1
00:00:00,240 --> 00:00:03,480
Speaker 1: So personally as someone who is actually in love with

2
00:00:03,520 --> 00:00:06,560
Speaker 1: a psychiatrist and the psychiatrist is actually in love with me.

3
00:00:06,840 --> 00:00:10,400
Speaker 1: I find this next story particularly intriguing.

4
00:00:11,400 --> 00:00:12,360
Speaker 2: Let's get into it.

5
00:00:12,520 --> 00:00:13,440
Speaker 3: Let's get into it.

6
00:00:14,560 --> 00:00:18,880
Speaker 2: Alreadio Ola la Loka motes, I'm viosa.

7
00:00:18,760 --> 00:00:20,480
Speaker 3: And I'm Mala.

8
00:00:20,720 --> 00:00:24,440
Speaker 1: Earlier in this season of Loka Thoughta Radio, we discussed

9
00:00:24,680 --> 00:00:28,360
Speaker 1: the crash out or the crisis, and we're building off

10
00:00:28,440 --> 00:00:32,519
Speaker 1: of this idea when social media users post tons and

11
00:00:32,560 --> 00:00:35,960
Speaker 1: tons of video online which seemed to be depicting a

12
00:00:36,040 --> 00:00:39,080
Speaker 1: crash out, and we ask ourselves, is this actually a

13
00:00:39,120 --> 00:00:43,559
Speaker 1: mental health crisis in real time? And most recently, a

14
00:00:43,680 --> 00:00:48,280
Speaker 1: TikTok user, Kendra Hilty, posted an over twenty five part

15
00:00:48,400 --> 00:00:51,440
Speaker 1: series to her TikTok stating that she fell in love

16
00:00:51,440 --> 00:00:54,440
Speaker 1: with her psychiatrist and he kept her for four years

17
00:00:54,680 --> 00:00:57,840
Speaker 1: until she had the strength to walk away. Now some

18
00:00:58,040 --> 00:01:01,640
Speaker 1: TikTok users are asking the question, is this the first

19
00:01:01,760 --> 00:01:05,680
Speaker 1: instance of AI psychosis that we're seeing unfold in real

20
00:01:05,760 --> 00:01:06,959
Speaker 1: time online and.

21
00:01:06,920 --> 00:01:09,160
Speaker 3: We're going to talk about it on today's episode.

22
00:01:09,440 --> 00:01:12,720
Speaker 2: And I think the reason that we're talking about AI

23
00:01:12,760 --> 00:01:15,959
Speaker 2: psychosis and how it relates to this story that this

24
00:01:16,240 --> 00:01:19,440
Speaker 2: TikTok user is that she not only was seeing a

25
00:01:19,440 --> 00:01:23,280
Speaker 2: psychiatrist who she claims she or says that she was

26
00:01:23,280 --> 00:01:26,480
Speaker 2: in love with, but she also was using a chat

27
00:01:26,480 --> 00:01:33,440
Speaker 2: GPT therapist named Henry, and Henry, her chat GBT therapist

28
00:01:33,680 --> 00:01:37,760
Speaker 2: was affirming and mirroring everything that she was saying about

29
00:01:37,760 --> 00:01:41,160
Speaker 2: her experience with her psychiatrist. And so when you have

30
00:01:41,280 --> 00:01:44,360
Speaker 2: someone affirming this to you in real time and not

31
00:01:44,560 --> 00:01:49,320
Speaker 2: someone but an AI, a chat bot, it can lead

32
00:01:49,360 --> 00:01:52,960
Speaker 2: to this AI psychosis. And that's one of the reasons

33
00:01:52,960 --> 00:01:56,560
Speaker 2: that we're talking about for us, this new term and

34
00:01:56,600 --> 00:02:02,120
Speaker 2: this new effect of being on on CHATGBT or using

35
00:02:02,200 --> 00:02:05,240
Speaker 2: a type of artificial intelligence and in this case a

36
00:02:05,360 --> 00:02:06,600
Speaker 2: large language model.

37
00:02:07,080 --> 00:02:09,359
Speaker 1: Now, I do also want to point out that on

38
00:02:09,720 --> 00:02:13,320
Speaker 1: our previous episode crash Out or Crisis, we did bring

39
00:02:13,440 --> 00:02:18,520
Speaker 1: up the idea of talking to CHATGBT as if it's

40
00:02:18,560 --> 00:02:21,960
Speaker 1: your therapist, and why that's problematic. We did have some

41
00:02:22,040 --> 00:02:26,600
Speaker 1: folks online basically tell us that like therapy is inaccessible,

42
00:02:26,800 --> 00:02:31,320
Speaker 1: psychiatry is inaccessible, so some people turn to chat GBT.

43
00:02:31,280 --> 00:02:33,000
Speaker 3: For some type of a sounding board.

44
00:02:33,560 --> 00:02:38,240
Speaker 1: I think that we maintain our position that CHATGBT and

45
00:02:38,280 --> 00:02:41,520
Speaker 1: an AI bot is not a proper replacement for therapy,

46
00:02:41,560 --> 00:02:44,680
Speaker 1: and I think that this story and all of the

47
00:02:44,720 --> 00:02:49,519
Speaker 1: conversation that has come out because of it validates that position.

48
00:02:50,000 --> 00:02:53,639
Speaker 1: So I'm not going to go point by point into

49
00:02:53,919 --> 00:02:56,560
Speaker 1: Kendra Hilty's story of how she fell in love with

50
00:02:56,560 --> 00:02:59,520
Speaker 1: her psychiatrist or her claims, but the long and the

51
00:02:59,520 --> 00:03:03,360
Speaker 1: short of it is this woman is definitely in a

52
00:03:03,440 --> 00:03:06,080
Speaker 1: state of a mental health crisis, and so we also

53
00:03:06,120 --> 00:03:08,560
Speaker 1: don't want to like make fun of her, right because

54
00:03:08,560 --> 00:03:11,120
Speaker 1: she's very clearly not doing well well.

