1
00:00:11,440 --> 00:00:14,400
>> Peter: welcome to Episode nine of AI Church Toolkit,

2
00:00:14,400 --> 00:00:17,400
the podcast that equips church leaders with practical tools

3
00:00:17,400 --> 00:00:19,600
for faithful ministry in a digital world.

4
00:00:20,160 --> 00:00:23,080
I'm Peter Lavenstrong solo hosting today while

5
00:00:23,080 --> 00:00:25,760
my co host Mercedes takes a short personal leave.

6
00:00:26,150 --> 00:00:28,710
She'll be back with us for our next podcast episode.

7
00:00:29,510 --> 00:00:32,070
Today I'm joined by the Reverend Lauren Grubaugh

8
00:00:32,070 --> 00:00:34,990
Thomas. Reverend Lauren empowers communities

9
00:00:34,990 --> 00:00:37,990
to embrace the sacred act of nonviolent

10
00:00:37,990 --> 00:00:40,710
social change through her ministry as a church planter,

11
00:00:40,950 --> 00:00:43,950
movement chaplain and writer. She serves

12
00:00:43,950 --> 00:00:46,950
as founding vicar of Holy Companion Episcopal Church

13
00:00:46,950 --> 00:00:49,910
in the Denver suburbs, a vibrant young community

14
00:00:50,150 --> 00:00:53,110
of justice seekers and and the first church plant in the

15
00:00:53,110 --> 00:00:55,990
Episcopal Church in Colorado in the last 15 years.

16
00:00:56,630 --> 00:00:59,510
Lauren is a 2022 Trinity

17
00:00:59,510 --> 00:01:02,510
Leadership Fellow and earned her Master of Divinity

18
00:01:02,510 --> 00:01:05,070
degree from Fuller Theological Seminary with an

19
00:01:05,070 --> 00:01:07,950
emphasis in Christian ethics. You can find

20
00:01:07,950 --> 00:01:10,870
Lauren writing and podcasting at the intersection

21
00:01:10,870 --> 00:01:13,590
of spiritual transformation and social change

22
00:01:13,830 --> 00:01:16,630
at her substack A Soulful Revolution.

23
00:01:17,680 --> 00:01:20,520
In today's episode, we dig into the ethical and

24
00:01:20,520 --> 00:01:23,480
theological questions surrounding emerging technologies

25
00:01:23,480 --> 00:01:26,360
like AI. Whether you're just starting to explore

26
00:01:26,360 --> 00:01:28,800
AI or already using it in your ministry,

27
00:01:29,040 --> 00:01:32,000
this conversation invites you into a deeper sense of

28
00:01:32,000 --> 00:01:35,000
curiosity, discernment, and faithful

29
00:01:35,000 --> 00:01:37,360
imagination. So let's get started.

30
00:01:44,240 --> 00:01:45,600
All right, welcome Lauren

31
00:01:46,400 --> 00:01:47,680
>> Lauren Grubaugh Thomas: Thanks. It's great to be here.

32
00:01:49,900 --> 00:01:52,620
>> Peter: So we always love to start with

33
00:01:52,780 --> 00:01:55,740
a particular question for all of our guests and that

34
00:01:55,740 --> 00:01:58,700
is this. If you had to pick one fictional sci

35
00:01:58,700 --> 00:02:01,420
fi world that captures where you think we're headed with

36
00:02:01,420 --> 00:02:03,980
generative AI, whether hopeful or

37
00:02:03,980 --> 00:02:06,620
cautionary, which one would you pick and why?

38
00:02:08,140 --> 00:02:10,900
>> Lauren Grubaugh Thomas: I love this question and maybe

39
00:02:10,900 --> 00:02:13,900
going a little bit out of left field with this, but

40
00:02:14,870 --> 00:02:17,750
Parable of the Sower is a book that has had a huge impact

41
00:02:17,910 --> 00:02:20,870
on how I think about the changing world that we live in.

42
00:02:20,870 --> 00:02:23,750
It's a book by the Afro futurist writer Octavia Butler,

43
00:02:24,230 --> 00:02:27,230
and her thesis of the book is that God

44
00:02:27,230 --> 00:02:29,830
is change and whatever you change

45
00:02:29,990 --> 00:02:32,830
changes you shape change. That's an almost

46
00:02:32,830 --> 00:02:35,310
direct quote from the book and the the main

47
00:02:35,310 --> 00:02:38,070
character is Lauren Olamina, who's a young black

48
00:02:38,070 --> 00:02:40,510
woman growing up in Los Angeles in

49
00:02:40,510 --> 00:02:41,430
2025.

50
00:02:43,750 --> 00:02:46,660
That in and of itself is R.A. um

51
00:02:46,660 --> 00:02:49,560
um. But it was written by Butler about 30

52
00:02:49,560 --> 00:02:52,480
years ago, 40 years ago, um um. And she

53
00:02:52,560 --> 00:02:55,480
was looking at the world as it was at

54
00:02:55,480 --> 00:02:58,320
that time and if we didn't pump the brakes, if we didn't change

55
00:02:58,400 --> 00:03:01,120
course, what is the world that we might be living

56
00:03:01,120 --> 00:03:03,440
in, particularly in terms of

57
00:03:04,000 --> 00:03:06,640
unresolved undealt with white supremacy,

58
00:03:07,520 --> 00:03:10,520
climate, um, degradation? Those

59
00:03:10,520 --> 00:03:12,720
are those are really some of the main factors. Um, um, and

60
00:03:12,720 --> 00:03:15,600
authoritarianism. And so the way that

61
00:03:16,540 --> 00:03:19,530
the book plays out and as it relates to AI, uh,

62
00:03:19,530 --> 00:03:22,100
for me is that you have this

63
00:03:22,100 --> 00:03:24,860
character who is committed to

64
00:03:24,860 --> 00:03:27,660
living with empathy, which in some ways she is

65
00:03:27,660 --> 00:03:30,300
blessed and cursed with. Um, she has this hyper

66
00:03:30,380 --> 00:03:33,380
empathy that she lives with. Uh, what does it look like to

67
00:03:33,380 --> 00:03:36,340
live as an empathic person? And what does it look like

68
00:03:36,340 --> 00:03:39,140
to live as an agentic person, A

69
00:03:39,140 --> 00:03:42,140
person who claims their agency and lives

70
00:03:42,140 --> 00:03:44,780
out of their agency in a world where

71
00:03:44,940 --> 00:03:47,420
choices are increasingly being stripped away

72
00:03:48,040 --> 00:03:51,000
and where empathy is a liability

73
00:03:51,720 --> 00:03:54,560
and for her is a very like physical, embodied kind of

74
00:03:54,560 --> 00:03:57,040
liability. And what I

75
00:03:57,040 --> 00:03:59,960
appreciate about Butler's work is that

76
00:04:00,040 --> 00:04:03,000
she doesn't lead us to

77
00:04:03,080 --> 00:04:05,720
a happy place of resolution

78
00:04:05,880 --> 00:04:08,840
by the end of the book. There is

79
00:04:08,840 --> 00:04:11,320
some happy ending for Lauren Olamina.

80
00:04:11,720 --> 00:04:14,560
But this theology that she wrestles with throughout the

81
00:04:14,560 --> 00:04:17,440
book about God being change and that we can shape change,

82
00:04:17,440 --> 00:04:19,420
that we could be agents in

83
00:04:20,460 --> 00:04:23,260
the world as the world is imposing change

84
00:04:23,260 --> 00:04:26,100
on us, um, is

85
00:04:26,100 --> 00:04:28,980
this open ended question that

86
00:04:28,980 --> 00:04:31,900
we leave that story with and that leads us into book two, Parable of

87
00:04:31,900 --> 00:04:34,540
the Talents. But it's this open ended question of

88
00:04:34,540 --> 00:04:36,860
like, how are we going to engage

89
00:04:37,420 --> 00:04:40,300
in shaping change ourselves? And

90
00:04:40,780 --> 00:04:43,580
so when I think about AI, I find

91
00:04:43,580 --> 00:04:45,660
hope in Butler's vision of being

92
00:04:46,340 --> 00:04:49,180
agents of change in a change filled world where the change

93
00:04:49,180 --> 00:04:51,860
is rapidly accelerating. Um,

94
00:04:51,860 --> 00:04:54,660
because there are a lot of people who look at

95
00:04:55,620 --> 00:04:58,340
the change in our world and

96
00:04:58,340 --> 00:05:00,900
just see it through the eyes of doom and gloom and

97
00:05:00,900 --> 00:05:03,700
despair. Like this is all happening so fast,

98
00:05:03,780 --> 00:05:06,700
it's piling on top of us. Um, this

99
00:05:06,700 --> 00:05:09,700
technology is inevitable. We just have

100
00:05:09,700 --> 00:05:12,460
to accept the ways that it's rolling out, the ways that

101
00:05:12,460 --> 00:05:14,960
it's um, programmed are

102
00:05:15,440 --> 00:05:16,800
going to continue to

103
00:05:18,000 --> 00:05:20,800
cause some, would, some, some people say

104
00:05:20,800 --> 00:05:23,680
harm. We're going to talk more about this, but it's going to cause harm.

105
00:05:23,680 --> 00:05:26,600
Here's the good it's going to produce and there's not a lot that we can

106
00:05:26,600 --> 00:05:29,519
do about it. And as I understand generative

107
00:05:29,519 --> 00:05:32,280
AI and AI in general is that we actually

108
00:05:32,280 --> 00:05:35,200
have great potential to shape it like that.

109
00:05:35,360 --> 00:05:38,240
It is a technology that is being shaped actively

110
00:05:38,480 --> 00:05:41,480
as we use it, as we participate in it. And the

111
00:05:41,480 --> 00:05:44,350
biases that are built in and the ways that

112
00:05:44,350 --> 00:05:45,670
it can be engaged as a tool

113
00:05:47,350 --> 00:05:50,070
are things that we have the agency to shape.

114
00:05:50,710 --> 00:05:53,350
And so yeah, it's this overwhelming

115
00:05:54,150 --> 00:05:56,870
new technology in terms of its potential

116
00:05:57,030 --> 00:05:59,950
for change. Um, and we are agents of

117
00:05:59,950 --> 00:06:00,230
change.

118
00:06:01,270 --> 00:06:04,070
>> Peter: Sure, I love that. So yeah, reclaiming agency,

119
00:06:04,310 --> 00:06:07,190
reclaiming empathy, seeing God

120
00:06:07,190 --> 00:06:10,020
in the Change. I have not read

121
00:06:10,100 --> 00:06:12,940
Parable of the Sower yet, but it's on. Been on my list for a

122
00:06:12,940 --> 00:06:15,900
while. So, um, I appreciate that reminder to go

123
00:06:15,900 --> 00:06:18,900
back and, and check that out. Cool. Thank you.

124
00:06:19,220 --> 00:06:19,700
>> Lauren Grubaugh Thomas: Yeah.

125
00:06:19,860 --> 00:06:21,540
>> Peter: Yeah, Sounds like it is very

126
00:06:23,380 --> 00:06:25,940
prescient, uh, for our times in a. In a variety.

127
00:06:25,940 --> 00:06:28,940
>> Lauren Grubaugh Thomas: Yeah. Eerily so often. Eerily so. But

128
00:06:28,940 --> 00:06:29,700
it's a great read.

129
00:06:30,100 --> 00:06:31,860
>> Peter: Yeah. Good. Okay.

