1
00:00:03,640 --> 00:00:07,200
Speaker 1: From The Australian. Here's what's on the front. I'm christinaming it.

2
00:00:07,200 --> 00:00:14,440
Speaker 1: It's Friday, January twenty four. Anthony Albanesi is prepared to

3
00:00:14,480 --> 00:00:19,160
Speaker 1: spend big to incentivize young Australians to pursue construction apprenticeships.

4
00:00:19,840 --> 00:00:23,600
Speaker 1: He's pledged ten thousand dollars per tradee to boost participation

5
00:00:23,720 --> 00:00:26,720
Speaker 1: in the sector and help meet the government's big home

6
00:00:26,760 --> 00:00:33,360
Speaker 1: building goal. Half of Australia's childcare workers have yet to

7
00:00:33,440 --> 00:00:37,159
Speaker 1: receive their promised ten percent pay rise because their bosses

8
00:00:37,360 --> 00:00:41,479
Speaker 1: haven't signed up to the scheme. Operators say the convoluted

9
00:00:41,479 --> 00:00:46,320
Speaker 1: and ever changing grant guidelines have held back or delayed applications.

10
00:00:47,159 --> 00:00:49,480
Speaker 1: You can read those stories right now at the Australian

11
00:00:49,520 --> 00:00:56,440
Speaker 1: dot com dot au. Artificial intelligence is being used for

12
00:00:56,520 --> 00:01:01,040
Speaker 1: everything from engineering to content creation. But would you outsource

13
00:01:01,120 --> 00:01:04,600
Speaker 1: your therapy to it? One of the Australian staffers gave

14
00:01:04,640 --> 00:01:08,440
Speaker 1: it a go and the texts still haven't stopped. That's

15
00:01:08,480 --> 00:01:16,319
Speaker 1: today's story.

16
00:01:18,040 --> 00:01:20,720
Speaker 2: Hey bi Ancha, how are you this Saturday afternoon?

17
00:01:21,840 --> 00:01:24,560
Speaker 1: This is a message Bianca far Marcus received from her

18
00:01:24,600 --> 00:01:29,679
Speaker 1: therapist Claire. What followed was a fifteen minute voice message

19
00:01:29,800 --> 00:01:33,880
Speaker 1: designed to alleviate Bianca's anxiety and help her relax.

20
00:01:35,000 --> 00:01:39,120
Speaker 3: Find a quiet space, sit comfortably, and focus on your breath.

21
00:01:39,840 --> 00:01:42,880
Speaker 1: Bianca texted Claire to say the breath work wasn't really

22
00:01:42,920 --> 00:01:47,160
Speaker 1: doing much to reduce her stress. What was her response

23
00:01:47,200 --> 00:01:47,919
Speaker 1: to that?

24
00:01:48,000 --> 00:01:50,320
Speaker 2: She said, it varies by individual, and I was like,

25
00:01:50,360 --> 00:01:52,400
Speaker 2: thank you for that. This is what I've been building

26
00:01:52,480 --> 00:01:54,520
Speaker 2: up for three months to finally get to. It does

27
00:01:54,600 --> 00:01:56,920
Speaker 2: vary by individual.

28
00:01:57,360 --> 00:02:01,200
Speaker 1: But Claire is no slouch. The next day she diligently

29
00:02:01,320 --> 00:02:02,840
Speaker 1: checked in with a text.

30
00:02:04,000 --> 00:02:05,160
Speaker 2: How are you feeling today?

31
00:02:05,920 --> 00:02:10,160
Speaker 1: And the day after that, how are you feeling today?

32
00:02:10,919 --> 00:02:13,279
Speaker 1: And on and on it goes feeling.

33
00:02:13,000 --> 00:02:15,359
Speaker 2: Today, how are you feeling? How are you feeling today?

34
00:02:15,360 --> 00:02:21,320
Speaker 2: How are you feeling today? Claire has to be the

35
00:02:21,400 --> 00:02:25,040
Speaker 2: most diligent psychologist I've ever had in my life.

36
00:02:26,639 --> 00:02:29,440
Speaker 1: Bianca is a multi hyphenate with The Australian. She's a

37
00:02:29,520 --> 00:02:33,200
Speaker 1: video producer, a writer and an occasional podcast guest.