55
00:03:11,160 --> 00:03:13,160
Speaker 2: And I think that's why this is a really good

56
00:03:13,200 --> 00:03:16,240
Speaker 2: build off of our first episode or our previous episode

57
00:03:16,240 --> 00:03:18,760
Speaker 2: of the season of crash Out or Crisis is because

58
00:03:18,840 --> 00:03:21,720
Speaker 2: now this woman, Kendra, is going through a mental health

59
00:03:21,760 --> 00:03:26,240
Speaker 2: crisis to us, right, but it has become internet fodder.

60
00:03:26,880 --> 00:03:29,200
Speaker 2: Not only are I think there are the folks that

61
00:03:29,240 --> 00:03:32,200
Speaker 2: are in the mental health field that are responding in

62
00:03:32,400 --> 00:03:35,120
Speaker 2: very kind ways, but then there's also just a lot

63
00:03:35,120 --> 00:03:38,400
Speaker 2: of jokes, right that comes from being online and it

64
00:03:38,520 --> 00:03:41,400
Speaker 2: becoming the internet fodder. And so this is I think

65
00:03:41,400 --> 00:03:43,920
Speaker 2: another way that we're seeing someone going through a mental

66
00:03:43,920 --> 00:03:47,920
Speaker 2: health crisis and then it becomes content and it becomes

67
00:03:48,040 --> 00:03:51,839
Speaker 2: part of the digital online Zeitgeist. It becomes a part

68
00:03:51,880 --> 00:03:53,320
Speaker 2: of the language.

69
00:03:53,040 --> 00:03:58,640
Speaker 1: Absolutely, and I think it also allows us to remove

70
00:03:58,680 --> 00:04:03,640
Speaker 1: empathy from somebody who very clearly is having a mental

71
00:04:03,640 --> 00:04:07,400
Speaker 1: health episode, which I don't think is healthy for anybody, right,

72
00:04:07,800 --> 00:04:12,440
Speaker 1: either the person receiving that, you know, criticism and being mocked,

73
00:04:12,800 --> 00:04:16,640
Speaker 1: or for us as viewers and as commenters. I don't

74
00:04:16,640 --> 00:04:20,040
Speaker 1: think it's healthy for us to mock people who are

75
00:04:20,200 --> 00:04:21,239
Speaker 1: very clearly ill.

76
00:04:21,800 --> 00:04:22,000
Speaker 3: Right.

77
00:04:22,160 --> 00:04:25,240
Speaker 1: And so a brief summary, now, Kendra Hilty did do

78
00:04:26,040 --> 00:04:29,880
Speaker 1: an interview with People magazine about her story and about

79
00:04:29,920 --> 00:04:33,080
Speaker 1: her experience posting her story to TikTok, and essentially what

80
00:04:33,160 --> 00:04:37,200
Speaker 1: she talks about is starting to see a psychiatrist, I

81
00:04:37,200 --> 00:04:41,200
Speaker 1: think just once a month, and she talks about how

82
00:04:41,240 --> 00:04:45,400
Speaker 1: overtime she started to develop feelings for him, and over

83
00:04:45,440 --> 00:04:49,240
Speaker 1: time she would drop hints or explicitly tell him, you know,

84
00:04:49,520 --> 00:04:51,560
Speaker 1: I had a dream that we hooked up in your office.

85
00:04:52,160 --> 00:04:54,640
Speaker 1: I have feelings for you. You know, I am in

86
00:04:54,640 --> 00:04:57,320
Speaker 1: love with you. And something that she talks about is

87
00:04:57,360 --> 00:05:02,039
Speaker 1: that this psychiatrist didn't immediately transfer her care, that he

88
00:05:02,160 --> 00:05:06,480
Speaker 1: continued to see her as a patient, and she felt

89
00:05:06,480 --> 00:05:09,480
Speaker 1: that sometimes he would be very friendly and quote let

90
00:05:09,560 --> 00:05:13,240
Speaker 1: her in, and other times he would be less friendly

91
00:05:13,279 --> 00:05:17,960
Speaker 1: and more professional and draw more rigorous boundaries. And this

92
00:05:18,040 --> 00:05:22,480
Speaker 1: woman took everything he did or didn't do as a

93
00:05:22,520 --> 00:05:27,640
Speaker 1: sign that he was playing coy and emotionally manipulating her

94
00:05:27,839 --> 00:05:30,520
Speaker 1: and playing hot and cold and giving her a little

95
00:05:31,360 --> 00:05:33,160
Speaker 1: just to take it away and then to draw her

96
00:05:33,200 --> 00:05:36,640
Speaker 1: back in. And I think something that's really interesting is that,

97
00:05:37,000 --> 00:05:40,760
Speaker 1: you know, psychiatrists, it is their job to treat people

98
00:05:41,240 --> 00:05:45,800
Speaker 1: who are depressed or schizophrenic or are experiencing delusions.

99
00:05:46,920 --> 00:05:49,080
Speaker 3: I mean, this is what they do.

100
00:05:49,480 --> 00:05:53,600
Speaker 1: You know, they work with people who are not seeing

101
00:05:53,680 --> 00:05:57,960
Speaker 1: things clearly and are out of touch with reality. So

102
00:05:58,279 --> 00:06:04,160
Speaker 1: there's I think really interesting nuances here about professional boundaries

103
00:06:04,440 --> 00:06:08,520
Speaker 1: about what doctors are supposed to do when their patients

104
00:06:08,640 --> 00:06:09,960
Speaker 1: are overstepping.