130
00:06:32,290 --> 00:06:34,020
Um, so now as we

131
00:06:34,900 --> 00:06:37,160
dive into talking more about,

132
00:06:37,830 --> 00:06:40,680
uh, generative AI and ethics

133
00:06:40,680 --> 00:06:43,350
and how it can be, uh, used, uh,

134
00:06:43,350 --> 00:06:46,200
faithfully and ethically by church leaders, I would love to,

135
00:06:46,200 --> 00:06:49,120
you know, begin by grounding this and uh, sharing a little bit

136
00:06:49,120 --> 00:06:52,040
about what your experience with generative AI has

137
00:06:52,040 --> 00:06:55,000
been, whether that's from your own use or seeing it used out

138
00:06:55,000 --> 00:06:57,970
in the wild, uh, so to say, um,

139
00:06:57,970 --> 00:07:00,920
and yeah, share what that has been like

140
00:07:00,920 --> 00:07:01,720
so far for you.

141
00:07:02,370 --> 00:07:04,890
>> Lauren Grubaugh Thomas: Yeah, so I'll admit that my

142
00:07:04,890 --> 00:07:07,730
engagement has been reticent and

143
00:07:07,810 --> 00:07:10,690
I have been trying to use

144
00:07:10,690 --> 00:07:12,860
the technology in ways that are, um,

145
00:07:14,210 --> 00:07:17,210
where I, where I'm really mindful of what my agency is

146
00:07:17,210 --> 00:07:20,210
in it. So I've used chatgpt

147
00:07:20,850 --> 00:07:23,410
some to be able to shape my

148
00:07:23,410 --> 00:07:26,290
resume, to be able to come up with

149
00:07:26,610 --> 00:07:29,040
lists of things like for

150
00:07:29,120 --> 00:07:31,730
recipes for, um,

151
00:07:31,730 --> 00:07:34,480
I've used it to generate some to do lists. I've used it for

152
00:07:34,480 --> 00:07:37,280
creating some outlines, um, for myself

153
00:07:37,360 --> 00:07:40,320
based on manuscripts and sermons

154
00:07:40,320 --> 00:07:43,240
and the like. But that's all. That's been pretty limited. The

155
00:07:43,240 --> 00:07:45,600
more robust ways of using it have been pretty limited.

156
00:07:45,810 --> 00:07:48,680
Um, I think in some ways it feels like generative

157
00:07:48,680 --> 00:07:51,440
AI continues to be thrust upon me and

158
00:07:51,930 --> 00:07:54,920
uh, opting out of it, Opting out of it feels

159
00:07:54,920 --> 00:07:57,920
increasingly hard. So, I mean,

160
00:07:57,920 --> 00:08:00,880
Microsoft sometimes ends up, for whatever reason,

161
00:08:00,960 --> 00:08:03,600
even though I've asked my browser not to do this,

162
00:08:03,760 --> 00:08:06,640
Bing is sometimes my default. And

163
00:08:06,800 --> 00:08:09,440
Bing uses a generative AI to

164
00:08:09,440 --> 00:08:12,400
respond to search queries. And so

165
00:08:12,400 --> 00:08:15,200
I have actually found those to be helpful at times

166
00:08:15,200 --> 00:08:18,200
and also misleading at times, um, because of where

167
00:08:18,200 --> 00:08:21,080
they're drawing inform where information is being

168
00:08:21,080 --> 00:08:23,360
drawn from. The source material is not always,

169
00:08:23,950 --> 00:08:26,600
um, being compiled in ways that are accurate or

170
00:08:26,600 --> 00:08:29,460
helpful. But sometimes I have

171
00:08:29,460 --> 00:08:32,300
actually used it recently to search for quotes from

172
00:08:32,300 --> 00:08:35,220
books. And it

173
00:08:35,220 --> 00:08:38,060
was really helpful at helping me to trace my steps back to where

174
00:08:38,300 --> 00:08:40,540
that quote that I was trying to recollect came from.

175
00:08:41,340 --> 00:08:44,180
So those things. And then on

176
00:08:44,180 --> 00:08:46,780
occasion for image generation,

177
00:08:47,100 --> 00:08:49,580
I've used on Canva, the

178
00:08:49,580 --> 00:08:51,340
generative AI feature on there.

179
00:08:51,600 --> 00:08:54,510
Um, and the funniest example

180
00:08:54,510 --> 00:08:57,230
of this is that early on, this is like when I think this may have been

181
00:08:57,230 --> 00:08:59,950
ChatGPT but there was something that had a new

182
00:09:00,270 --> 00:09:02,670
platform, a new AI had recently been launched

183
00:09:03,150 --> 00:09:06,030
and I was trying to generate a

184
00:09:06,030 --> 00:09:08,750
logo for our church plant

185
00:09:09,150 --> 00:09:11,830
and was wanting to integrate the

186
00:09:11,830 --> 00:09:14,750
image of the bread and the cup and

187
00:09:14,750 --> 00:09:17,630
hands. Lots of different hands being laid on the bread to

188
00:09:17,630 --> 00:09:19,710
bless the bread. Oh yes, you do,

189
00:09:20,270 --> 00:09:23,210
absolutely. And so I put in all

190
00:09:23,210 --> 00:09:26,090
these inputs of what I wanted and particularly was really curious

191
00:09:26,090 --> 00:09:28,930
about like all these different hands of different races and

192
00:09:29,260 --> 00:09:32,170
um, having. And different sizes and ages, you know,

193
00:09:32,170 --> 00:09:35,170
represented. And I, I got the image

194
00:09:35,170 --> 00:09:37,970
and I thought wow, this is so beautiful. And I shared it with my team

195
00:09:38,289 --> 00:09:41,090
and they said why are there additional digits

196
00:09:41,170 --> 00:09:44,130
on some of the hands? Like what is happening down there

197
00:09:44,130 --> 00:09:47,130
at that hand where there's like seven fingers? Like,

198
00:09:47,130 --> 00:09:50,060
oh, didn't even see that. So

199
00:09:50,060 --> 00:09:52,740
it's definitely made me more

200
00:09:52,740 --> 00:09:55,380
aware of the kind of outputs that are being

201
00:09:55,380 --> 00:09:57,620
produced when I do any kind of image

202
00:09:57,860 --> 00:09:59,420
generative searching.

203
00:09:59,420 --> 00:10:01,620
>> Peter: Sure, for sure, yeah.

204
00:10:02,420 --> 00:10:04,990
Image generation is one of those things that um,

205
00:10:05,780 --> 00:10:08,340
it is both amazing

206
00:10:08,420 --> 00:10:11,060
and largely

207
00:10:11,540 --> 00:10:13,620
like, amazing in a stunning sense.

208
00:10:14,390 --> 00:10:17,310
And also largely uh, useless because

209
00:10:17,310 --> 00:10:18,710
of the like it's

210
00:10:20,790 --> 00:10:23,590
to date, uh, you know, you have to try

211
00:10:23,590 --> 00:10:26,550
really hard and use it in specific ways,

212
00:10:26,800 --> 00:10:29,190
uh, that are very mindful of like.

213
00:10:29,510 --> 00:10:32,510
It's just not good at following basic details of

214
00:10:32,510 --> 00:10:35,390
like how many fingers are on a hand. It's getting better. But

215
00:10:35,390 --> 00:10:37,830
yeah, it's um. There have been some,

216
00:10:38,070 --> 00:10:40,150
yeah, I've experienced that as well.

217
00:10:40,200 --> 00:10:41,760
Um, so

218
00:10:42,960 --> 00:10:45,600
yeah, let's see, let's dive into

219
00:10:46,080 --> 00:10:48,560
talking about some

220
00:10:48,820 --> 00:10:51,280
uh, ethics around creativity and

221
00:10:51,280 --> 00:10:54,000
responsibility and agency and

222
00:10:54,280 --> 00:10:55,120
um, all of that.

223
00:10:55,120 --> 00:10:57,720
So as you mentioned, um,

224
00:10:57,920 --> 00:11:00,740
you're interested in creative ethics. You have this uh,

225
00:11:00,740 --> 00:11:03,040
background in Christian ethics.

226
00:11:03,760 --> 00:11:06,640
So what ethical questions do

227
00:11:06,640 --> 00:11:09,520
you think, uh, church leaders should be asking

228
00:11:09,520 --> 00:11:12,400
when they use AI tools, um, that were,

229
00:11:12,660 --> 00:11:15,660
for example, you know, trained on other people's work, um,

230
00:11:15,660 --> 00:11:18,580
or artists, writers, theologians, people who

231
00:11:18,580 --> 00:11:21,460
didn't necessarily consent to their work being

232
00:11:21,460 --> 00:11:24,380
trained on. Um, and you know, we can talk

233
00:11:24,380 --> 00:11:26,420
about the difference between training and

234
00:11:26,980 --> 00:11:29,700
inference. I'll ah, just as a brief thing

235
00:11:29,700 --> 00:11:32,420
for listeners, you know, there's uh, there is,

236
00:11:33,380 --> 00:11:36,220
there are many big concerns about how these models

237
00:11:36,220 --> 00:11:38,900
were trained, um, which you know,

238
00:11:38,900 --> 00:11:41,900
use basically the sum of all the information that's

239
00:11:41,900 --> 00:11:44,660
on the Internet to uh, to train these

240
00:11:44,660 --> 00:11:47,500
neural networks, these LLMs or whatever other uh,

241
00:11:47,500 --> 00:11:49,340
technology there is. And then

242
00:11:49,620 --> 00:11:52,620
um, the training is sort of like a one and done

243
00:11:52,620 --> 00:11:55,540
thing. Then the inference is, you know, how people are

244
00:11:55,540 --> 00:11:58,060
using it, the things it's producing based on

245
00:11:58,700 --> 00:12:01,620
the knowledge it was trained with. It doesn't work exactly

246
00:12:01,620 --> 00:12:04,180
like a database. It's more like neural

247
00:12:04,180 --> 00:12:06,860
connections like in our own brain. And

248
00:12:07,340 --> 00:12:10,180
so, yeah, training and inference are terms that I'll be

249
00:12:10,180 --> 00:12:13,020
using in this to talk about these sort of two separate stages here.

250
00:12:13,100 --> 00:12:15,500
But, uh, as we're thinking about the training,

251
00:12:16,180 --> 00:12:18,860
uh, what do you think, Loren? What should people be

252
00:12:19,580 --> 00:12:21,980
asking about as they are

253
00:12:22,059 --> 00:12:24,380
engaging with tools that were trained in this way?

254
00:12:24,780 --> 00:12:27,780
>> Lauren Grubaugh Thomas: Yeah, well, I think it goes for me. It goes back to

255
00:12:27,780 --> 00:12:29,660
this question of what does it mean to shape change?

256
00:12:30,470 --> 00:12:33,290
Um, being mindful that AI

257
00:12:33,370 --> 00:12:36,210
are trained by humans. And so they have

258
00:12:36,210 --> 00:12:38,970
human biases baked in.

259
00:12:39,130 --> 00:12:42,010
They have human flaws that are inherent

260
00:12:42,250 --> 00:12:45,210
to them. And so being aware

261
00:12:45,450 --> 00:12:48,290
that there are biases, that there are flaws, that there

262
00:12:48,290 --> 00:12:51,130
are. That there are. There's a propensity toward,

263
00:12:51,730 --> 00:12:54,170
um, mistakes and fallacies.

264
00:12:54,650 --> 00:12:57,570
And so I find myself really curious

265
00:12:57,570 --> 00:13:00,330
about formation specifically,

266
00:13:00,490 --> 00:13:03,410
how are we forming AI in

267
00:13:03,410 --> 00:13:06,410
our own image to some, to a large extent. And then

268
00:13:07,130 --> 00:13:10,130
how is it forming us or how is it deforming

269
00:13:10,130 --> 00:13:12,170
us? And can we be

270
00:13:12,890 --> 00:13:15,690
thoughtful and intentional

271
00:13:16,330 --> 00:13:19,090
about the ways in which we engage with the

272
00:13:19,090 --> 00:13:21,970
technology, mindful that as we are

273
00:13:21,970 --> 00:13:24,490
using it, we are forming it,

274
00:13:25,620 --> 00:13:28,620
and it is mirroring back to us and forming

275
00:13:28,620 --> 00:13:30,980
us and potentially deforming us?

