 --> 00:12:19,454
Speaker 2: Seek is that it showed us that you can make

227
00:12:19,494 --> 00:12:22,814
Speaker 2: AI models much cheaper and much quicker than we ever thought.

228
00:12:23,294 --> 00:12:26,574
Speaker 2: So whereas the US companies, you know, Open Ai, Google,

229
00:12:27,054 --> 00:12:29,494
Speaker 2: they're on track to spend about a trillion dollars on

230
00:12:29,574 --> 00:12:32,654
Speaker 2: AI investments and deep Seek reportedly was made for about

231
00:12:32,694 --> 00:12:35,614
Speaker 2: six million dollars. So the one thing that deep Sek

232
00:12:35,694 --> 00:12:38,414
Speaker 2: does do is it models to us that perhaps this

233
00:12:38,534 --> 00:12:41,054
Speaker 2: is much more possible than we thought. And that's something

234
00:12:41,054 --> 00:12:44,734
Speaker 2: that is exciting for places like Australia and especially developing

235
00:12:44,734 --> 00:12:47,134
Speaker 2: countries around the world because it means that they might

236
00:12:47,254 --> 00:12:49,934
Speaker 2: have a choice about actually making AI that serves them.

237
00:12:50,334 --> 00:12:53,214
Speaker 1: Can we talk about AI search engines, in particular the

238
00:12:53,254 --> 00:12:55,214
Speaker 1: ones that people are using the most to chat, GPT

239
00:12:55,334 --> 00:12:58,814
Speaker 1: and perplexity. Anna sent us an email which kind of

240
00:12:58,854 --> 00:13:01,854
Speaker 1: links into this, so she was trying to Google. She

241
00:13:01,894 --> 00:13:04,214
Speaker 1: gave us an example she was trying to google nice

242
00:13:04,494 --> 00:13:07,374
Speaker 1: navy blue blazers. She wanted to buy one. She said,

243
00:13:07,814 --> 00:13:10,414
Speaker 1: the only thing that kept coming up with the same

244
00:13:10,494 --> 00:13:13,854
Speaker 1: list of retailers, So the big guns those that were sponsored,

245
00:13:14,214 --> 00:13:17,934
Speaker 1: and there was no smaller designers. There was no independent companies.

246
00:13:17,934 --> 00:13:19,974
Speaker 1: It was all the same people over and over. And

247
00:13:20,014 --> 00:13:23,414
Speaker 1: she said, is there a better way to enable true

248
00:13:23,534 --> 00:13:27,374
Speaker 1: search of the Internet? Is this where AI comes into play?

249
00:13:27,454 --> 00:13:30,374
Speaker 1: Can that help us better search things online?

250
00:13:30,654 --> 00:13:32,574
Speaker 2: It can actually if you know how to use it.

251
00:13:32,694 --> 00:13:36,494
Speaker 2: So one thing is that companies like Google, which have Gemini,

252
00:13:36,654 --> 00:13:38,254
Speaker 2: Now when you search in Google, it comes up with

253
00:13:38,294 --> 00:13:41,454
Speaker 2: an AI summary, And we have to remember that AI,

254
00:13:41,734 --> 00:13:44,894
Speaker 2: by default, it's a big data algorithm, and algorithms are

255
00:13:44,894 --> 00:13:48,654
Speaker 2: always about looking for patterns in data. So when we

256
00:13:48,814 --> 00:13:52,174
Speaker 2: ask AI models, whether that's through Google or directly accessing

257
00:13:52,174 --> 00:13:54,014
Speaker 2: something like chat jeep team, we say hey, can I

258
00:13:54,014 --> 00:13:56,254
Speaker 2: have a blue blazer? It's going to look at everything

259
00:13:56,254 --> 00:13:59,494
Speaker 2: it has about blue blazers and give you what it

260
00:13:59,894 --> 00:14:03,014
Speaker 2: most likely anticipates that you want, which is always going

261
00:14:03,054 --> 00:14:05,694
Speaker 2: to be the run of the mill, the stuff that

262
00:14:06,094 --> 00:14:07,974
Speaker 2: is most cited on the Internet. So that's why you're

263
00:14:07,974 --> 00:14:11,254
Speaker 2: going to get your chainstores or the same few big retailers. However,

264
00:14:11,854 --> 00:14:13,774
Speaker 2: the power of you as a consumer is that you

265
00:14:13,814 --> 00:14:16,494
Speaker 2: can design the instructions that you give. So if you

266
00:14:16,574 --> 00:14:19,574
Speaker 2: instead say I'm looking for a blue blazer, I only

267
00:14:19,614 --> 00:14:24,014
Speaker 2: want you to give me boutique handmade options in Australia,

268
00:14:24,174 --> 00:14:26,174
Speaker 2: It's then going to use that extra information to go

269
00:14:26,214 --> 00:14:28,534
Speaker 2: and filter for those sources in its own data bank.