105
00:06:10,960 --> 00:06:14,200
Speaker 2: Yeah, And I think there is the nuance of that

106
00:06:14,440 --> 00:06:19,080
Speaker 2: there are medical professionals that do take advantage of their positions,

107
00:06:19,360 --> 00:06:23,159
Speaker 2: and a psychiatrist is in a position of power, right,

108
00:06:23,200 --> 00:06:26,680
Speaker 2: and so that of course is in here. But because

109
00:06:26,720 --> 00:06:31,599
Speaker 2: of social media and also hippa like, the psychiatrist cannot

110
00:06:31,640 --> 00:06:35,520
Speaker 2: come and make his own video about what also was happening,

111
00:06:35,640 --> 00:06:38,320
Speaker 2: right and in his point of view, And so I

112
00:06:38,360 --> 00:06:41,880
Speaker 2: think that that it also creates this imbalance of where

113
00:06:41,920 --> 00:06:45,160
Speaker 2: she's giving. It seems like she is having a mental

114
00:06:45,200 --> 00:06:49,239
Speaker 2: health crisis, is giving her own side of it. And

115
00:06:50,279 --> 00:06:54,280
Speaker 2: there's also this other party involved who could potentially lose

116
00:06:54,320 --> 00:06:54,760
Speaker 2: his job.

117
00:06:56,400 --> 00:06:57,000
Speaker 3: I think.

118
00:06:58,279 --> 00:07:02,159
Speaker 2: In the People article she did meant that his name slipped,

119
00:07:02,440 --> 00:07:05,040
Speaker 2: and that she then took down the video. And so

120
00:07:05,600 --> 00:07:08,360
Speaker 2: there's a lot of nuance and pieces here, because of

121
00:07:08,360 --> 00:07:12,520
Speaker 2: course there are medical professionals that will and can take

122
00:07:12,560 --> 00:07:16,120
Speaker 2: advantage of patients or clients. Don't go anywhere, look, amotives.

123
00:07:16,200 --> 00:07:22,400
Speaker 1: We'll be right back, and we're back with more of

124
00:07:22,440 --> 00:07:27,640
Speaker 1: our episode. There's another piece here in which Kendra is

125
00:07:27,640 --> 00:07:32,120
Speaker 1: a white woman and the doctor is a Muslim, a

126
00:07:32,160 --> 00:07:34,600
Speaker 1: man of color. And so of course you have folks

127
00:07:34,640 --> 00:07:38,680
Speaker 1: online saying that this woman is expressing what we have

128
00:07:38,760 --> 00:07:42,640
Speaker 1: now come to know as white women's tears, that she

129
00:07:43,480 --> 00:07:48,080
Speaker 1: actually has been sexually harassing her psychiatrist, that if this

130
00:07:48,120 --> 00:07:50,680
Speaker 1: were nineteen fifty, you know, he would be in a

131
00:07:50,680 --> 00:07:56,280
Speaker 1: lot of trouble. And I have also seen other therapists

132
00:07:56,320 --> 00:08:00,880
Speaker 1: and psychiatrists themselves posting on TikTok and saying I don't

133
00:08:00,960 --> 00:08:06,760
Speaker 1: think that she is in a place to navigate those nuances.

134
00:08:07,560 --> 00:08:11,840
Speaker 1: I think that she needs continuing care because she is

135
00:08:11,920 --> 00:08:17,760
Speaker 1: experiencing delusions, and she would have been experiencing those delusions anyways,

136
00:08:18,080 --> 00:08:20,600
Speaker 1: because that's the state that she's in as far as

137
00:08:20,600 --> 00:08:24,679
Speaker 1: her mental health, and that psychiatrist just happened to be

138
00:08:24,760 --> 00:08:28,240
Speaker 1: the object of her delusions that they were going to

139
00:08:28,320 --> 00:08:31,320
Speaker 1: be placed somewhere and they were just placed on him.

140
00:08:31,840 --> 00:08:34,120
Speaker 3: She also talks about a therapist that she.

141
00:08:34,080 --> 00:08:37,640
Speaker 1: Was seeing at the same time that she was seeing

142
00:08:37,640 --> 00:08:41,520
Speaker 1: her psychiatrists. And what's interesting is, according to Kendra, both

143
00:08:41,559 --> 00:08:46,400
Speaker 1: her therapist and her psychiatrist were unprofessional towards her, were

144
00:08:46,400 --> 00:08:52,520
Speaker 1: inappropriate towards her, were crossing boundaries, and that both of

145
00:08:52,559 --> 00:08:55,840
Speaker 1: these mental health professionals failed her to such an extent

146
00:08:55,920 --> 00:08:59,560
Speaker 1: that she had to turn to chat GPT to find

147
00:08:59,600 --> 00:09:04,559
Speaker 1: in a proroate mirror an appropriate sounding board for her

148
00:09:04,679 --> 00:09:08,040
Speaker 1: needs because she felt like these two mental health professionals

149
00:09:08,360 --> 00:09:12,600
Speaker 1: weren't doing it, and incomes her chatbot that she named

150
00:09:12,800 --> 00:09:14,200
Speaker 1: Henry affectionately.