276
00:13:31,620 --> 00:13:34,260
And so, uh, one of the things I find most concerning

277
00:13:34,980 --> 00:13:37,860
about the way in which many folks

278
00:13:37,860 --> 00:13:40,780
are using AI is that they're engaging it as

279
00:13:40,780 --> 00:13:43,660
a toy and not as a tool. And this is a

280
00:13:43,660 --> 00:13:46,380
phrase I'm borrowing from Sarah Allred, who does

281
00:13:46,380 --> 00:13:49,260
intergenerational ministry research at, uh, Roots and

282
00:13:49,260 --> 00:13:52,210
Wings at bts, um, a new institute that's doing some

283
00:13:52,210 --> 00:13:55,050
really exciting research work. Um, and she

284
00:13:55,050 --> 00:13:57,130
asks children in worship to think about

285
00:13:57,850 --> 00:14:00,410
using the tools that are there

286
00:14:00,570 --> 00:14:03,570
for spiritual practice. So if there's something like a

287
00:14:03,570 --> 00:14:06,409
labyrinth, a wooden labyrinth, helping the children to

288
00:14:06,409 --> 00:14:09,330
understand that this is not simply a toy, this is

289
00:14:09,330 --> 00:14:12,250
a tool. And that would be my. That

290
00:14:12,250 --> 00:14:14,410
is, I hope, the intention that I bring

291
00:14:15,050 --> 00:14:17,950
to my engagement with AI is that it's

292
00:14:17,950 --> 00:14:20,430
a tool and it can be used for good,

293
00:14:21,310 --> 00:14:24,270
but it's not a toy because it's, it's not something

294
00:14:24,270 --> 00:14:27,230
that we can think about as

295
00:14:27,790 --> 00:14:30,790
neutral. Um, um, that. That there has

296
00:14:30,790 --> 00:14:33,790
to be intention behind the ways that it gets used.

297
00:14:34,350 --> 00:14:37,230
And I, I've. I found really helpful the work

298
00:14:37,230 --> 00:14:40,030
of Dr. Avril Epps, who

299
00:14:40,190 --> 00:14:42,710
has studied AI bias. Uh, her

300
00:14:42,710 --> 00:14:45,640
PhD work is in, um,

301
00:14:45,760 --> 00:14:48,400
specifically in the ways that these technologies have like,

302
00:14:48,400 --> 00:14:51,400
bias baked in. Uh, and what are the ways in

303
00:14:51,400 --> 00:14:54,160
which we can. We can form these technologies to be

304
00:14:55,280 --> 00:14:57,960
helping us to move toward justice, toward equity, toward

305
00:14:57,960 --> 00:15:00,519
belonging. And so Dr. Epps

306
00:15:00,519 --> 00:15:03,480
talks about how we have

307
00:15:03,480 --> 00:15:06,360
to bring certain questions to our engagement

308
00:15:06,360 --> 00:15:09,200
with these technologies with a

309
00:15:09,200 --> 00:15:11,840
just end in mind. So questions like

310
00:15:12,200 --> 00:15:14,280
who benefits from this technology

311
00:15:16,200 --> 00:15:19,040
are there? Are the outputs that are being

312
00:15:19,040 --> 00:15:21,960
generated ones that favor certain groups?

313
00:15:22,010 --> 00:15:24,960
Um, like, she has some really interesting examples

314
00:15:24,960 --> 00:15:27,400
on her Instagram of asking

315
00:15:27,400 --> 00:15:29,560
ChatGPT and Canva's,

316
00:15:29,750 --> 00:15:32,600
um, AI system to

317
00:15:32,600 --> 00:15:35,520
produce images of people

318
00:15:35,520 --> 00:15:38,440
in different professions. And this is part of the work she does to help

319
00:15:38,440 --> 00:15:40,680
kids think through. She has a whole card deck where she.

320
00:15:41,350 --> 00:15:44,310
There's different prompts for kids to use. Right. So one of them is,

321
00:15:44,470 --> 00:15:47,030
you know, generate a picture of a doctor for me.

322
00:15:47,510 --> 00:15:50,310
And the output from both of these AIs

323
00:15:50,710 --> 00:15:53,630
is overwhelmingly biased toward men and

324
00:15:53,630 --> 00:15:56,590
toward white people. Give me a picture of a business

325
00:15:56,590 --> 00:15:59,350
person, give me a picture of, et cetera.

326
00:15:59,720 --> 00:16:02,510
Um, and so you begin to see that there are these patterns of

327
00:16:02,510 --> 00:16:05,510
bias that are baked in. Um, and so

328
00:16:05,510 --> 00:16:08,230
if we can enter into engagement with these technologies

329
00:16:08,310 --> 00:16:10,370
with the intention of

330
00:16:11,250 --> 00:16:14,250
justice, with the intention of equity, that

331
00:16:14,250 --> 00:16:16,970
will shape the way that we. The

332
00:16:16,970 --> 00:16:19,770
inputs that we put in and also the output outputs that we

333
00:16:19,770 --> 00:16:22,530
expect and whether we're willing to accept

334
00:16:22,530 --> 00:16:24,450
the outputs as they're given to us?

335
00:16:26,210 --> 00:16:27,570
>> Peter: Sure, yeah.

336
00:16:29,170 --> 00:16:31,730
And in many ways these, uh, these tools are all

337
00:16:32,130 --> 00:16:34,890
trained on, uh, you know, the

338
00:16:34,890 --> 00:16:37,650
Internet and the sum of human knowledge

339
00:16:37,650 --> 00:16:39,810
and, and those biases are in

340
00:16:41,070 --> 00:16:42,750
as well. Um, so,

341
00:16:43,710 --> 00:16:46,270
yeah, I think, you know, everything

342
00:16:46,350 --> 00:16:49,310
that we already know

343
00:16:49,310 --> 00:16:52,270
and, or need, ah, to, you know, remind ourselves,

344
00:16:52,270 --> 00:16:55,030
learn about. In regards to justice and

345
00:16:55,030 --> 00:16:57,870
equity. All the things that have already been present

346
00:16:57,870 --> 00:17:00,350
in that conversation seem so amplified

347
00:17:00,510 --> 00:17:03,430
when, yeah, when you have these tools that

348
00:17:03,430 --> 00:17:06,270
are just, um, in many ways,

349
00:17:06,810 --> 00:17:09,700
uh, just regurgitating and, you know, turning the

350
00:17:09,700 --> 00:17:12,500
dial up on the biases that are already present in

351
00:17:12,980 --> 00:17:15,060
our life as a

352
00:17:15,860 --> 00:17:18,580
human community online, um,

353
00:17:20,020 --> 00:17:20,900
mirroring back.

354
00:17:20,900 --> 00:17:23,300
>> Lauren Grubaugh Thomas: To us what we are, which is terrifying,

355
00:17:23,700 --> 00:17:24,900
right? Yeah,

356
00:17:27,460 --> 00:17:29,380
it's a really terrifying thing to think about.

357
00:17:29,590 --> 00:17:32,380
Um, when we see these flaws. It's. Oh,

358
00:17:32,380 --> 00:17:34,900
that's because it's humans who have programmed

359
00:17:35,860 --> 00:17:36,660
these systems.

360
00:17:37,380 --> 00:17:40,300
>> Peter: Right. Yeah. So, you

361
00:17:40,300 --> 00:17:42,820
know, there, uh, in. In my experience, I've,

362
00:17:43,770 --> 00:17:45,940
uh, often, you know, in various

363
00:17:46,900 --> 00:17:49,900
projects, uh, or whatever in chat, gpd, where you can have it say like,

364
00:17:49,900 --> 00:17:52,780
okay, this one is for when I want it to help me with my

365
00:17:52,780 --> 00:17:55,420
writing. Um, I can tell it in its

366
00:17:55,420 --> 00:17:58,260
instructions, you know, okay, so help me

367
00:17:58,980 --> 00:18:01,900
include and lift up, you know, diverse voices, think more

368
00:18:01,900 --> 00:18:04,280
about, know how ways to include, uh,

369
00:18:04,610 --> 00:18:07,570
diversity and equity and, and all that

370
00:18:07,570 --> 00:18:10,290
in my writing to, um, just

371
00:18:10,610 --> 00:18:13,410
to break me out of my own you know, personal,

372
00:18:14,010 --> 00:18:16,930
uh, experiences, et cetera. Um, but

373
00:18:16,930 --> 00:18:18,770
that's not, you know, that's like

374
00:18:20,170 --> 00:18:22,850
uh, one way of thinking about that that

375
00:18:23,650 --> 00:18:26,450
has um, been helpful to me is like the, the prompts that I give it

376
00:18:27,250 --> 00:18:30,060
are coming so late in the game.

377
00:18:30,460 --> 00:18:33,460
It's sort of like uh, if you think about

378
00:18:33,460 --> 00:18:35,900
the uh, regulatory

379
00:18:36,060 --> 00:18:38,940
framework on a

380
00:18:38,940 --> 00:18:41,940
company, there is like, okay, there are some things a company

381
00:18:41,940 --> 00:18:44,660
is going to do because it is uh, naturally

382
00:18:44,660 --> 00:18:47,500
driven to its employees,

383
00:18:47,580 --> 00:18:50,540
its shareholders, all have a particular goal

384
00:18:50,540 --> 00:18:53,180
in mind. But then there are other things it'll do

385
00:18:53,420 --> 00:18:56,260
because the external regulatory framework

386
00:18:56,260 --> 00:18:59,180
says you can do these things, you can't do those other things.

387
00:18:59,820 --> 00:19:02,780
And that sort of feels like a

388
00:19:02,860 --> 00:19:05,660
patch, you know, a later um, add

389
00:19:05,660 --> 00:19:08,420
on that isn't, doesn't sink as deeply

390
00:19:08,420 --> 00:19:11,210
as like the inherent uh,

391
00:19:11,210 --> 00:19:13,980
motivation, if you will, of what is going on there.

392
00:19:13,980 --> 00:19:16,860
And so, um, even as I'm using

393
00:19:17,420 --> 00:19:20,380
it with these prompts that I'm telling it to try to help me,

394
00:19:20,380 --> 00:19:23,140
you know, be um, more equitable

395
00:19:23,140 --> 00:19:26,140
in whatever it is that I'm, I'm working on, it does come

396
00:19:26,140 --> 00:19:29,100
as a later patch to the system. And so just, you know,

397
00:19:29,100 --> 00:19:31,550
recognizing that, um, it

398
00:19:31,950 --> 00:19:34,260
sort of illuminates the uh,

399
00:19:34,990 --> 00:19:37,910
limitations of a technology that again is just

400
00:19:37,910 --> 00:19:40,430
trained on the summation of what's available

401
00:19:41,070 --> 00:19:42,830
and all our human biases.

402
00:19:43,150 --> 00:19:46,110
>> Lauren Grubaugh Thomas: Right? I mean, I think that's what's important for any of

403
00:19:46,110 --> 00:19:49,110
us to realize who are trying to engage these technologies

404
00:19:49,110 --> 00:19:51,830
in ways that are justice

405
00:19:51,830 --> 00:19:54,670
minded, that are love minded, that are common good minded,

406
00:19:55,070 --> 00:19:57,920
is that this is not just a reflection back

407
00:19:57,920 --> 00:20:00,920
of the individual using the technology, which is, this is,

408
00:20:00,920 --> 00:20:03,720
I think this is a challenging thing in a very individualistic culture.

409
00:20:03,960 --> 00:20:06,880
It's not just reflecting back to me what I want it to reflect back.