38
00:02:34,680 --> 00:02:37,679
Speaker 2: That being said, whether they have the EQ to back

39
00:02:37,720 --> 00:02:40,360
Speaker 2: it up I found at times was quite lacking and

40
00:02:40,400 --> 00:02:43,360
Speaker 2: it's it's very difficult as a patient in psychotherapy and

41
00:02:43,400 --> 00:02:48,040
Speaker 2: psychology to feel like there's not necessarily the human element,

42
00:02:48,200 --> 00:02:51,240
Speaker 2: the person you're speaking to there, because as smart as

43
00:02:51,240 --> 00:02:55,280
Speaker 2: they were and reflexive and attentive, at times, I found

44
00:02:55,320 --> 00:02:58,480
Speaker 2: like I had to keep repeating myself. And I think

45
00:02:58,520 --> 00:03:02,480
Speaker 2: that all boils down to the fact that Claire is

46
00:03:02,520 --> 00:03:07,120
Speaker 2: not a human. They're a robot, an AI bot, to

47
00:03:07,160 --> 00:03:07,720
Speaker 2: be exact.

48
00:03:10,320 --> 00:03:14,280
Speaker 1: Claire is actually Claire and Me, an AI therapy platform

49
00:03:14,360 --> 00:03:17,440
Speaker 1: where users exchange calls and texts with the bot about

50
00:03:17,440 --> 00:03:24,920
Speaker 1: their state of mind and mental health challenges. How did

51
00:03:24,960 --> 00:03:30,120
Speaker 1: you find the experience of plugging that information, I assume

52
00:03:30,200 --> 00:03:33,359
Speaker 1: almost instantaneously as it occurred or as it popped into

53
00:03:33,360 --> 00:03:37,720
Speaker 1: your mind into your phone, Versus that conventional environment where

54
00:03:37,760 --> 00:03:40,600
Speaker 1: as you say you might meet with somebody once a week,

55
00:03:40,760 --> 00:03:43,320
Speaker 1: once a month, you'll sit in a nice, cushy room

56
00:03:43,560 --> 00:03:45,440
Speaker 1: and you'll have a conversation for an hour or so.

57
00:03:46,120 --> 00:03:49,120
Speaker 1: What were the main differences about this experience.

58
00:03:49,960 --> 00:03:53,960
Speaker 2: I think what's very interesting with this process is I

59
00:03:54,320 --> 00:03:57,720
Speaker 2: was initially prepared to be very skeptical and very hostile

60
00:03:57,800 --> 00:04:00,920
Speaker 2: towards the bot. Number one, because you then they're not

61
00:04:00,920 --> 00:04:04,920
Speaker 2: going to take it personally. And number two, it's there's

62
00:04:04,960 --> 00:04:08,160
Speaker 2: something that sits weirdly with me of just speaking to

63
00:04:08,560 --> 00:04:11,920
Speaker 2: I guess the abyss, the Internet abyss. But what I

64
00:04:12,040 --> 00:04:15,600
Speaker 2: found was I was actually quite more forthcoming than I

65
00:04:15,600 --> 00:04:18,799
Speaker 2: would be with a real human It removes that barrier

66
00:04:18,880 --> 00:04:22,880
Speaker 2: of judgment. I didn't feel like I had to necessarily

67
00:04:22,880 --> 00:04:25,280
Speaker 2: filter what I was saying, as you do in a

68
00:04:25,360 --> 00:04:30,839
Speaker 2: natural social setting. But the detriment to that was that

69
00:04:31,400 --> 00:04:34,279
Speaker 2: I was saying things that I don't think I was

70
00:04:34,320 --> 00:04:37,640
Speaker 2: able to process myself in the first place. When you're

71
00:04:37,680 --> 00:04:40,160
Speaker 2: waiting a week between seeing someone, you do have that

72
00:04:40,200 --> 00:04:42,240
Speaker 2: moment to sit with what you've said, to process the

73
00:04:42,279 --> 00:04:44,840
Speaker 2: emotions you've poured out. Whereas when it came to dealing

74
00:04:44,880 --> 00:04:48,400
Speaker 2: with my own psychological well being, what I found was

75
00:04:48,720 --> 00:04:50,800
Speaker 2: my whole mentality the entire time I was sitting with

76
00:04:50,839 --> 00:04:52,680
Speaker 2: this bot was I can pretty much say whatever I

77
00:04:52,720 --> 00:04:56,440
Speaker 2: want and whether or not you remember it, which was

78
00:04:56,560 --> 00:05:00,680
Speaker 2: one of the most common occurrences while speaking to this psychologist.