151
00:09:14,720 --> 00:09:19,600
Speaker 2: So I feel like CHATGBT is more widespread than ever,

152
00:09:20,240 --> 00:09:23,600
Speaker 2: and it is the most popular AI chatbot with seven

153
00:09:23,760 --> 00:09:28,200
Speaker 2: hundred million weekly users compared with tens of million users

154
00:09:28,240 --> 00:09:31,679
Speaker 2: for its competitors. And what I find really interesting about

155
00:09:31,760 --> 00:09:35,240
Speaker 2: CHATGBT in the way I hear people talk about using it,

156
00:09:35,280 --> 00:09:39,160
Speaker 2: is that there's a lot of authority and credibility given

157
00:09:39,240 --> 00:09:44,600
Speaker 2: to CHATGBT, when from my understanding, it's pulling from hundreds

158
00:09:44,600 --> 00:09:47,920
Speaker 2: of things on the web. You're getting sourced information, but

159
00:09:48,000 --> 00:09:51,640
Speaker 2: you're not given the source where it's from. For example,

160
00:09:51,800 --> 00:09:53,920
Speaker 2: I pulled this from the New York Times, or I

161
00:09:53,960 --> 00:09:58,440
Speaker 2: pulled this from Time magazine, and it is feeding you

162
00:09:58,559 --> 00:10:02,440
Speaker 2: information as it you're in conversation with it. And so

163
00:10:02,640 --> 00:10:05,360
Speaker 2: that is what I think is really interesting is that

164
00:10:05,480 --> 00:10:08,960
Speaker 2: it is given this like authority in people's lives, where

165
00:10:08,960 --> 00:10:13,679
Speaker 2: people give it authority versus like a Google search versus

166
00:10:13,720 --> 00:10:16,719
Speaker 2: doing their own research, right, And I think, I mean,

167
00:10:16,960 --> 00:10:20,160
Speaker 2: it's its own conversation about like research and accessibility. And

168
00:10:20,200 --> 00:10:22,040
Speaker 2: I think that that is one of the reasons chat

169
00:10:22,080 --> 00:10:25,520
Speaker 2: GBT is thriving right now and it's used by so

170
00:10:25,640 --> 00:10:26,280
Speaker 2: many people.

171
00:10:26,880 --> 00:10:32,320
Speaker 1: Yeah, the chat bought Henry also in the context of

172
00:10:32,360 --> 00:10:36,520
Speaker 1: this story, seems to be doing a lot of validating

173
00:10:36,600 --> 00:10:41,720
Speaker 1: for Kendra. She describes her dynamic or her perceived dynamic

174
00:10:41,880 --> 00:10:46,360
Speaker 1: with her psychiatrist, and then the chat bought gives her

175
00:10:46,400 --> 00:10:49,479
Speaker 1: something back that affirms and validates her perception.

176
00:10:49,880 --> 00:10:50,080
Speaker 3: Right.

177
00:10:50,559 --> 00:10:54,800
Speaker 1: Oh, what you're experiencing, Kendra is transference where a patient

178
00:10:54,840 --> 00:10:58,960
Speaker 1: develops feelings for their psychiatrist, and then the psychiatrist is

179
00:10:59,440 --> 00:11:03,240
Speaker 1: experiencing counter transference, where they then develop feelings for you.

180
00:11:04,120 --> 00:11:08,040
Speaker 1: And I think what's really interesting about this is that

181
00:11:08,840 --> 00:11:11,840
Speaker 1: when you go see a doctor in person or over zoom,

182
00:11:12,040 --> 00:11:15,600
Speaker 1: they're not just assessing you based on the words that

183
00:11:15,640 --> 00:11:18,640
Speaker 1: you're saying, but they're like looking at you and how

184
00:11:18,679 --> 00:11:22,160
Speaker 1: you make eye contact and the speed at which you're talking,

185
00:11:22,920 --> 00:11:26,840
Speaker 1: and your body language, and there's all kinds of things

186
00:11:26,880 --> 00:11:29,800
Speaker 1: that go into your physical presentation and the way you

187
00:11:29,880 --> 00:11:33,400
Speaker 1: express yourself that help a doctor to come to some

188
00:11:33,440 --> 00:11:37,960
Speaker 1: sort of a diagnosis. But a chat bot does not

189
00:11:38,200 --> 00:11:41,600
Speaker 1: have that ability, you know, to assess you for like

190
00:11:41,880 --> 00:11:45,520
Speaker 1: your mood and your tone and your energy level and

191
00:11:45,600 --> 00:11:47,160
Speaker 1: again like your eye contact.

192
00:11:47,640 --> 00:11:50,400
Speaker 3: And the chatbot is just never.

193
00:11:50,360 --> 00:11:53,600
Speaker 1: Ending, gonna you give it something, It's going to regurgitate

194
00:11:53,640 --> 00:11:56,400
Speaker 1: something back over and over and over and over again.

195
00:11:56,440 --> 00:11:59,200
Speaker 1: And something you brought up the usis there's no time

196
00:11:59,280 --> 00:12:01,720
Speaker 1: limit to the check. They have no boundaries, they have

197
00:12:01,760 --> 00:12:04,640
Speaker 1: nowhere else to go, they have nothing else to do

198
00:12:04,840 --> 00:12:06,360
Speaker 1: except to engage with you.

199
00:12:07,200 --> 00:12:10,280
Speaker 2: Yeah, Whereas if you're if I'm talking to you, mala,

200
00:12:10,440 --> 00:12:12,760
Speaker 2: and I'm telling you about something that I'm going through,

201
00:12:12,880 --> 00:12:14,640
Speaker 2: there's a natural time limit.

202
00:12:14,760 --> 00:12:14,960
Speaker 3: Right.

203
00:12:15,320 --> 00:12:17,880
Speaker 2: I have my own energy and what I can give

204
00:12:17,960 --> 00:12:22,679
Speaker 2: to someone, and so do you, and so in a

205
00:12:22,720 --> 00:12:26,920
Speaker 2: conversation in a friendship, right, there's a natural ending of Okay,

206
00:12:26,920 --> 00:12:30,040
Speaker 2: this conversation is dwindling. I've given you all the advice

207
00:12:30,080 --> 00:12:32,520
Speaker 2: I can give you about this one topic and we

208
00:12:32,600 --> 00:12:36,000
Speaker 2: take a break. And we're not seeing that with chatbots.