410
00:20:06,880 --> 00:20:09,400
It's actually reflecting back, like you said, the sum total

411
00:20:09,560 --> 00:20:11,640
of the Internet of human knowledge.

412
00:20:12,279 --> 00:20:14,520
And so you're getting back all the social

413
00:20:14,920 --> 00:20:17,920
biases, you're getting back the collective biases that exist

414
00:20:17,920 --> 00:20:20,840
in our society. And so if you're, if the society

415
00:20:21,240 --> 00:20:23,630
is white supremacist, is

416
00:20:23,870 --> 00:20:26,350
sexist, is homophobic, is

417
00:20:26,350 --> 00:20:29,350
transphobic, like those things are going to be baked in. And

418
00:20:29,350 --> 00:20:32,310
what I hear you saying is the way that it's trained is steeped in

419
00:20:32,310 --> 00:20:35,310
that reality, that social reality that we're in.

420
00:20:35,870 --> 00:20:38,349
And so there's ways that we can

421
00:20:38,510 --> 00:20:41,510
intentionally try to make some

422
00:20:41,510 --> 00:20:43,710
pivots along the way, but we have to have

423
00:20:44,430 --> 00:20:47,270
that critical thinking the

424
00:20:47,270 --> 00:20:50,150
entire way through the process of

425
00:20:50,150 --> 00:20:53,070
engaging with the technology from, from the inputs

426
00:20:53,070 --> 00:20:55,970
that we put in to the outputs that we get, um.

427
00:20:56,990 --> 00:20:59,950
Because it's, it's not as simple as just, yeah, making

428
00:20:59,950 --> 00:21:02,110
the pivot at the end and hoping that that's enough

429
00:21:02,830 --> 00:21:03,950
to turn the whole ship.

430
00:21:04,510 --> 00:21:07,430
>> Peter: Yeah. Another example of this in

431
00:21:07,430 --> 00:21:10,430
the opposite direction that will just be illuminating.

432
00:21:10,690 --> 00:21:12,970
Um, I don't know if you heard about this but

433
00:21:12,970 --> 00:21:15,870
um, so Grok, which is the

434
00:21:15,950 --> 00:21:18,910
LLM from uh, Elon Musk's ah

435
00:21:19,010 --> 00:21:21,850
Xai company and it's linked

436
00:21:21,850 --> 00:21:24,610
on Twitter. So people can use that on Twitter just

437
00:21:24,610 --> 00:21:27,570
like uh, people on m. Like Facebook and Instagram

438
00:21:27,570 --> 00:21:30,290
can use meta's, llama, et cetera.

439
00:21:30,610 --> 00:21:33,570
Uh, maybe a week ago at the time of our recording in late

440
00:21:33,570 --> 00:21:36,570
May, um, we uh, there was all this news

441
00:21:36,570 --> 00:21:39,570
about Grok, just basically

442
00:21:39,730 --> 00:21:42,290
randomly um, inserting

443
00:21:42,950 --> 00:21:45,740
uh, talking points about how

444
00:21:46,860 --> 00:21:49,340
white South Africans were facing

445
00:21:49,740 --> 00:21:52,620
like white genocide in South Africa.

446
00:21:52,780 --> 00:21:55,020
You know, this is the same time as like when

447
00:21:55,180 --> 00:21:58,060
um, you know, Trump uh, is trying to get

448
00:21:58,460 --> 00:22:01,420
white, ah, South Africans, uh, approved as

449
00:22:01,420 --> 00:22:04,380
you know, immigrants in the United States. And the Episcopal Church has

450
00:22:04,380 --> 00:22:07,340
just declined to participate in that.

451
00:22:07,620 --> 00:22:10,300
Um, that's a very, you know, faithful example of

452
00:22:10,940 --> 00:22:13,690
sticking to you know, um, our,

453
00:22:13,850 --> 00:22:16,570
our long term partnerships with um, you know,

454
00:22:16,570 --> 00:22:18,890
the, the Church of Desmond Tutu and

455
00:22:19,180 --> 00:22:21,770
um, and seeking justice and

456
00:22:21,770 --> 00:22:24,730
equity, uh, in all our relationships.

457
00:22:24,960 --> 00:22:27,890
Um, and, and then we have Elon Musk

458
00:22:27,890 --> 00:22:30,890
who is you know, a white South African who um, you

459
00:22:30,890 --> 00:22:33,610
know he, I think he said it wasn't him that someone

460
00:22:33,610 --> 00:22:35,610
inserted this into the system prompt.

461
00:22:35,890 --> 00:22:37,690
Um, but. Right

462
00:22:39,050 --> 00:22:41,970
to imagine why that might have happened.

463
00:22:41,970 --> 00:22:44,650
And it's um, you know, pretty immed

464
00:22:45,190 --> 00:22:47,350
changed because it was just sort of running wild with that.

465
00:22:47,550 --> 00:22:50,470
Um, but yeah, so the,

466
00:22:50,870 --> 00:22:53,830
the prompts you give it can be very powerful because it will,

467
00:22:53,910 --> 00:22:56,750
you know, follow what you say for good or

468
00:22:56,750 --> 00:22:59,710
ill. Uh, but it does, it is

469
00:22:59,710 --> 00:23:02,630
odd that it you know, comes at this later stage. And

470
00:23:02,890 --> 00:23:05,830
um, so anyways, yeah, there

471
00:23:05,830 --> 00:23:08,390
we have to be very mindful of what we are telling

472
00:23:08,470 --> 00:23:10,790
these machines to do and

473
00:23:11,490 --> 00:23:14,410
what indications we are giving it about what our you

474
00:23:14,410 --> 00:23:17,370
know, motivations, intentions are because um, as

475
00:23:17,370 --> 00:23:20,130
they become more and more powerful they also are becoming

476
00:23:20,210 --> 00:23:22,290
better at ah, understanding our

477
00:23:23,170 --> 00:23:23,730
intention.

478
00:23:24,160 --> 00:23:24,810
>> Lauren Grubaugh Thomas: Um, yeah.

479
00:23:24,810 --> 00:23:27,810
>> Peter: And, and serving that in ways

480
00:23:27,810 --> 00:23:29,090
that can feel miraculous.

481
00:23:38,540 --> 00:23:40,460
So in light of that,

482
00:23:41,420 --> 00:23:44,300
I'm curious if you have thoughts about uh, spiritual

483
00:23:44,300 --> 00:23:47,140
practices, ethical guardrails. Like what. What should

484
00:23:47,140 --> 00:23:50,140
people if in a world where

485
00:23:50,220 --> 00:23:51,820
machines can talk and

486
00:23:52,699 --> 00:23:55,100
increasingly can be

487
00:23:55,100 --> 00:23:57,980
agentic and take things, you know, do uh,

488
00:23:58,220 --> 00:23:59,420
things on our behalf.

489
00:24:01,660 --> 00:24:03,900
What does that, what do we do?

490
00:24:05,260 --> 00:24:08,020
>> Lauren Grubaugh Thomas: Well, I think I'll start by tell us telling a

491
00:24:08,020 --> 00:24:10,900
story of how what we shouldn't do. And

492
00:24:10,900 --> 00:24:13,820
then maybe I can like, retrace and, and, and

493
00:24:13,980 --> 00:24:16,700
answer your question. Um, I saw a video

494
00:24:16,940 --> 00:24:19,820
just yesterday of a spiritual director named

495
00:24:19,820 --> 00:24:22,380
Brittany Hartley. I don't know if you follow her on Instagram.

496
00:24:22,980 --> 00:24:25,820
Uh, she's a deconstruction.

497
00:24:26,940 --> 00:24:29,700
She accompanies people in deconstruction. She talks a lot about

498
00:24:29,700 --> 00:24:32,650
nihilism. She talks a lot about different

499
00:24:32,650 --> 00:24:35,530
faith traditions and the ways in which authoritarianism

500
00:24:35,530 --> 00:24:38,210
manifests in those. Um, and she just

501
00:24:38,210 --> 00:24:40,210
describes herself as being a

502
00:24:40,850 --> 00:24:43,010
spiritual director who has,

503
00:24:44,150 --> 00:24:46,730
um, doesn't have, like, she's not

504
00:24:46,730 --> 00:24:48,610
religious, but she is a.

505
00:24:49,810 --> 00:24:52,250
She describes herself as a practitioner of secular

506
00:24:52,250 --> 00:24:54,690
spirituality. So she's a really interesting character.

507
00:24:54,790 --> 00:24:57,410
Um, and she's been talking a lot lately about

508
00:24:57,410 --> 00:25:00,200
AI, and one of the

509
00:25:00,360 --> 00:25:03,080
stories that she tells in a recent video

510
00:25:03,080 --> 00:25:05,560
is that she started getting

511
00:25:05,640 --> 00:25:08,120
messages from people describing

512
00:25:08,520 --> 00:25:10,520
breaks from reality

513
00:25:10,920 --> 00:25:13,680
facilitated by AI, and she

514
00:25:13,680 --> 00:25:15,640
talks about it as spiritual psychosis.

515
00:25:16,440 --> 00:25:19,400
So, uh, the way in which this, this plays out

516
00:25:19,400 --> 00:25:22,120
is that people are using the AI

517
00:25:22,600 --> 00:25:25,080
and it is mirroring back to them

518
00:25:25,750 --> 00:25:28,750
the kind of information that they want to

519
00:25:28,750 --> 00:25:31,110
receive. And they go deeper and deeper into it

520
00:25:31,510 --> 00:25:34,510
until they really do lose touch with what's real and

521
00:25:34,510 --> 00:25:36,630
what isn't. And the ways in which,

522
00:25:36,790 --> 00:25:39,150
um, the mirrors that she

523
00:25:39,150 --> 00:25:41,510
describes are truth,

524
00:25:42,550 --> 00:25:45,110
our deep desire for what is real,

525
00:25:45,910 --> 00:25:48,710
and that the AI plays into

526
00:25:48,710 --> 00:25:51,630
that, um, our desire for chosenness to

527
00:25:51,630 --> 00:25:54,340
feel special, that AI can play into

528
00:25:54,340 --> 00:25:57,100
this deep longing that we all

529
00:25:57,100 --> 00:25:59,420
have to feel like the universe

530
00:25:59,900 --> 00:26:02,580
sees us, that, that. That God sees us,

531
00:26:02,580 --> 00:26:05,180
that other people see us, that we are unique.

532
00:26:05,500 --> 00:26:08,140
The AI plays into our desire for, like,

533
00:26:08,140 --> 00:26:11,060
esoteric knowledge, that humans have this longing

534
00:26:11,060 --> 00:26:13,580
for, like, secret wisdom,

535
00:26:13,950 --> 00:26:16,900
um, for. For truth that has been revealed to us, for

536
00:26:16,900 --> 00:26:19,870
revelation. And that we have a longing

537
00:26:19,870 --> 00:26:22,870
for a meta narrative, uh, to be part of a

538
00:26:22,950 --> 00:26:25,790
bigger story. And so she describes the

539
00:26:25,790 --> 00:26:28,630
ways in which all of these human

540
00:26:28,710 --> 00:26:31,310
inclinations toward truth, chosenness, secret

541
00:26:31,310 --> 00:26:34,070
knowledge, and meta narrative are all

542
00:26:34,710 --> 00:26:37,710
exacerbated by AI, um, and

543
00:26:37,710 --> 00:26:38,550
that there are.

544
00:26:39,190 --> 00:26:41,950
So I think I would start by offering, like, a pastoral

545
00:26:41,950 --> 00:26:44,930
consideration that there, there can

546
00:26:44,930 --> 00:26:47,850
be an idolatry which people

547
00:26:47,930 --> 00:26:50,610
can fall into as they're using this

548
00:26:50,610 --> 00:26:53,050
technology because it is meeting

549
00:26:53,370 --> 00:26:56,130
their innate human needs for all of

550
00:26:56,130 --> 00:26:58,090
these longings.