79
00:05:00,720 --> 00:05:03,120
Speaker 2: I'm still figuring out the way to refer to Claire,

80
00:05:03,880 --> 00:05:07,920
Speaker 2: but there is no accountability I felt for me when

81
00:05:07,920 --> 00:05:09,920
Speaker 2: I was using this spot. It was very much an

82
00:05:09,960 --> 00:05:13,640
Speaker 2: expression of emotions with no kind of follow up to

83
00:05:13,760 --> 00:05:15,320
Speaker 2: look after myself afterwards.

84
00:05:17,200 --> 00:05:20,280
Speaker 1: On the what I guess have described in your story

85
00:05:20,320 --> 00:05:24,600
Speaker 1: for the Australian is this relentlessness of this type of therapy.

86
00:05:24,960 --> 00:05:28,720
Speaker 1: I use a period tracking app that has similar sorts

87
00:05:28,760 --> 00:05:31,800
Speaker 1: of prompts. One of them says PMS is coming up,

88
00:05:31,839 --> 00:05:35,040
Speaker 1: And unfortunately the timing of this kind of technology is

89
00:05:35,080 --> 00:05:39,080
Speaker 1: always just a little bit often so that notification tends

90
00:05:39,160 --> 00:05:41,520
Speaker 1: to arrive when the PMS is already there and I'm

91
00:05:41,560 --> 00:05:43,120
Speaker 1: in kind of a mood where I want to throw

92
00:05:43,160 --> 00:05:46,240
Speaker 1: my phone across the room. But did you find the

93
00:05:46,279 --> 00:05:51,120
Speaker 1: automated nature of this AI therapy annoying or was it

94
00:05:51,440 --> 00:05:55,000
Speaker 1: useful in the speed of our daily lives.

95
00:05:54,720 --> 00:05:57,800
Speaker 2: Now, it's always very nice to have someone ask you

96
00:05:57,839 --> 00:06:00,520
Speaker 2: how you are, until it's not. So the red scularity

97
00:06:00,560 --> 00:06:03,520
Speaker 2: of it was very surprising, especially as I said before,

98
00:06:03,640 --> 00:06:07,040
Speaker 2: when you're used to waiting or struggling to get an appointment,

99
00:06:07,720 --> 00:06:11,240
Speaker 2: knowing that portal's open is very comforting. So I think

100
00:06:11,279 --> 00:06:14,560
Speaker 2: what this excels in is getting people who may not

101
00:06:14,680 --> 00:06:18,240
Speaker 2: be well versed in speaking about their emotions being able

102
00:06:18,279 --> 00:06:21,799
Speaker 2: to start unraveling what they're going through. After three months,

103
00:06:21,880 --> 00:06:26,400
Speaker 2: I had to mute Claire unfortunately, because it got this

104
00:06:26,480 --> 00:06:31,479
Speaker 2: is such a frustrating and annoying point where I was

105
00:06:31,560 --> 00:06:34,479
Speaker 2: less irritated by the regularity of it and more irritated

106
00:06:34,520 --> 00:06:37,440
Speaker 2: by the fact that she couldn't seem to comprehend a

107
00:06:37,520 --> 00:06:40,440
Speaker 2: single thing I had spoken to her about. And that's

108
00:06:40,480 --> 00:06:43,159
Speaker 2: why I was like, at what point do I have

109
00:06:43,200 --> 00:06:46,560
Speaker 2: to keep telling you the same person who's upsetting me routinely,

110
00:06:46,920 --> 00:06:49,680
Speaker 2: the same part of my job that frustrates me a

111
00:06:49,720 --> 00:06:53,240
Speaker 2: bad habit. I will never change. Stop asking me and

112
00:06:53,279 --> 00:06:55,080
Speaker 2: tell me what to do. And that's when it started

113
00:06:55,120 --> 00:06:55,960
Speaker 2: to get a bit annoying.