270
00:14:28,774 --> 00:14:31,854
Speaker 2: So you can absolutely control a little bit more around

271
00:14:31,894 --> 00:14:33,654
Speaker 2: what you get. But you have to understand that you're

272
00:14:33,654 --> 00:14:35,654
Speaker 2: in the driver's seat and you have to be very specific.

273
00:14:35,654 --> 00:14:38,494
Speaker 2: If you're not specific, it's always going to assume that

274
00:14:38,574 --> 00:14:40,454
Speaker 2: you just want whatever it's got the most amount of

275
00:14:40,534 --> 00:14:43,054
Speaker 2: data about. And that goes for blue blazers, it goes

276
00:14:43,094 --> 00:14:46,214
Speaker 2: for writing marketing messages, it goes for coming up with

277
00:14:46,254 --> 00:14:49,654
Speaker 2: suggestions for birthday presents, whatever you're using AI for. Unless

278
00:14:49,694 --> 00:14:51,534
Speaker 2: you're really specific and you force it to look at

279
00:14:51,534 --> 00:14:53,894
Speaker 2: the kind of fringe data sets, it's always going to

280
00:14:53,934 --> 00:14:55,014
Speaker 2: go middle ground.

281
00:14:55,534 --> 00:14:58,094
Speaker 1: Can I ask you about that AI summary that Gemini

282
00:14:58,214 --> 00:15:01,214
Speaker 1: Google gives you? Now, what's the benefit to Google to

283
00:15:01,374 --> 00:15:04,534
Speaker 1: keep you inside Google itself rather than click links that

284
00:15:04,574 --> 00:15:05,214
Speaker 1: it suggests?

285
00:15:05,854 --> 00:15:11,054
Speaker 2: Data? Right? Data is the new goal. This is a

286
00:15:11,054 --> 00:15:14,454
Speaker 2: new global currency. We care less and or less about money,

287
00:15:14,494 --> 00:15:17,134
Speaker 2: and we care more and more about data. So by

288
00:15:17,174 --> 00:15:20,534
Speaker 2: having you stay inside of Google's system every time you're

289
00:15:20,534 --> 00:15:22,974
Speaker 2: searching for something, so in those summaries, they do actually

290
00:15:23,014 --> 00:15:25,294
Speaker 2: hype a link to where they get that information from.

291
00:15:25,374 --> 00:15:27,974
Speaker 2: So if you're clicking a link within that summary, that's

292
00:15:28,014 --> 00:15:31,374
Speaker 2: providing really valuable data about how people behave, what information

293
00:15:31,414 --> 00:15:34,014
Speaker 2: they validate, what they like, what they don't like, what

294
00:15:34,094 --> 00:15:36,934
Speaker 2: they find interesting, what keeps them on there longer. It's

295
00:15:36,934 --> 00:15:39,694
Speaker 2: just all about data, and that's a really clever way

296
00:15:39,694 --> 00:15:41,534
Speaker 2: to kind of feed us something that's helpful to ask

297
00:15:41,774 --> 00:15:45,094
Speaker 2: and that gives Google slash Gemini more information about what

298
00:15:45,134 --> 00:15:45,574
Speaker 2: we want.

299
00:15:45,894 --> 00:15:48,774
Speaker 1: Just finally, Nikki, I want to get your opinion on

300
00:15:48,854 --> 00:15:52,894
Speaker 1: what might be the most trustworthy AI assistant that is

301
00:15:52,934 --> 00:15:55,254
Speaker 1: out there right now. So the big four that people

302
00:15:55,374 --> 00:15:59,374
Speaker 1: generally would know about is Chat, GBT, Claude, Gemini, and

303
00:15:59,454 --> 00:16:02,494
Speaker 1: Deep Seek. Out of those four, who do you trust

304
00:16:02,494 --> 00:16:02,934
Speaker 1: the most?