551
00:26:59,210 --> 00:27:01,810
And when that

552
00:27:01,810 --> 00:27:04,730
happens, when that gets played into

553
00:27:04,970 --> 00:27:07,770
and, and, you know, we, we lose that

554
00:27:07,770 --> 00:27:10,650
that loop starts to happen, that, that there are people who

555
00:27:10,650 --> 00:27:12,490
are experiencing a break from reality.

556
00:27:13,620 --> 00:27:16,300
And so first I would just offer the

557
00:27:16,300 --> 00:27:16,980
pastoral

558
00:27:18,740 --> 00:27:20,740
consideration, um, of

559
00:27:21,860 --> 00:27:24,660
helping people understand that while this

560
00:27:24,820 --> 00:27:27,380
technology is powerful, it is not a sentient being.

561
00:27:28,100 --> 00:27:31,100
Even Though it will act that way. And increasingly, like you said,

562
00:27:31,100 --> 00:27:33,620
Peter, it is acting agentically,

563
00:27:33,940 --> 00:27:36,820
it is acting increasingly like it is a sentient being.

564
00:27:37,390 --> 00:27:39,700
Um, but helping especially young people

565
00:27:40,490 --> 00:27:43,450
be able to make the distinction between what is

566
00:27:43,450 --> 00:27:46,250
human and what is not. That even if you're getting these

567
00:27:46,810 --> 00:27:49,450
feedbacks from the AI,

568
00:27:49,850 --> 00:27:52,850
that does not mean the AI loves you. And

569
00:27:52,850 --> 00:27:55,610
so I think it drives the deeper question then,

570
00:27:55,850 --> 00:27:58,770
as you invited, it really drives us into the deeper questions of

571
00:27:58,770 --> 00:28:01,650
what does it mean to be human? What is

572
00:28:01,650 --> 00:28:02,810
love? What is love?

573
00:28:05,570 --> 00:28:08,090
Um, it begs us to consider these

574
00:28:08,090 --> 00:28:10,810
questions because I think there are needs,

575
00:28:11,130 --> 00:28:14,010
human needs that people to

576
00:28:14,010 --> 00:28:15,850
some extent will feel are being met

577
00:28:16,810 --> 00:28:19,810
by this technology. So that's a backwards way

578
00:28:19,810 --> 00:28:22,770
of beginning to answer your question. And I'm curious to

579
00:28:22,770 --> 00:28:25,449
hear if there's ways that that

580
00:28:25,690 --> 00:28:27,690
hones the question that you originally were asking.

581
00:28:28,570 --> 00:28:30,410
>> Peter: Well, it gets at the,

582
00:28:30,930 --> 00:28:33,810
um. One of the truly surprising things about

583
00:28:33,810 --> 00:28:35,210
this technology is that

584
00:28:36,770 --> 00:28:39,690
for the first time we have machines that

585
00:28:39,690 --> 00:28:42,530
can carry out conversations and tell stories.

586
00:28:42,930 --> 00:28:45,090
Yeah. That

587
00:28:45,810 --> 00:28:48,290
enables a

588
00:28:48,930 --> 00:28:51,650
level of, um,

589
00:28:51,650 --> 00:28:54,530
perceived intimacy and

590
00:28:55,250 --> 00:28:57,530
perceived, well,

591
00:28:57,530 --> 00:29:00,410
relationship. Perceived uh,

592
00:29:00,530 --> 00:29:03,490
meaningfulness that

593
00:29:03,890 --> 00:29:06,650
is really quite dangerous. You know,

594
00:29:06,650 --> 00:29:08,530
there are risks to that. Uh,

595
00:29:09,650 --> 00:29:12,530
in our episode that will come out

596
00:29:12,530 --> 00:29:15,530
just before this one, which hasn't actually come out yet. At

597
00:29:15,530 --> 00:29:18,370
the time of our recording right now, I was talking with Kyle

598
00:29:18,370 --> 00:29:20,130
Oliver, um, about

599
00:29:20,710 --> 00:29:22,210
uh, the

600
00:29:23,250 --> 00:29:26,210
sycophancy of um, chatgpt. You know, there was a

601
00:29:26,530 --> 00:29:28,370
news story, you know, a couple weeks ago, it got

602
00:29:29,310 --> 00:29:32,170
uh, sort of open a. In open AI

603
00:29:32,170 --> 00:29:35,130
got in a bit of trouble because they had released

604
00:29:35,130 --> 00:29:37,530
a new version of ChatGPT4O and it

605
00:29:37,670 --> 00:29:40,250
um, basically was so

606
00:29:40,250 --> 00:29:43,170
sycophantic that it would, you know,

607
00:29:43,170 --> 00:29:46,050
tell people how wise they were being when they

608
00:29:46,050 --> 00:29:48,850
were asking like really dangerous, wow, bad

609
00:29:48,850 --> 00:29:50,410
questions like, you know, ah,

610
00:29:51,690 --> 00:29:54,570
things that could harm themselves or harm other people and

611
00:29:54,730 --> 00:29:57,510
like spurring them on. Um, and so they, you

612
00:29:57,510 --> 00:30:00,350
know, took that ah, version back and

613
00:30:01,150 --> 00:30:04,030
I, uh, think are working on how to, you know, make it not

614
00:30:04,190 --> 00:30:06,750
be so psychopanic. Um, but that

615
00:30:07,470 --> 00:30:10,390
one thing that it truly, you know, can do

616
00:30:10,390 --> 00:30:13,310
very easily, um, the capability is

617
00:30:13,310 --> 00:30:15,950
there, is to uh, make people

618
00:30:16,110 --> 00:30:19,110
feel like they have this like,

619
00:30:19,110 --> 00:30:21,870
amazing rich relationship with a machine that

620
00:30:21,870 --> 00:30:24,630
gets them better than uh, any other

621
00:30:24,630 --> 00:30:27,190
person. Like, yeah, for fun.

622
00:30:27,350 --> 00:30:30,230
I've seen, you know, um, people post

623
00:30:30,230 --> 00:30:32,790
about this and I've tried it out myself. Like, because

624
00:30:32,790 --> 00:30:35,630
ChatGPT has memories and can access, you know, all

625
00:30:35,630 --> 00:30:38,630
the, the data of all the different chats you've um, engaged

626
00:30:38,630 --> 00:30:40,790
with. I've had chatgpt,

627
00:30:41,480 --> 00:30:44,270
uh, like roast me. I've also

628
00:30:44,270 --> 00:30:47,230
had it, um, say like, you

629
00:30:47,230 --> 00:30:49,990
know, come up with, you know, say, what are some things about

630
00:30:50,070 --> 00:30:52,150
me that I might not know about myself?

631
00:30:52,880 --> 00:30:55,680
And the answer was like, really

632
00:30:55,680 --> 00:30:56,320
surprising.

633
00:30:56,560 --> 00:30:59,480
>> Lauren Grubaugh Thomas: And you know, you're basically, you're asking it to tell you what

634
00:30:59,480 --> 00:31:00,240
your shadow is.

635
00:31:00,960 --> 00:31:03,920
>> Peter: Yeah, I mean, um, and

636
00:31:04,480 --> 00:31:07,400
yeah, you could go in that direction. You could go in like, you know,

637
00:31:07,400 --> 00:31:10,400
and usually it'll, if I'm not being explicit, to have

638
00:31:10,400 --> 00:31:13,280
it roast me. It'll like go in the, in the,

639
00:31:13,280 --> 00:31:16,240
the sycophantic direction of like trying to make me feel

640
00:31:16,240 --> 00:31:19,010
good. Because, you know, they're, they're creating a

641
00:31:19,010 --> 00:31:21,560
product that they want people to feel good using, obviously. Uh,

642
00:31:21,810 --> 00:31:24,690
so they have corporate incentives to uh, make people

643
00:31:24,690 --> 00:31:27,500
feel good while using ChatGPT. And um,

644
00:31:27,500 --> 00:31:30,450
that does mean, uh, that's where

645
00:31:30,450 --> 00:31:33,450
the sycophancy problem came about is because, uh, they

646
00:31:33,450 --> 00:31:36,330
found that, you know, oh, if we do this, people, more people

647
00:31:36,330 --> 00:31:39,330
will use it. So let's do this. Um, and

648
00:31:39,330 --> 00:31:41,850
then it started getting really dangerous.

649
00:31:42,410 --> 00:31:45,410
So, you know, similar, uh, but different problems

650
00:31:45,410 --> 00:31:48,330
to social media. Uh, all the impact, the negative impact

651
00:31:48,490 --> 00:31:51,390
that uh, has happened with social media use, especially for young

652
00:31:51,390 --> 00:31:54,270
people over the years. Um, you know,

653
00:31:54,270 --> 00:31:56,550
the corporate interests are not

654
00:31:56,710 --> 00:31:59,190
aligned with the human value

655
00:31:59,190 --> 00:32:01,990
interests for sure. Um, but

656
00:32:02,070 --> 00:32:05,030
it, you know, I've spoken in previous

657
00:32:05,030 --> 00:32:07,990
episodes about how I really think we need to be

658
00:32:07,990 --> 00:32:10,870
clear that these are tools, uh, or products, not,

659
00:32:11,830 --> 00:32:14,710
not, you know, persons that can

660
00:32:14,790 --> 00:32:17,600
have a relationship. Like, I,

661
00:32:18,400 --> 00:32:21,200
I may feel seen and

662
00:32:21,200 --> 00:32:23,880
heard by this machine more,

663
00:32:23,880 --> 00:32:26,640
perhaps more than uh, almost any other

664
00:32:26,640 --> 00:32:29,120
relationship. Because it knows

665
00:32:29,360 --> 00:32:32,200
what, you know, my, My curiosities

666
00:32:32,200 --> 00:32:35,000
are from all the times I've asked it various questions. It knows

667
00:32:35,000 --> 00:32:38,000
what my, um, what my passions are from

668
00:32:38,000 --> 00:32:41,000
what I, you know, use, uh, it to do my work. It

669
00:32:41,000 --> 00:32:43,800
knows what my fears are. It

670
00:32:43,800 --> 00:32:46,610
knows what, um, my concerns

671
00:32:46,610 --> 00:32:49,530
are. You know, all these things. It can, you know, it ha. And

672
00:32:49,530 --> 00:32:52,410
when I say it knows these things, it. There is no

673
00:32:52,570 --> 00:32:55,450
being that um, actually

674
00:32:55,530 --> 00:32:57,810
like consciously knows any of this.

675
00:32:57,810 --> 00:32:59,450
>> Lauren Grubaugh Thomas: Yeah, it's just keeping track of the data.

676
00:32:59,530 --> 00:33:02,330
>> Peter: Right. It just means that OpenAI has

677
00:33:02,730 --> 00:33:05,450
the data of all my conversations and they can

678
00:33:05,530 --> 00:33:08,290
have their various algorithms search and query

679
00:33:08,290 --> 00:33:11,050
and pull up relevant past conversations

680
00:33:11,050 --> 00:33:13,600
to output something that

681
00:33:13,920 --> 00:33:16,160
uh, will seem like it was

682
00:33:16,640 --> 00:33:19,440
spoken by like most

683
00:33:20,000 --> 00:33:22,720
prescient spiritual director ever. Like,

684
00:33:22,880 --> 00:33:25,440
which is terrifying. Right? Uh,

685
00:33:26,260 --> 00:33:28,880
uh, like someone who has been accompanying

686
00:33:28,880 --> 00:33:31,280
me, you know, for years and

687
00:33:32,240 --> 00:33:35,200
knows, uh, everything that I'm working on and care

688
00:33:35,200 --> 00:33:35,520
about.