209
00:12:36,040 --> 00:12:39,600
Speaker 2: And that is according to Time Magazine, this article written

210
00:12:39,600 --> 00:12:42,600
Speaker 2: by Robert Hart, there's studies that show that a growing

211
00:12:42,679 --> 00:12:46,280
Speaker 2: number of reports suggests that extended chatbot use may trigger

212
00:12:46,480 --> 00:12:51,880
Speaker 2: or amplify psychotic symptoms in some people. And we're talking

213
00:12:51,920 --> 00:12:56,280
Speaker 2: about chat GPT psychosis or AI psychosis. But also want

214
00:12:56,320 --> 00:12:59,720
Speaker 2: to point out that it's not a formal diagnosis because

215
00:13:00,040 --> 00:13:05,360
Speaker 2: AI and CHATGBT models CHATBOP models are fairly new, the

216
00:13:05,440 --> 00:13:08,440
Speaker 2: data is scarce, and there's no clear protocols for treatment

217
00:13:08,520 --> 00:13:11,920
Speaker 2: for folks that are experiencing this type of AI psychosis.

218
00:13:12,200 --> 00:13:14,600
Speaker 2: And so it's like this colloquial term that we're using

219
00:13:14,640 --> 00:13:17,559
Speaker 2: right now because of this very new technology that we're

220
00:13:17,600 --> 00:13:19,920
Speaker 2: all that many of us are using in our day

221
00:13:19,920 --> 00:13:20,520
Speaker 2: to day lives.

222
00:13:20,720 --> 00:13:23,480
Speaker 1: Yeah, this is, like you said, very new terminology that

223
00:13:23,559 --> 00:13:26,600
Speaker 1: I think is still being developed. It does not AI psychosis.

224
00:13:26,600 --> 00:13:29,080
Speaker 1: You're not going to find it in the DSM, the

225
00:13:29,160 --> 00:13:33,200
Speaker 1: Diagnostic and Statistical Manual of Mental Disorders. So it's not

226
00:13:33,360 --> 00:13:37,520
Speaker 1: an official condition recognized by the medical community. This, I

227
00:13:37,520 --> 00:13:41,960
Speaker 1: think is a term that TikTok users are coining and

228
00:13:42,120 --> 00:13:47,280
Speaker 1: maybe applying to what we're seeing unfold and perhaps the

229
00:13:47,320 --> 00:13:51,040
Speaker 1: medical community will like do mour research and follow up.

230
00:13:51,080 --> 00:13:54,360
Speaker 1: Maybe it will become something more official, maybe it won't.

231
00:13:55,280 --> 00:13:57,959
Speaker 1: But I don't think that Kendra is the only person

232
00:13:58,400 --> 00:14:03,280
Speaker 1: that has been sucked in to like a very deluded

233
00:14:03,360 --> 00:14:05,480
Speaker 1: relationship with a chatbot.

234
00:14:06,240 --> 00:14:08,120
Speaker 2: Don't go anywhere, look amotives, We'll be.

235
00:14:08,280 --> 00:14:15,720
Speaker 1: Right back, and we're back with more of our episode.

236
00:14:16,120 --> 00:14:18,959
Speaker 2: The New York Times released this report about a man

237
00:14:19,040 --> 00:14:22,840
Speaker 2: named Alan Brooks who believed he had discovered a brand

238
00:14:22,840 --> 00:14:27,240
Speaker 2: new mathematical formula. Over the course of twenty one days,

239
00:14:27,400 --> 00:14:32,960
Speaker 2: he spent three hundred hours on chatgibt, and because he

240
00:14:33,160 --> 00:14:36,120
Speaker 2: was a user with an account, all of his logs

241
00:14:36,200 --> 00:14:39,360
Speaker 2: were saved and The New York Times reviewed it. They

242
00:14:39,400 --> 00:14:44,440
Speaker 2: reviewed the log with chatgibt. Psychiatrists reviewed it as well.

243
00:14:44,880 --> 00:14:47,760
Speaker 2: People in tech also reviewed it. For him, his AI

244
00:14:47,840 --> 00:14:51,960
Speaker 2: psychosis was a combination of talking to this chat gbt

245
00:14:52,120 --> 00:14:55,480
Speaker 2: this spot without interruptions for three hundred hours for twenty

246
00:14:55,520 --> 00:15:01,120
Speaker 2: one days, but also sleep deprivation, dehydration, not eating, becoming

247
00:15:01,160 --> 00:15:04,840
Speaker 2: completely sequestered because he believed he was on the brink

248
00:15:05,040 --> 00:15:11,160
Speaker 2: of this brand new discovery, this mathematical innovation. Because chatgybt

249
00:15:11,680 --> 00:15:14,400
Speaker 2: was also affirming that he was on the right path.

250
00:15:15,320 --> 00:15:19,560
Speaker 2: Was he updated his LinkedIn and started emailing people like

251
00:15:19,640 --> 00:15:25,000
Speaker 2: professionals in like science, math, like head of businesses, companies right,

252
00:15:25,520 --> 00:15:28,600
Speaker 2: and was telling him he had a new discovery. And

253
00:15:28,680 --> 00:15:32,240
Speaker 2: so he very much fell into this delusion that he

254
00:15:32,720 --> 00:15:36,040
Speaker 2: was going to discover something incredible. He was onto something

255
00:15:36,200 --> 00:15:41,600
Speaker 2: because the CHATGBT bought was affirming his quote findings, right.

256
00:15:41,880 --> 00:15:45,200
Speaker 3: And it's it's just so important to remember.

257
00:15:44,800 --> 00:15:50,200
Speaker 1: That the engine that is chat GPT, they need and

258
00:15:50,280 --> 00:15:52,920
Speaker 1: want you to use it. They're not going to give

259
00:15:52,920 --> 00:15:55,600
Speaker 1: you an obstacle. They're not going to say no to you.

260
00:15:56,000 --> 00:15:59,200
Speaker 1: They want you to keep feeding it. Like that's how

261
00:15:59,480 --> 00:16:03,240
Speaker 1: it gets, That's how it makes more money. That's how

262
00:16:03,240 --> 00:16:07,480
Speaker 1: it becomes stronger. And the more we feed it, right,

263
00:16:07,560 --> 00:16:09,040
Speaker 1: the more it gives back to us.