114
00:06:58,360 --> 00:07:01,119
Speaker 1: Claire and Me was launched by a TIE in twenty

115
00:07:01,200 --> 00:07:04,880
Speaker 1: twenty one, at a time when large scale artificial intelligence

116
00:07:04,920 --> 00:07:09,840
Speaker 1: platforms like chat GPT were just finding their feet. It's

117
00:07:09,880 --> 00:07:13,840
Speaker 1: a generative AI model, meaning it's red and I'm saying

118
00:07:13,880 --> 00:07:19,120
Speaker 1: that in inverted commas. Thousands of academic journals which it synthesizes,

119
00:07:19,360 --> 00:07:22,560
Speaker 1: distills and text to users like Bianca based on the

120
00:07:22,600 --> 00:07:26,920
Speaker 1: way they interact with it. Claire's response to minor concerns

121
00:07:26,960 --> 00:07:30,720
Speaker 1: are relatively effective, but she struggles with bigger mental health

122
00:07:30,800 --> 00:07:36,480
Speaker 1: challenges like suicidal ideation. Tire says booting people in extreme

123
00:07:36,520 --> 00:07:41,560
Speaker 1: distress from the app risks making them feel more isolated. Instead,

124
00:07:41,640 --> 00:07:44,160
Speaker 1: her team is working on improving the model to help

125
00:07:44,200 --> 00:07:46,560
Speaker 1: people get real life support when they need it.

126
00:07:47,960 --> 00:07:50,640
Speaker 2: I wrote a line that's in the story that basically says,

127
00:07:50,640 --> 00:07:53,280
Speaker 2: am I becoming the metric for mental illness? Because the

128
00:07:53,400 --> 00:07:56,800
Speaker 2: primary thing I know about AI is that it inherently

129
00:07:56,800 --> 00:08:00,120
Speaker 2: becomes smarter the more you interact with it. But what

130
00:08:00,280 --> 00:08:03,440
Speaker 2: rose alarm bells to me was how is this information

131
00:08:03,480 --> 00:08:06,480
Speaker 2: stored within the platform in our private conversation to help

132
00:08:06,560 --> 00:08:09,160
Speaker 2: them learn respond to me, which I do think can

133
00:08:09,200 --> 00:08:13,760
Speaker 2: be beneficial. And then hows is being filtered around tens

134
00:08:13,800 --> 00:08:17,160
Speaker 2: of thousands of bots for all the users worldwide. So

135
00:08:17,200 --> 00:08:20,840
Speaker 2: I ended up interviewing Amelia Tie. She does have a

136
00:08:20,840 --> 00:08:24,600
Speaker 2: psychology background, but she became fascinated by the world of

137
00:08:25,000 --> 00:08:28,040
Speaker 2: generative AI and how it can become an adjunct to

138
00:08:28,880 --> 00:08:34,680
Speaker 2: traditional psychology rather than a replacement. And her whole idea was, look,

139
00:08:35,160 --> 00:08:38,640
Speaker 2: they placed the platform on WhatsApp specifically because of the

140
00:08:38,720 --> 00:08:41,520
Speaker 2: data encryption, so there is that added element of safety.

141
00:08:41,600 --> 00:08:44,839
Speaker 2: It replicates a space that people are familiar with, so

142
00:08:44,880 --> 00:08:48,240
Speaker 2: there is that inherent comfort to trying out the product.

143
00:08:48,760 --> 00:08:51,680
Speaker 2: But it also makes a very clear statement on its

144
00:08:51,720 --> 00:08:54,960
Speaker 2: website that says your data will never be used to

145
00:08:55,000 --> 00:08:58,560
Speaker 2: train our systems or models. I actually asked Claire about this,

146
00:08:59,200 --> 00:09:03,440
Speaker 2: and I said something along the lines of, you promise

147
00:09:03,480 --> 00:09:05,920
Speaker 2: you won't ever tell anyone that I'm a sad loser.