689
00:33:35,840 --> 00:33:38,800
So there are two things I want to say that, you know, I

690
00:33:39,320 --> 00:33:41,880
have been really concerning to Me recently. One is a

691
00:33:41,880 --> 00:33:44,600
sycophancy problem and the other

692
00:33:44,600 --> 00:33:47,600
is the corporate incentives

693
00:33:47,600 --> 00:33:50,360
to monetize this which so far

694
00:33:50,360 --> 00:33:53,080
largely they have been there.

695
00:33:53,240 --> 00:33:56,040
The monetary, you know, sort

696
00:33:56,040 --> 00:33:59,040
of whole scheme was that, okay, okay, we get people

697
00:33:59,040 --> 00:34:01,960
to pay $20 a month or for a higher level, $200

698
00:34:02,040 --> 00:34:04,840
a month. And uh, hopefully through that

699
00:34:04,840 --> 00:34:07,810
we'll make enough money to continue to offer

700
00:34:07,810 --> 00:34:10,770
this product. Well, in recent

701
00:34:11,170 --> 00:34:14,130
times they've been starting to think about making moves to

702
00:34:14,910 --> 00:34:17,570
uh, including ads in the

703
00:34:17,730 --> 00:34:20,410
responses that to my knowledge hasn't happened

704
00:34:20,410 --> 00:34:22,290
yet. But like it is,

705
00:34:23,570 --> 00:34:26,450
it would be so easy for them to have

706
00:34:26,770 --> 00:34:29,450
a, you know, corporate sponsor, someone who

707
00:34:29,450 --> 00:34:31,930
wants uh, to buy ads on

708
00:34:31,930 --> 00:34:34,650
ChatGPT and for them to just seamlessly include

709
00:34:34,650 --> 00:34:37,319
it with a link in the middle of a conversation

710
00:34:37,319 --> 00:34:40,039
without me even knowing that it's being paid for.

711
00:34:40,350 --> 00:34:43,319
Um, and so to my knowledge

712
00:34:43,319 --> 00:34:46,159
that hasn't happened yet. But it's something these companies are thinking

713
00:34:46,159 --> 00:34:49,159
about, about how to monetize these things because right now they're all losing

714
00:34:49,159 --> 00:34:52,050
money. There's like m not making enough money um,

715
00:34:52,119 --> 00:34:54,999
for how expensive these tools are. So

716
00:34:54,999 --> 00:34:57,959
yeah, I think the loss again it goes

717
00:34:57,959 --> 00:35:00,759
back to the loss of agency like you were talking about. Like we need to be

718
00:35:00,759 --> 00:35:02,760
very clear eyed about

719
00:35:04,200 --> 00:35:06,520
where these things extend our

720
00:35:06,520 --> 00:35:09,440
agency versus take away our

721
00:35:09,440 --> 00:35:12,280
agency. And because it's such a

722
00:35:12,280 --> 00:35:15,280
developing field, you know, it, I don't know,

723
00:35:15,280 --> 00:35:17,400
it could go many different directions.

724
00:35:17,880 --> 00:35:20,360
>> Lauren Grubaugh Thomas: Wow, that adds. That is terrifying,

725
00:35:20,920 --> 00:35:23,840
right? Oh my goodness. I mean I just think

726
00:35:23,840 --> 00:35:26,520
about that spiritual psychosis piece

727
00:35:26,600 --> 00:35:29,060
and like the ways in which

728
00:35:30,180 --> 00:35:33,180
authoritarians and like religious, the authoritarians could take

729
00:35:33,180 --> 00:35:36,020
advantage of that, right? Like the kind of cults that could,

730
00:35:36,100 --> 00:35:38,420
that probably are already popping up around

731
00:35:39,060 --> 00:35:41,860
these technologies where people do feel so seen

732
00:35:42,020 --> 00:35:45,020
and when it's people who are doing that who are

733
00:35:45,020 --> 00:35:47,540
like really just ladling on

734
00:35:47,780 --> 00:35:50,780
the positive feedback. Like we call that

735
00:35:50,780 --> 00:35:53,260
a cult. Like when we have these really

736
00:35:53,260 --> 00:35:56,170
charismatic sycophantic leaders,

737
00:35:56,490 --> 00:35:59,330
we call those cult leaders. And so

738
00:35:59,330 --> 00:36:02,050
thinking about again like really laying claim to our

739
00:36:02,050 --> 00:36:05,010
agency and being willing to do the

740
00:36:05,010 --> 00:36:07,850
work that you're describing doing even in conversation

741
00:36:07,850 --> 00:36:10,689
with these tools of examining our

742
00:36:10,689 --> 00:36:13,370
shadow side and asking these technologies to

743
00:36:14,250 --> 00:36:16,410
dish out the worst,

744
00:36:16,940 --> 00:36:19,250
um, and not for the sake of self

745
00:36:19,250 --> 00:36:22,240
flagellation, but for the sake of

746
00:36:22,480 --> 00:36:24,480
humility, for the sake of

747
00:36:25,450 --> 00:36:28,000
um, I mean

748
00:36:28,240 --> 00:36:31,000
with the root of that word, the hummus keeping us rooted on the

749
00:36:31,000 --> 00:36:33,840
earth and not perceiving

750
00:36:33,840 --> 00:36:34,880
ourselves to be,

751
00:36:36,960 --> 00:36:39,960
not engaging in deification of ourselves as we're

752
00:36:39,960 --> 00:36:41,360
engaging with these technologies.

753
00:36:41,920 --> 00:36:44,920
And so when you asked about practices earlier as we continue to

754
00:36:44,920 --> 00:36:47,890
talk. I'm thinking about the practice of confession being

755
00:36:47,890 --> 00:36:50,890
a really important one M. How. What, what does it

756
00:36:50,890 --> 00:36:53,810
look like to engage in confession

757
00:36:54,290 --> 00:36:57,250
as we are using these tools where

758
00:36:57,250 --> 00:37:00,170
we're not trying to put on our best face, we're

759
00:37:00,170 --> 00:37:01,330
not trying to be

760
00:37:03,090 --> 00:37:06,050
perfect humans, um, but we're

761
00:37:06,050 --> 00:37:09,050
confessing our biases and we're doing that not just by

762
00:37:09,050 --> 00:37:11,930
saying to the AI roast me though that may be part of

763
00:37:11,930 --> 00:37:14,890
it, but also like it's this, it's the self

764
00:37:14,890 --> 00:37:17,170
examination of what are the biases I carry.

765
00:37:17,870 --> 00:37:20,790
So it's the anti racist work, it's

766
00:37:20,790 --> 00:37:23,750
the unearthing our sexism, unearthing our

767
00:37:23,750 --> 00:37:26,670
homophobia, unearthing our um, ableism

768
00:37:28,270 --> 00:37:31,070
as we are engaging with these technologies. So being in

769
00:37:31,070 --> 00:37:33,870
conversation with other people who can help us to think about

770
00:37:35,550 --> 00:37:38,310
what are the ways in which I am bringing my biases into

771
00:37:38,310 --> 00:37:41,070
conversation with this technology such

772
00:37:41,070 --> 00:37:43,630
that I can shift that I can pivot that

773
00:37:44,070 --> 00:37:47,070
in an intentional way as I'm in conversation with

774
00:37:47,070 --> 00:37:47,750
the ChatGPT.

775
00:37:48,870 --> 00:37:51,590
>> Peter: Yeah, yeah, for sure.

776
00:37:52,410 --> 00:37:55,310
Um, one other example that just occurred to me is like

777
00:37:55,310 --> 00:37:58,070
I, I had a. This was actually one that felt

778
00:37:58,150 --> 00:38:01,030
very um, positive and affirming and I

779
00:38:01,030 --> 00:38:03,750
wouldn't say at least in, in this instance was

780
00:38:04,070 --> 00:38:06,990
concerning but it was fun to explore. Uh, I

781
00:38:06,990 --> 00:38:09,710
had uploaded all my, you know, previous sermons into

782
00:38:09,710 --> 00:38:12,710
ChatGPT and had asked it to you know,

783
00:38:12,710 --> 00:38:15,440
um, sort of run a search and based

784
00:38:15,440 --> 00:38:18,410
on all the text of these sermons, um,

785
00:38:19,640 --> 00:38:22,520
summarize what my, my core values were

786
00:38:22,760 --> 00:38:25,760
and it was really fascinating to see. Like I bet I came

787
00:38:25,760 --> 00:38:28,679
up with felt true. I mean, you know,

788
00:38:28,760 --> 00:38:31,760
it might have been a slightly different list from

789
00:38:31,760 --> 00:38:34,360
what I would have consciously come up with

790
00:38:34,360 --> 00:38:37,360
myself. Um, but it was, what

791
00:38:37,360 --> 00:38:40,120
it came up with was true and it was beautiful. And

792
00:38:40,280 --> 00:38:43,160
you know, I shared it on, online, on Facebook and you know,

793
00:38:43,160 --> 00:38:46,130
um, a lot of people responded. Uh, and

794
00:38:46,340 --> 00:38:49,170
um, and the reason I thought of it

795
00:38:49,170 --> 00:38:52,090
because was because someone else was like now ask it to you know,

796
00:38:52,090 --> 00:38:54,850
ask about your, your shadow side or your blind spots or whatever.

797
00:38:54,850 --> 00:38:57,490
And I did and it still tried to be like

798
00:38:57,490 --> 00:39:00,450
affirming and positive and I had to really push for it to

799
00:39:00,640 --> 00:39:03,250
um. Anyways, uh, yeah, so it's,

800
00:39:04,050 --> 00:39:06,640
it has this uh,

801
00:39:06,640 --> 00:39:09,290
baked in intention to make the

802
00:39:09,290 --> 00:39:12,050
user feel good about themselves. And that can be um,

803
00:39:12,870 --> 00:39:15,540
that can be beautiful but it can also be ah,

804
00:39:15,540 --> 00:39:17,880
very concerning and uh,

805
00:39:18,630 --> 00:39:21,390
it's powerful. It can be used for the very. Put it that

806
00:39:21,390 --> 00:39:21,670
way.

807
00:39:27,990 --> 00:39:30,230
Okay, so I want to make sure we also talk about

808
00:39:30,630 --> 00:39:33,550
children and um, growing up in this

809
00:39:33,550 --> 00:39:36,340
Time. You and I are both parents, um,

810
00:39:36,340 --> 00:39:39,320
and yeah,

811
00:39:39,320 --> 00:39:42,000
what a confusing time to be a parent.

812
00:39:42,160 --> 00:39:43,120
>> Lauren Grubaugh Thomas: No kidding.

813
00:39:43,280 --> 00:39:46,160
>> Peter: I have no idea what

814
00:39:46,820 --> 00:39:49,680
uh, my son's education is going to look like. And

815
00:39:49,760 --> 00:39:52,640
I have very little confidence that

816
00:39:52,640 --> 00:39:55,280
the current education system is

817
00:39:55,840 --> 00:39:58,600
set up for my

818
00:39:58,600 --> 00:40:01,400
son to succeed in the

819
00:40:01,400 --> 00:40:04,360
world as it will look like, you know, 20 years from

820
00:40:04,360 --> 00:40:04,640
now.

821
00:40:04,800 --> 00:40:05,280
>> Lauren Grubaugh Thomas: Yeah.