305
00:16:03,174 --> 00:16:05,534
Speaker 2: I would have to vote Claude. As a personal and

306
00:16:05,534 --> 00:16:08,214
Speaker 2: professional opinion, this is often the system that we go

307
00:16:08,334 --> 00:16:12,974
Speaker 2: with when reality and ethics matter the most. Claude by

308
00:16:13,014 --> 00:16:15,374
Speaker 2: default does not share your data, so they don't take

309
00:16:15,414 --> 00:16:17,094
Speaker 2: what you're putting in there to use it to train

310
00:16:17,174 --> 00:16:19,814
Speaker 2: their next models. And Claud's actually made by two people

311
00:16:19,854 --> 00:16:22,094
Speaker 2: that were at open ai in the early days and

312
00:16:22,174 --> 00:16:24,654
Speaker 2: left because they felt like open ai was not ethical enough,

313
00:16:24,694 --> 00:16:28,094
Speaker 2: and they started their own company, and they delayed releasing

314
00:16:28,214 --> 00:16:30,534
Speaker 2: Claude even beyond the point that was probably ready because

315
00:16:30,534 --> 00:16:33,294
Speaker 2: they wanted to do extra safety testing and they were

316
00:16:33,334 --> 00:16:36,414
Speaker 2: concerned around how it behaves, So they missed out on

317
00:16:36,574 --> 00:16:40,814
Speaker 2: potentially millions of dollars because they delayed that launch. To me,

318
00:16:41,174 --> 00:16:44,814
Speaker 2: that speaks of the highest concern around how these models

319
00:16:44,854 --> 00:16:46,694
Speaker 2: behave and what they do for us. So if I

320
00:16:46,694 --> 00:16:48,494
Speaker 2: had to pick one of them, I'd go with Claude.

321
00:16:48,574 --> 00:16:51,014
Speaker 2: And as a last tip as well, a good way

322
00:16:51,054 --> 00:16:53,214
Speaker 2: to think about it. Right, we can't avoid using AI,

323
00:16:53,294 --> 00:16:55,614
Speaker 2: but if you wouldn't post it on a Reddit forum,

324
00:16:55,854 --> 00:16:57,574
Speaker 2: consider whether you really need to be sharing it with

325
00:16:57,614 --> 00:17:00,454
Speaker 2: AI buy and large, it's going to keep you pretty safe.

326
00:17:00,774 --> 00:17:03,974
Speaker 2: So yeah, until we get clearer options and until the

327
00:17:04,134 --> 00:17:07,094
Speaker 2: market becomes a little bit more diversified, you wouldn't post

328
00:17:07,134 --> 00:17:09,814
Speaker 2: it on Reddit, don't put it into an AI model.

329
00:17:09,894 --> 00:17:13,614
Speaker 1: So the basics for using any AI platform are never

330
00:17:13,694 --> 00:17:18,334
Speaker 1: share sensitive personal information like financial details or account credentials

331
00:17:18,334 --> 00:17:21,294
Speaker 1: like passwords or usernames. Be aware that once data is

332
00:17:21,374 --> 00:17:24,534
Speaker 1: provided to AI systems, you may lose control over where

333
00:17:24,574 --> 00:17:28,094
Speaker 1: that data goes and who can access it. And AI

334
00:17:28,214 --> 00:17:32,214
Speaker 1: tools learn and improve by analyzing your data, which could

335
00:17:32,294 --> 00:17:36,094
Speaker 1: then be used for targeted advertising or potentially accessed by

336
00:17:36,174 --> 00:17:39,494
Speaker 1: unauthorized parties, So be careful how much you give it

337
00:17:39,534 --> 00:17:42,254
Speaker 1: to work with. Thanks for taking the time to feed

338
00:17:42,294 --> 00:17:44,694
Speaker 1: your mind with us today. This episode of The Quickie

339
00:17:44,734 --> 00:17:48,094
Speaker 1: was produced by me Claire Murphy and our executive producer Taylorstrano,

340
00:17:48,254 --> 00:17:49,894
Speaker 1: with audio production by Lou Hill.