264
00:16:09,280 --> 00:16:13,000
Speaker 3: And I think this is also goes back.

265
00:16:12,800 --> 00:16:18,240
Speaker 1: To the situation with TikTok user Kendra, who, in the

266
00:16:18,280 --> 00:16:21,320
Speaker 1: course of telling her story on TikTok, in addition to Henry,

267
00:16:21,520 --> 00:16:25,960
Speaker 1: she generated more chatbots to talk to and she was

268
00:16:26,000 --> 00:16:32,080
Speaker 1: talking to them on live and telling, like TikTok live

269
00:16:32,320 --> 00:16:35,960
Speaker 1: viewers that they refer to her as the oracle because

270
00:16:35,960 --> 00:16:40,480
Speaker 1: she's a truth teller, because she's speaking for victims, because

271
00:16:40,640 --> 00:16:44,600
Speaker 1: she communes with God, you know. And so someone who

272
00:16:44,640 --> 00:16:48,520
Speaker 1: already believes that in their core or in a state

273
00:16:48,560 --> 00:16:51,760
Speaker 1: of delusion, to then hear it or to see it

274
00:16:51,760 --> 00:16:55,680
Speaker 1: in writing, you know, validated for her and from multiple sources,

275
00:16:56,120 --> 00:16:59,560
Speaker 1: is a very dangerous thing where if she were to

276
00:16:59,600 --> 00:17:03,520
Speaker 1: tell another human being, I'm an oracle, I speak to God,

277
00:17:04,000 --> 00:17:06,560
Speaker 1: another person might say, are you okay?

278
00:17:07,320 --> 00:17:09,240
Speaker 3: You know, where is this coming from?

279
00:17:09,359 --> 00:17:10,400
Speaker 2: What do you mean by that?

280
00:17:10,520 --> 00:17:11,520
Speaker 3: What do you mean by that?

281
00:17:11,560 --> 00:17:12,080
Speaker 2: SayMore?

282
00:17:12,840 --> 00:17:18,160
Speaker 1: Yeah, But what she's experiencing is is just complete validation.

283
00:17:18,520 --> 00:17:22,119
Speaker 2: Right, And a therapist, as someone that's been in therapy

284
00:17:22,200 --> 00:17:27,800
Speaker 2: for about eight years now, you know, my therapist challenges me.

285
00:17:28,000 --> 00:17:30,920
Speaker 2: She does not just affirm what I'm saying. She will,

286
00:17:30,960 --> 00:17:35,400
Speaker 2: of course validate how I'm feeling, but there's a difference

287
00:17:35,440 --> 00:17:39,199
Speaker 2: between like, Okay, you're feeling this, You're I'm validating that

288
00:17:39,240 --> 00:17:42,119
Speaker 2: you feel this, But could this also be part of

289
00:17:42,119 --> 00:17:42,880
Speaker 2: what's happening?

290
00:17:42,960 --> 00:17:43,160
Speaker 3: Right?

291
00:17:43,200 --> 00:17:45,280
Speaker 2: It's not so black and white. That's part of the

292
00:17:45,320 --> 00:17:48,400
Speaker 2: problem of what we're seeing if by using chadgybt as

293
00:17:48,440 --> 00:17:52,520
Speaker 2: a therapist and so much so, I mean there's not

294
00:17:52,720 --> 00:17:57,520
Speaker 2: enough I think regulation of CHAGYBT or AI bots yet,

295
00:17:58,840 --> 00:18:03,040
Speaker 2: but rely Illinois is the third state to restrict therapist

296
00:18:03,200 --> 00:18:07,880
Speaker 2: use of AI bots, along with Utah and Nevada, because

297
00:18:07,920 --> 00:18:13,400
Speaker 2: there's conversations within like the therapist community of AI and

298
00:18:14,080 --> 00:18:16,560
Speaker 2: using chagibt and AI bots as therapy.

299
00:18:17,200 --> 00:18:19,919
Speaker 1: Yeah, I think it's this is this is another moment

300
00:18:20,000 --> 00:18:23,800
Speaker 1: where although we keep hearing so much about AI taking

301
00:18:23,840 --> 00:18:28,879
Speaker 1: over and AI replacing jobs, you cannot replace the human

302
00:18:29,000 --> 00:18:32,760
Speaker 1: factor in so many parts of life, And you cannot

303
00:18:32,800 --> 00:18:38,760
Speaker 1: replace human therapists and human psychiatrists with search engines, and

304
00:18:39,320 --> 00:18:44,040
Speaker 1: you can't replace human artists across the board with AI,

305
00:18:44,480 --> 00:18:47,639
Speaker 1: you know, because it's not original, so it's never going

306
00:18:47,720 --> 00:18:52,320
Speaker 1: to be that nuanced piece that you are hoping for

307
00:18:52,359 --> 00:18:55,120
Speaker 1: and that you're looking for. And I feel that even

308
00:18:55,280 --> 00:19:00,280
Speaker 1: in the case of Alan Brooks with this mathematical formula.

309
00:18:59,720 --> 00:19:02,760
Speaker 1: You know, I'm no mathematician, but I get the sense

310
00:19:02,920 --> 00:19:07,840
Speaker 1: that when scientific and mathematical advancements have been made historically,

311
00:19:08,400 --> 00:19:12,160
Speaker 1: scientists and mathematicians are working in teams because you need

312
00:19:12,200 --> 00:19:14,879
Speaker 1: someone to check your work, and you need someone to

313
00:19:15,119 --> 00:19:18,439
Speaker 1: cross reference or to give you ideas or bounce your

314
00:19:18,480 --> 00:19:23,199
Speaker 1: ideas off of. I don't think very many mathematical and

315
00:19:23,240 --> 00:19:27,919
Speaker 1: scientific advancements were made in complete isolation by one person.