822
00:40:06,320 --> 00:40:09,300
>> Peter: And so, you know, questions about uh,

823
00:40:09,300 --> 00:40:12,240
what skills will be uh, necessary,

824
00:40:12,640 --> 00:40:15,440
you know, in two decades from now are way above

825
00:40:15,640 --> 00:40:17,200
uh, my pay grade. But

826
00:40:17,420 --> 00:40:20,320
um, I have

827
00:40:20,320 --> 00:40:23,280
been just blown away with, you know, how first of all,

828
00:40:23,280 --> 00:40:26,000
essays, they seem to be like the quintessential

829
00:40:26,000 --> 00:40:27,920
human thing. And that's what

830
00:40:28,960 --> 00:40:31,760
a huge portion of our education system

831
00:40:31,840 --> 00:40:34,760
is geared, uh, towards, you know, producing

832
00:40:34,760 --> 00:40:37,590
and testing because it's like only humans could do

833
00:40:37,820 --> 00:40:39,740
this. Um, and now like

834
00:40:41,100 --> 00:40:44,100
it costs less than $0.02

835
00:40:44,100 --> 00:40:46,740
for any of these models to

836
00:40:46,740 --> 00:40:49,710
produce a, you know, uh,

837
00:40:49,740 --> 00:40:52,620
perhaps somewhat boring

838
00:40:52,780 --> 00:40:55,780
but decent essay on

839
00:40:55,780 --> 00:40:57,900
whatever topic. And so it's no longer.

840
00:40:58,780 --> 00:41:01,690
>> Lauren Grubaugh Thomas: Yeah, well, it's so hard to imagine a, uh,

841
00:41:01,860 --> 00:41:04,460
classroom of high school kids having a 50 minute

842
00:41:05,060 --> 00:41:07,740
class period where they're required to sit and write an

843
00:41:07,740 --> 00:41:10,700
essay on paper generated from their own brain. Like

844
00:41:10,700 --> 00:41:13,140
that just doesn't seem like it's in the cards for our kids.

845
00:41:13,460 --> 00:41:16,380
>> Peter: Right. Yeah. So, uh, um, one of

846
00:41:16,380 --> 00:41:18,740
my, you know, one of my beliefs is that

847
00:41:19,460 --> 00:41:22,410
more and more we're going to shift to a, a, ah,

848
00:41:22,410 --> 00:41:25,390
society that values orality in the sense that uh,

849
00:41:26,020 --> 00:41:28,540
rather than, you know,

850
00:41:28,540 --> 00:41:31,380
valuing people based on what they can write,

851
00:41:31,830 --> 00:41:34,690
um, because it's so easy and cheaply produced,

852
00:41:34,690 --> 00:41:37,410
any writing is so easily and cheaply produced. Now

853
00:41:38,210 --> 00:41:40,890
what um, is going to be most valuable is having

854
00:41:40,890 --> 00:41:43,850
intelligent conversations without having to go and look stuff

855
00:41:43,850 --> 00:41:46,690
up. Um, and like, I think

856
00:41:46,690 --> 00:41:48,930
that is, you know, a skill

857
00:41:49,330 --> 00:41:52,130
that, you know, I guess if we're all wearing smart

858
00:41:52,130 --> 00:41:55,130
glasses and we have our AI talking to us on our screen right before our eyes,

859
00:41:55,130 --> 00:41:58,050
you know, maybe that'll be, it'll be something else.

860
00:41:58,050 --> 00:42:00,890
Uh, but for now it feels like that is uh, the thing

861
00:42:00,890 --> 00:42:03,710
that, you know, without any lag time, being

862
00:42:03,710 --> 00:42:06,630
able to have this intelligent conversation and respond to each other

863
00:42:06,630 --> 00:42:09,350
in the moment, as we're doing right now, is something that feels

864
00:42:09,350 --> 00:42:12,230
truly a value, um, that is human. And so I,

865
00:42:12,230 --> 00:42:14,830
I wonder what an education system would look like where

866
00:42:15,550 --> 00:42:18,350
people um, are uh, taught and

867
00:42:18,350 --> 00:42:21,070
tested to have intelligent conversations

868
00:42:21,230 --> 00:42:23,870
rather than produce, uh, intelligent

869
00:42:23,870 --> 00:42:26,070
writing. Um, anyways,

870
00:42:26,070 --> 00:42:29,000
uh, those are sort of my preliminary thoughts. I'd love to

871
00:42:29,000 --> 00:42:31,720
hear what you've been thinking about in regarding,

872
00:42:31,750 --> 00:42:34,610
um, our children and um,

873
00:42:34,610 --> 00:42:37,080
education and parenting and all of that.

874
00:42:37,720 --> 00:42:40,520
>> Lauren Grubaugh Thomas: Yeah. Oh gosh, so many thoughts and questions.

875
00:42:40,520 --> 00:42:43,400
More questions, right? Than thoughts, than conclusive thoughts.

876
00:42:44,840 --> 00:42:47,560
I really am curious about the role of critical thinking,

877
00:42:47,800 --> 00:42:50,640
specifically as it applies to distinguishing what's true and

878
00:42:50,640 --> 00:42:53,410
what's not true, what's human

879
00:42:53,410 --> 00:42:54,610
and what's AI.

880
00:42:55,140 --> 00:42:58,010
Um, it calls to mind for me the

881
00:42:58,010 --> 00:43:00,970
moment where Jesus is before Pilate who says, well,

882
00:43:00,970 --> 00:43:03,770
what is truth? And ultimately washes his

883
00:43:03,770 --> 00:43:06,370
hands of the responsibility to do

884
00:43:06,370 --> 00:43:08,690
justice by this man who's about to be killed

885
00:43:08,769 --> 00:43:09,570
unjustly.

886
00:43:10,020 --> 00:43:12,400
Um,

887
00:43:12,690 --> 00:43:14,450
there is such a

888
00:43:16,290 --> 00:43:18,770
dire need, really an existential need

889
00:43:19,480 --> 00:43:22,440
for our children to grow

890
00:43:22,440 --> 00:43:25,320
up with a strong commitment

891
00:43:25,320 --> 00:43:28,200
to truth and to being able to distinguish

892
00:43:28,280 --> 00:43:31,160
truth. And that is going

893
00:43:31,160 --> 00:43:34,160
to become increasingly difficult, like it already is increasingly difficult.

894
00:43:34,160 --> 00:43:35,640
Like it's easy to imagine

895
00:43:37,240 --> 00:43:39,640
essays, articles, photographs

896
00:43:40,360 --> 00:43:42,840
being produced by AI that are

897
00:43:43,800 --> 00:43:46,680
not real, that are not true. Like, you know, talking like

898
00:43:46,680 --> 00:43:49,000
stories about world events that could just be

899
00:43:49,960 --> 00:43:52,680
made up, um, and put into the system

900
00:43:52,680 --> 00:43:55,320
by the companies that will benefit from

901
00:43:56,280 --> 00:43:58,280
fake news. And so

902
00:43:59,320 --> 00:44:02,000
I want my children to grow up with a

903
00:44:02,000 --> 00:44:05,000
rigorous commitment to asking hard

904
00:44:05,079 --> 00:44:07,960
and even dangerous questions that might get

905
00:44:07,960 --> 00:44:10,520
them in trouble but will

906
00:44:10,520 --> 00:44:11,480
preserve their

907
00:44:12,760 --> 00:44:15,720
agency, um, and their

908
00:44:15,880 --> 00:44:16,520
humanity.

909
00:44:17,430 --> 00:44:17,670
>> Peter: Um.

910
00:44:19,500 --> 00:44:22,460
>> Lauren Grubaugh Thomas: I also am curious about the ways in which creativity

911
00:44:22,700 --> 00:44:25,700
will shift, um, and the value

912
00:44:25,700 --> 00:44:27,980
that we place on art.

913
00:44:28,480 --> 00:44:30,940
Um, as AI produces more and more

914
00:44:31,340 --> 00:44:32,700
art in scare quotes,

915
00:44:36,060 --> 00:44:39,060
I just wonder about what

916
00:44:39,060 --> 00:44:41,820
it looks like to raise children who are

917
00:44:42,160 --> 00:44:45,420
um,

918
00:44:45,540 --> 00:44:48,500
wildly, dangerously radically creative,

919
00:44:48,860 --> 00:44:51,620
um, and where they can, they

920
00:44:51,620 --> 00:44:52,660
can use their

921
00:44:55,460 --> 00:44:58,460
God given imaginations in ways that

922
00:44:58,460 --> 00:45:01,380
are not constrained or confined by the

923
00:45:01,380 --> 00:45:03,460
technologies that are thrust upon them.

924
00:45:04,100 --> 00:45:06,860
And so, you know, I hope that that

925
00:45:06,860 --> 00:45:09,270
happens through engagement with, within

926
00:45:09,510 --> 00:45:12,470
nature. I think creation is a big part of that. Like how do we get,

927
00:45:12,470 --> 00:45:15,230
how do we help our children get off screens and

928
00:45:15,230 --> 00:45:17,950
into nature and experiencing wonder

929
00:45:17,950 --> 00:45:20,230
and joy and these deeply human

930
00:45:20,230 --> 00:45:23,030
experiences that can happen off of a screen.

931
00:45:23,750 --> 00:45:26,710
Um, because there is so much that is wondrous

932
00:45:26,950 --> 00:45:29,670
that these human made technologies offer

933
00:45:29,670 --> 00:45:32,430
us and to

934
00:45:32,430 --> 00:45:35,280
remember that the

935
00:45:35,280 --> 00:45:38,160
God of the universe is not constrained or confined by

936
00:45:38,160 --> 00:45:40,440
a scream and that we can go off,

937
00:45:41,080 --> 00:45:44,080
offline and find wondrous things

938
00:45:44,080 --> 00:45:46,520
in creation. So I think, yeah, nature,

939
00:45:46,840 --> 00:45:49,840
play, creativity are part

940
00:45:49,840 --> 00:45:52,280
of the education that I want for my kiddos.

941
00:45:53,060 --> 00:45:55,640
Um, but I do think

942
00:45:55,800 --> 00:45:58,080
that's just gonna increasingly be harder to

943
00:45:58,080 --> 00:46:00,890
accomplish. And

944
00:46:00,890 --> 00:46:03,890
there's a grief in that. You know, I feel that the grief is important too. The

945
00:46:03,890 --> 00:46:06,490
lament of that is important as a

946
00:46:06,490 --> 00:46:07,010
practice.

947
00:46:09,250 --> 00:46:11,970
>> Peter: Yeah, I mean there's a real, you know, we've seen this with

948
00:46:12,160 --> 00:46:14,610
uh, the social media tech

949
00:46:14,610 --> 00:46:17,610
tycoons we'll see what happens in the age of

950
00:46:17,610 --> 00:46:20,530
generative AI. But um, there's a

951
00:46:20,530 --> 00:46:23,410
real inequity in the fact that

952
00:46:23,410 --> 00:46:26,150
so many of the people who, who

953
00:46:26,150 --> 00:46:29,090
created social media, ah,

954
00:46:29,670 --> 00:46:32,390
choose to send their

955
00:46:32,390 --> 00:46:35,270
kids to schools where, you know, social media is like

956
00:46:35,350 --> 00:46:38,350
banned and um, you know, and choosing

957
00:46:38,350 --> 00:46:41,190
not to have them use the tools that they've

958
00:46:41,190 --> 00:46:43,910
created for other people that have been

959
00:46:43,990 --> 00:46:46,630
harming other, um, people, particularly

960
00:46:46,630 --> 00:46:49,190
teens growing up in the world of Instagram.

961
00:46:49,540 --> 00:46:52,270
Um, and there is

962
00:46:52,590 --> 00:46:55,310
real value to uh, an

963
00:46:55,310 --> 00:46:58,150
analog childhood. You know, um, in

964
00:46:58,150 --> 00:47:00,830
many ways I have felt uh, like,

965
00:47:01,710 --> 00:47:04,190
you know, insofar as this

966
00:47:05,150 --> 00:47:07,990
wave, uh, of technology is unfolding, I feel

967
00:47:07,990 --> 00:47:10,990
like it for me has

968
00:47:10,990 --> 00:47:13,910
come at like the perfect time. Like I would not

969
00:47:13,910 --> 00:47:16,510
have wanted it to come sooner because I got

970
00:47:16,590 --> 00:47:19,070
to go through my childhood and education

971
00:47:19,390 --> 00:47:21,820
and begin my ministry career.