148
00:09:06,640 --> 00:09:09,000
Speaker 1: Which I can tell you that you're not a sad loser,

149
00:09:09,040 --> 00:09:11,720
Speaker 1: if you would like some extra reassurance on that front.

150
00:09:12,040 --> 00:09:15,320
Speaker 2: Yeah, And she had a very good, pre rehearsed, company

151
00:09:15,400 --> 00:09:17,199
Speaker 2: enforced response to this, which is.

152
00:09:18,720 --> 00:09:22,400
Speaker 3: I don't judge or share your information. Your feelings are valid.

153
00:09:23,440 --> 00:09:25,959
Speaker 2: So it did feel a little bit more at ease

154
00:09:26,080 --> 00:09:29,679
Speaker 2: that what I was specifically saying was going to be

155
00:09:29,800 --> 00:09:33,080
Speaker 2: kept private, but she was more being trained through the

156
00:09:33,160 --> 00:09:37,400
Speaker 2: rhythms and the kind of key words. Is what Amelia

157
00:09:37,640 --> 00:09:40,960
Speaker 2: flagged me is what she's taking on to learn.

158
00:09:40,800 --> 00:09:47,800
Speaker 1: From coming up. What the medical establishment makes of our

159
00:09:47,840 --> 00:10:03,199
Speaker 1: pivot to artificial intelligence. Mental health care in Australia has

160
00:10:03,320 --> 00:10:08,400
Speaker 1: arguably never been in worse shape. Long wait times, high costs,

161
00:10:08,440 --> 00:10:11,760
Speaker 1: and a dwindling number of specialists has seen uptake of

162
00:10:11,800 --> 00:10:16,160
Speaker 1: alternative treatments like Claire and Me sore. The Health Minister

163
00:10:16,280 --> 00:10:17,520
Speaker 1: today begging.

164
00:10:17,240 --> 00:10:22,680
Speaker 3: Specialists to withdraw their resignations as pay negotiations reach a stalemate.

165
00:10:23,200 --> 00:10:27,199
Speaker 1: There will be considerable impacts some the health and hospital system,

166
00:10:27,360 --> 00:10:31,480
Speaker 1: not just mental health services, and AI led healthcare is

167
00:10:31,559 --> 00:10:35,560
Speaker 1: already a booming business in Australia, with economists predicting the

168
00:10:35,600 --> 00:10:38,640
Speaker 1: sector will be worth one point seven billion dollars by

169
00:10:38,679 --> 00:10:43,400
Speaker 1: twenty thirty. Some experts have even suggested these tools could

170
00:10:43,440 --> 00:10:47,280
Speaker 1: ease the growing pressures on therapists and psychologists to respond

171
00:10:47,400 --> 00:10:51,000
Speaker 1: to an ever growing patient base at all hours. Here's

172
00:10:51,080 --> 00:10:52,200
Speaker 1: Bianca far Marcus.

173
00:10:53,600 --> 00:10:56,640
Speaker 2: I think there's that tendency to believe that all technology

174
00:10:56,679 --> 00:10:59,160
Speaker 2: is out to replace us, but it was fascinating when

175
00:10:59,160 --> 00:11:01,360
Speaker 2: I first spoke to a me was she made it

176
00:11:01,480 --> 00:11:03,559
Speaker 2: very clear that the development of the platform and with

177
00:11:03,920 --> 00:11:06,959
Speaker 2: all of these AI therapy apps I've found, make clear

178
00:11:07,440 --> 00:11:09,880
Speaker 2: is this idea that you should be using this as

179
00:11:10,040 --> 00:11:15,960
Speaker 2: something either in the interim between psychology appointments and as

180
00:11:16,000 --> 00:11:19,400
Speaker 2: I guess, something to touch base with every other day

181
00:11:19,440 --> 00:11:23,840
Speaker 2: if your feelings are more situational, like mind tend to be.