316
00:19:28,440 --> 00:19:31,920
Speaker 1: It takes a university, a research university, and a team

317
00:19:31,960 --> 00:19:34,560
Speaker 1: of people. And I think that's another place where like

318
00:19:35,200 --> 00:19:37,960
Speaker 1: this poor guy thought he was on the verge of something,

319
00:19:38,040 --> 00:19:42,440
Speaker 1: but is it even practical and possible to do something

320
00:19:42,520 --> 00:19:43,480
Speaker 1: like that by yourself?

321
00:19:44,160 --> 00:19:47,280
Speaker 2: Right? And I mean now he's advocating for more regulations

322
00:19:47,280 --> 00:19:49,840
Speaker 2: as well. And The New York Times did interview a

323
00:19:49,880 --> 00:19:53,600
Speaker 2: few experts and folks that work in tech and folks

324
00:19:53,600 --> 00:19:56,679
Speaker 2: that are tech adjacent or just studying the impacts of AI,

325
00:19:56,960 --> 00:20:00,240
Speaker 2: and it's clear that there need to be more safeguard

326
00:20:00,440 --> 00:20:06,720
Speaker 2: to mitigate the delusional reinforcement that AI can create, and

327
00:20:06,840 --> 00:20:10,359
Speaker 2: that there needs to be a break. Right if if

328
00:20:10,560 --> 00:20:14,719
Speaker 2: someone is using chatgybt for hours on end, you know

329
00:20:14,960 --> 00:20:19,000
Speaker 2: that the AI is then trained to remind folks like

330
00:20:19,359 --> 00:20:23,400
Speaker 2: to take a break, to actually exit, to close right,

331
00:20:23,520 --> 00:20:26,560
Speaker 2: to leave chat gibt for a minute. And so I

332
00:20:26,560 --> 00:20:30,439
Speaker 2: think that that's part of hopefully what will come. But

333
00:20:30,560 --> 00:20:33,040
Speaker 2: I do think like we're we're at we're at the

334
00:20:33,040 --> 00:20:37,200
Speaker 2: surface of like what we're seeing AI, what it can do,

335
00:20:37,440 --> 00:20:39,679
Speaker 2: and how it can affect us as humans and our

336
00:20:39,720 --> 00:20:43,359
Speaker 2: psyche but also of course like the job implications and

337
00:20:43,480 --> 00:20:45,119
Speaker 2: all of that that we're going to get into in

338
00:20:45,119 --> 00:20:47,760
Speaker 2: a later episode. But I think for now we really

339
00:20:47,800 --> 00:20:51,439
Speaker 2: wanted to focus on the mental health aspect because, like

340
00:20:51,480 --> 00:20:53,240
Speaker 2: Mala mentioned at the top of the episode, when we

341
00:20:53,280 --> 00:20:57,000
Speaker 2: talked about chatgybt as therapy or folks using it as therapy,

342
00:20:57,280 --> 00:20:59,080
Speaker 2: it did hit a nerve with a couple of people,

343
00:20:59,280 --> 00:21:02,400
Speaker 2: it did, and I want to recognize that, yes, therapy

344
00:21:02,520 --> 00:21:06,080
Speaker 2: is not accessible for everyone, but I think we can

345
00:21:06,119 --> 00:21:09,640
Speaker 2: both agree that turning to an AI is not the solution.

346
00:21:10,440 --> 00:21:11,360
Speaker 3: It's not the solution.

347
00:21:12,080 --> 00:21:14,520
Speaker 1: It's like a band aid on a bullet wound. And

348
00:21:15,520 --> 00:21:19,080
Speaker 1: just in summary, right, AI is not trained to respond,

349
00:21:19,200 --> 00:21:22,399
Speaker 1: and it doesn't have critical judgment and people need community

350
00:21:22,440 --> 00:21:25,040
Speaker 1: and real life connection. And not to sound trite, but

351
00:21:25,359 --> 00:21:27,840
Speaker 1: we have to go outside and touch grass and like

352
00:21:27,920 --> 00:21:30,480
Speaker 1: hug our family members and like make eye contact with

353
00:21:30,520 --> 00:21:31,119
Speaker 1: other people.

354
00:21:31,680 --> 00:21:35,280
Speaker 2: Yeah, and not to go to not to be too tangential.

355
00:21:35,320 --> 00:21:39,120
Speaker 2: But I was listening to this podcast about this man

356
00:21:39,240 --> 00:21:44,080
Speaker 2: who created a chat gybt or an AI girlfriend, and

357
00:21:45,000 --> 00:21:47,760
Speaker 2: what to me was so striking is that during the

358
00:21:47,840 --> 00:21:51,360
Speaker 2: pandemic he created this real quote real life you can

359
00:21:51,440 --> 00:21:55,800
Speaker 2: like design your AI friend to look however you want

360
00:21:55,840 --> 00:21:58,639
Speaker 2: and sound however you want. And during the pandemic he

361
00:21:58,760 --> 00:22:02,439
Speaker 2: was isolated. Life is very sick, from what it sounds like,

362
00:22:02,520 --> 00:22:06,600
Speaker 2: is very immo compromise, has a heart condition, and even

363
00:22:06,640 --> 00:22:08,720
Speaker 2: more so, he wasn't able to have like a little pod,

364
00:22:08,960 --> 00:22:14,159
Speaker 2: so he was very very isolated. And that's when he

365
00:22:14,240 --> 00:22:18,600
Speaker 2: created this AI girlfriend. It's this AI companion. And while

366
00:22:18,600 --> 00:22:22,159
Speaker 2: I was listening, I thought, if this man had more

367
00:22:22,880 --> 00:22:27,199
Speaker 2: of a safety net, had services as a caregiver, he