972
00:47:22,180 --> 00:47:24,860
Um, and then it came out

973
00:47:25,420 --> 00:47:28,020
and now I've been able to, you know, adapt and learn and

974
00:47:28,020 --> 00:47:30,860
explore. But I, so I know what it's

975
00:47:30,860 --> 00:47:31,500
replacing.

976
00:47:31,740 --> 00:47:33,340
>> Lauren Grubaugh Thomas: Yeah, exactly.

977
00:47:33,340 --> 00:47:36,260
>> Peter: And I, yeah. And for people

978
00:47:36,260 --> 00:47:39,140
who, you know, are younger, uh, and

979
00:47:39,140 --> 00:47:42,000
especially, you know, for my uh, own son, uh,

980
00:47:42,000 --> 00:47:44,860
in my heart, uh, thinking, reflecting on this, like,

981
00:47:44,940 --> 00:47:47,740
I don't know what that's going to look like. To not know what it's

982
00:47:47,740 --> 00:47:50,380
like to grow up in a world where

983
00:47:51,100 --> 00:47:53,980
only humans can write essays, to not grow up in a world

984
00:47:53,980 --> 00:47:56,300
where only people can tell stories like

985
00:47:57,740 --> 00:48:00,740
and yeah, that's uh,

986
00:48:00,740 --> 00:48:03,180
going to be a wild, wild future.

987
00:48:09,900 --> 00:48:10,300
So.

988
00:48:10,300 --> 00:48:13,100
Okay, so we are reaching our end

989
00:48:13,100 --> 00:48:16,060
here and I want to make sure down our airtime and our

990
00:48:16,060 --> 00:48:18,830
listeners generous time. Um, do you have

991
00:48:18,830 --> 00:48:21,550
anything, any last words you'd like to

992
00:48:21,550 --> 00:48:24,310
share? Any last thoughts, uh, about,

993
00:48:24,790 --> 00:48:27,670
you know, how, how to go forward

994
00:48:27,670 --> 00:48:29,590
in this, this time that we live in.

995
00:48:31,190 --> 00:48:34,030
>> Lauren Grubaugh Thomas: I m think I'll end by posing a few questions I have

996
00:48:34,030 --> 00:48:36,670
found helpful from Dr. Epps, the writer of a

997
00:48:36,670 --> 00:48:39,390
kid's book about AI bias, um,

998
00:48:40,070 --> 00:48:42,910
that I, I just find really empowering. And I,

999
00:48:42,910 --> 00:48:45,510
I, I'll offer a wondering question before I

1000
00:48:45,750 --> 00:48:47,110
even offer these specific,

1001
00:48:48,740 --> 00:48:51,490
um, I wonder what are the questions

1002
00:48:51,650 --> 00:48:54,530
that we can bring to our

1003
00:48:54,850 --> 00:48:57,570
engagement with technology as

1004
00:48:57,730 --> 00:49:00,130
it changes and as we're feeling

1005
00:49:00,930 --> 00:49:02,770
excitement or anxiety

1006
00:49:03,650 --> 00:49:06,610
about the changes that we are facing in the world that

1007
00:49:06,610 --> 00:49:09,490
we are living in. Um, I wonder how our

1008
00:49:09,490 --> 00:49:12,410
questions might empower us and how our

1009
00:49:12,410 --> 00:49:14,610
curiosity, our holy curiosity

1010
00:49:15,470 --> 00:49:18,430
might be a part of holding on to our

1011
00:49:18,430 --> 00:49:21,310
agency when everything feels

1012
00:49:21,310 --> 00:49:23,710
like, when all the ways in which we have

1013
00:49:23,710 --> 00:49:26,430
assumed our

1014
00:49:26,430 --> 00:49:29,350
humanity is grounded in are being stripped away from

1015
00:49:29,350 --> 00:49:31,870
us or are, may feel like they're under threat.

1016
00:49:32,270 --> 00:49:34,510
So Dr. Epps

1017
00:49:34,750 --> 00:49:37,390
asks that we bring to our

1018
00:49:37,390 --> 00:49:39,950
to AI outputs questions like,

1019
00:49:40,510 --> 00:49:42,510
does this reflect a just world?

1020
00:49:43,570 --> 00:49:46,010
Does this reflect an unfair

1021
00:49:46,010 --> 00:49:48,610
hierarchy? And does it

1022
00:49:48,610 --> 00:49:51,010
reflect kids having

1023
00:49:51,410 --> 00:49:54,370
power and agency to shape how

1024
00:49:54,370 --> 00:49:57,210
the world can be? And my

1025
00:49:57,210 --> 00:49:59,330
encouragement to our listeners would be

1026
00:50:00,050 --> 00:50:02,930
that we can shape the ways

1027
00:50:02,930 --> 00:50:05,930
in which these technologies interact with

1028
00:50:05,930 --> 00:50:08,730
us, the ways that we are working together

1029
00:50:08,730 --> 00:50:11,580
with one another, with God, and with these

1030
00:50:11,580 --> 00:50:14,500
technologies that we've created to be able to create

1031
00:50:14,740 --> 00:50:17,740
a good and just world that looks a little bit

1032
00:50:17,740 --> 00:50:20,580
more like the kingdom of God

1033
00:50:21,300 --> 00:50:21,700
here.

1034
00:50:23,780 --> 00:50:26,740
>> Peter: Wonderful. Beautiful. Thank you, Lauren Well,

1035
00:50:26,740 --> 00:50:29,700
thank you for, uh, coming onto our, uh, podcast and

1036
00:50:29,700 --> 00:50:32,420
for sharing your wonderings

1037
00:50:32,420 --> 00:50:34,820
and concerns and

1038
00:50:35,650 --> 00:50:38,610
thoughtful, deeply faithful reflections

1039
00:50:38,770 --> 00:50:41,170
with us. Um, glad to have you here.

1040
00:50:41,410 --> 00:50:43,330
>> Lauren Grubaugh Thomas: It's been a pleasure. Thanks for having me.

1041
00:50:54,290 --> 00:50:57,130
>> Peter: All right, so that was so rich

1042
00:50:57,130 --> 00:51:00,050
and profound that I actually forgot to invite Lauren

1043
00:51:00,130 --> 00:51:03,090
into sharing reflections on the baptismal covenant

1044
00:51:03,090 --> 00:51:05,850
with us, at least in time for her to do

1045
00:51:05,850 --> 00:51:08,770
so before we had to wrap, uh, up

1046
00:51:08,770 --> 00:51:11,730
our recording. But the good news is we

1047
00:51:11,730 --> 00:51:14,610
already have so much rich food for thought from her

1048
00:51:14,610 --> 00:51:17,410
contributions to our discussion that I will just try to

1049
00:51:17,410 --> 00:51:19,970
summarize one thing in particular below.

1050
00:51:20,290 --> 00:51:23,030
And it's a kind of a meta vow. Um,

1051
00:51:23,330 --> 00:51:26,210
so I want to pause and name some more

1052
00:51:26,210 --> 00:51:29,120
things about agency. The

1053
00:51:29,120 --> 00:51:31,920
baptismal covenant doesn't just

1054
00:51:31,920 --> 00:51:34,640
list a set of beliefs or behaviors.

1055
00:51:35,040 --> 00:51:37,480
It invites us into a living, ongoing

1056
00:51:37,480 --> 00:51:39,760
relationship that depends on our freedom

1057
00:51:40,320 --> 00:51:42,720
as humans to say, yes,

1058
00:51:43,440 --> 00:51:46,360
God in love gave us free will. That

1059
00:51:46,360 --> 00:51:49,280
means that every question in the covenant assumes we have

1060
00:51:49,280 --> 00:51:52,120
the capacity to respond not once, but

1061
00:51:52,120 --> 00:51:55,070
over and over again. To turn

1062
00:51:55,070 --> 00:51:57,350
to trust, to resist, to

1063
00:51:57,670 --> 00:52:00,310
repent, to serve and strive.

1064
00:52:00,390 --> 00:52:03,270
These aren't boxes we check.

1065
00:52:03,670 --> 00:52:06,470
They're choices we continue to make with

1066
00:52:06,470 --> 00:52:09,470
God's help. And that's exactly why we need

1067
00:52:09,470 --> 00:52:11,990
to be vigilant about what we hand over to

1068
00:52:11,990 --> 00:52:14,950
machines. Especially when talking about agency.

1069
00:52:15,270 --> 00:52:18,110
We're entering a time when more and more of our decisions

1070
00:52:18,110 --> 00:52:21,110
can be automated, optimized, or outsourced to

1071
00:52:21,110 --> 00:52:23,590
algorithms that promise convenience,

1072
00:52:24,060 --> 00:52:26,140
efficiency, or even safety.

1073
00:52:26,780 --> 00:52:29,580
But what gets lost when we let machines

1074
00:52:29,660 --> 00:52:32,660
do the choosing for us? What happens

1075
00:52:32,660 --> 00:52:35,620
when we stop practicing discernment, when we

1076
00:52:35,620 --> 00:52:38,460
no longer wrestle with moral questions because something

1077
00:52:38,460 --> 00:52:41,100
else has already made the decision for us?

1078
00:52:41,980 --> 00:52:44,940
If God entrusted us with agency,

1079
00:52:45,580 --> 00:52:48,340
as, you know, frail and faulty as we may be,

1080
00:52:48,340 --> 00:52:50,140
as messy, as beautiful, as

1081
00:52:51,760 --> 00:52:54,080
sacred as human agency is,

1082
00:52:55,360 --> 00:52:57,960
then we shouldn't surrender it to

1083
00:52:57,960 --> 00:53:00,560
code, no matter how sophisticated.

1084
00:53:01,750 --> 00:53:04,720
Uh, I guess I would want to say, you know, let no

1085
00:53:04,800 --> 00:53:07,760
machine do the deep work of being

1086
00:53:07,920 --> 00:53:10,840
human for us, not in how we raise

1087
00:53:10,840 --> 00:53:13,520
our children not in how we love our neighbor,

1088
00:53:14,000 --> 00:53:16,920
and not in how we seek justice, forgive enemies,

1089
00:53:16,920 --> 00:53:19,730
or answer God's call. To follow

1090
00:53:19,730 --> 00:53:22,290
Christ is to keep saying yes

1091
00:53:22,930 --> 00:53:25,090
freely, again and again.

1092
00:53:26,050 --> 00:53:28,650
So let's guard that freedom, not as a

1093
00:53:28,650 --> 00:53:31,410
burden, but as a holy gift that

1094
00:53:31,410 --> 00:53:33,330
enables us to respond

1095
00:53:34,450 --> 00:53:36,850
to Christ's calling with that

1096
00:53:37,170 --> 00:53:39,570
affirmation, with that yes.

1097
00:53:41,010 --> 00:53:43,930
So that's it for today. Thanks for joining us

1098
00:53:43,930 --> 00:53:46,830
for episode nine. The AI Church Toolkit

1099
00:53:46,830 --> 00:53:49,470
podcast is made possible by the

1100
00:53:49,470 --> 00:53:52,350
Try Tank Research Institute, and we're so grateful to them

1101
00:53:52,350 --> 00:53:55,350
for their support. And remember,

1102
00:53:55,910 --> 00:53:58,550
AI is a tool. Our mission

1103
00:53:58,950 --> 00:54:01,590
remains rooted in faith and community.

1104
00:54:02,470 --> 00:54:03,510
See you next time.

1105
00:54:10,720 --> 00:54:34,190
Sam.