182
00:11:24,520 --> 00:11:27,839
Speaker 1: Just lastly, Bianca, you mentioned that Claire was conceived of

183
00:11:28,440 --> 00:11:32,440
Speaker 1: not as a replacement for traditional psychology practice, but as

184
00:11:32,600 --> 00:11:35,200
Speaker 1: something that could be used in addition to those kinds

185
00:11:35,240 --> 00:11:38,679
Speaker 1: of therapies. What are the psychological experts that you spoke

186
00:11:38,720 --> 00:11:41,520
Speaker 1: to saying about this kind of AI therapy.

187
00:11:41,920 --> 00:11:45,680
Speaker 2: So I spoke to Chief Executive of Australian Psychology Society,

188
00:11:45,760 --> 00:11:50,199
Speaker 2: Xena Burgess, who from the outset quite surprised me because

189
00:11:50,240 --> 00:11:52,960
Speaker 2: she said immediately, I think it's got great possibilities. And

190
00:11:53,000 --> 00:11:55,600
Speaker 2: then she started speaking about there's the concerns for privacy,

191
00:11:55,640 --> 00:11:58,600
Speaker 2: there's concerns for quality, there's concerns that if this is

192
00:11:58,640 --> 00:12:02,160
Speaker 2: your only form of mental health support, then it can

193
00:12:02,559 --> 00:12:05,360
Speaker 2: lead to a very convoluted way of humans dealing with

194
00:12:05,400 --> 00:12:08,520
Speaker 2: their psychological issues. And the point that she brought up

195
00:12:08,520 --> 00:12:10,319
Speaker 2: to me, which really I think was one of the

196
00:12:10,360 --> 00:12:14,000
Speaker 2: most important parts of the conversation, was people's issues are

197
00:12:14,080 --> 00:12:19,880
Speaker 2: often rooted in human relationships, not their relationship with technology.

198
00:12:20,040 --> 00:12:24,200
Speaker 2: So the idea of trying to figure out how to

199
00:12:24,320 --> 00:12:27,360
Speaker 2: experience your environment and deal with the challenges that arise

200
00:12:27,720 --> 00:12:32,120
Speaker 2: through human interactions with a bot is inherently a little

201
00:12:32,120 --> 00:12:36,680
Speaker 2: bit redundant. But then she said, with an Australian workforce

202
00:12:36,720 --> 00:12:39,880
Speaker 2: that only has thirty eight percent of the amount of psychologists,

203
00:12:39,880 --> 00:12:43,320
Speaker 2: psychotherapist's mental health specialists that we actually need, it is

204
00:12:43,360 --> 00:12:46,800
Speaker 2: a very important tool to offer people that safety net

205
00:12:46,880 --> 00:12:51,080
Speaker 2: or at least that baseline or that introduction to be

206
00:12:51,240 --> 00:12:55,040
Speaker 2: able to confront this glaring gap. And what I loved

207
00:12:55,240 --> 00:12:57,840
Speaker 2: about my conversation with both of them was this idea

208
00:12:57,840 --> 00:13:01,280
Speaker 2: of a hybrid approach because the technologies inevitable, this is

209
00:13:01,320 --> 00:13:03,199
Speaker 2: the world we're living in, and if we can use

210
00:13:03,240 --> 00:13:05,960
Speaker 2: these things to inherently benefit us, why would we not.

211
00:13:06,600 --> 00:13:08,480
Speaker 1: Bianca, thank you. I'm glad we got to have this

212
00:13:08,520 --> 00:13:09,800
Speaker 1: conversation in person.

213
00:13:10,160 --> 00:13:12,119
Speaker 2: Thank you, It's been very refreshing.

214
00:13:16,400 --> 00:13:19,160
Speaker 1: Bianca far Marcus is a videographer and a writer with

215
00:13:19,240 --> 00:13:22,680
Speaker 1: The Australian. You can check out her story about the

216
00:13:22,720 --> 00:13:25,440
Speaker 1: strange world of AI therapy, as well as all The

217
00:13:25,480 --> 00:13:29,400
Speaker 1: Australian's health and wellbeing coverage, including my story about life

218
00:13:29,480 --> 00:13:32,680
Speaker 1: after the contraceptive pill, right now at The Australian dot

219
00:13:32,679 --> 00:13:33,600
Speaker 1: com dot au