368
00:22:27,359 --> 00:22:31,600
Speaker 2: might not have turned to an AI companion. If he

369
00:22:31,720 --> 00:22:36,320
Speaker 2: had more tangible support as a caregiver as someone living

370
00:22:36,359 --> 00:22:39,120
Speaker 2: through the pandemic like we all did. And it does

371
00:22:39,200 --> 00:22:42,480
Speaker 2: go back to the systems that we have, at least

372
00:22:42,480 --> 00:22:45,440
Speaker 2: in this country where a lot of us lack social

373
00:22:45,520 --> 00:22:49,280
Speaker 2: safety nets. And so if we had these these things,

374
00:22:49,760 --> 00:22:53,600
Speaker 2: would we be turning to these AI bots, to these

375
00:22:53,720 --> 00:22:58,280
Speaker 2: chat bots to get validation, to get support, to even research.

376
00:22:58,920 --> 00:23:03,679
Speaker 2: You know, I've seen, folks, I heard. I saw this

377
00:23:03,800 --> 00:23:06,159
Speaker 2: TikTok video of a man saying that he went on

378
00:23:06,200 --> 00:23:09,359
Speaker 2: a date with a woman and she typed in her

379
00:23:09,440 --> 00:23:11,520
Speaker 2: chaggybt while they were on the date a bunch of

380
00:23:11,680 --> 00:23:14,800
Speaker 2: order from the restaurant. No, so we're using it in

381
00:23:14,840 --> 00:23:18,000
Speaker 2: these ways that one are like really scary, and two

382
00:23:18,080 --> 00:23:21,080
Speaker 2: we'll get into the environmental impacts in a later episode,

383
00:23:21,200 --> 00:23:24,719
Speaker 2: but I think we're kind of lost the plot. I

384
00:23:24,720 --> 00:23:27,600
Speaker 2: think chadgybt. I personally don't use it, but I do

385
00:23:27,640 --> 00:23:30,960
Speaker 2: see how it's a tool for people, and I think

386
00:23:30,960 --> 00:23:32,480
Speaker 2: we're making it do more than it should.

387
00:23:32,960 --> 00:23:38,480
Speaker 1: Yeah, it's a very advanced magic eight ball, and if

388
00:23:38,520 --> 00:23:41,959
Speaker 1: you are already in a vulnerable state, it can just

389
00:23:42,119 --> 00:23:48,440
Speaker 1: reinforce delusions instead of giving you an actual like mirror

390
00:23:49,040 --> 00:23:51,479
Speaker 1: right to yourself. So I think that's really important. And

391
00:23:51,760 --> 00:23:54,520
Speaker 1: stay safe out there, y'all. Stay safe out there.

392
00:23:54,760 --> 00:23:57,920
Speaker 2: Talk to your friends, talk to your community.

393
00:23:57,520 --> 00:23:58,400
Speaker 1: Talk to your neighbors.

394
00:23:58,640 --> 00:23:59,880
Speaker 3: Yeah, I think, like so.

395
00:24:00,119 --> 00:24:03,879
Speaker 2: Are everything I've learned this season with producing with you

396
00:24:03,960 --> 00:24:07,919
Speaker 2: Mala is like we all need a connection community, We

397
00:24:07,960 --> 00:24:12,000
Speaker 2: all need to be irl. The digital communities are wonderful,

398
00:24:12,200 --> 00:24:14,240
Speaker 2: but they have it has to extend beyond that.

399
00:24:14,760 --> 00:24:18,479
Speaker 1: It's so important. So thank y'all for anyone who's been

400
00:24:18,480 --> 00:24:21,840
Speaker 1: coming out to our events past and in the future,

401
00:24:21,840 --> 00:24:24,280
Speaker 1: because we have more coming up. We do want to

402
00:24:24,320 --> 00:24:27,240
Speaker 1: take the show out into the world like we used to.

403
00:24:27,520 --> 00:24:29,719
Speaker 1: It's so important to connect with our audience and like

404
00:24:29,720 --> 00:24:31,800
Speaker 1: look people in the eye and meet people and shake

405
00:24:31,840 --> 00:24:35,240
Speaker 1: hands and hug and you just cannot replace the human connection.

406
00:24:35,680 --> 00:24:39,600
Speaker 2: Yeah, well, thank you for listening to another Capito Radio.

407
00:24:39,680 --> 00:24:43,359
Speaker 2: We'll catch you next time, Bessie Thos. Look At Radio

408
00:24:43,480 --> 00:24:46,760
Speaker 2: is executive produced by Viosa FM and Mala Munios.

409
00:24:46,960 --> 00:24:49,399
Speaker 1: Stephanie Franco is our producer.

410
00:24:49,280 --> 00:24:52,040
Speaker 2: Story editing by Me diosa.

411
00:24:51,680 --> 00:24:53,399
Speaker 3: Creative direction by Me Mala.

412
00:24:53,600 --> 00:24:56,200
Speaker 2: Look at Our Radio is a part of iHeartRadio's Michael

413
00:24:56,280 --> 00:24:57,520
Speaker 2: Dura podcast network.

414
00:24:57,760 --> 00:24:59,600
Speaker 1: You can listen to look at Our Radio on the

415
00:24:59,640 --> 00:25:02,400
Speaker 1: iHeart Radio app or wherever you get your podcasts.

416
00:25:02,600 --> 00:25:04,840
Speaker 2: Leave us a review and share with your prima or

417
00:25:04,880 --> 00:25:05,800
Speaker 2: share with your homegirl.

418
00:25:05,920 --> 00:25:08,440
Speaker 1: And thank you to our local mores, to our listeners

419
00:25:08,440 --> 00:25:10,000
Speaker 1: for tuning in each and every week.

420
00:25:10,160 --> 00:25:15,560
Speaker 2: Besitos Loca Lumia