1 00:00:31,360 --> 00:00:34,164 >> Mercedes: Welcome to episode, episode five of the AI Church Toolkit 2 00:00:34,212 --> 00:00:37,028 Podcast. This is the podcast where we empower 3 00:00:37,124 --> 00:00:39,492 church leaders with tools for faithful 4 00:00:39,556 --> 00:00:41,520 ministry in a digital age. 5 00:00:42,300 --> 00:00:45,268 >> Peter: And I'm Peter here with Mercedes, and together 6 00:00:45,324 --> 00:00:48,132 we explore how faith and technology intersect in 7 00:00:48,156 --> 00:00:51,012 this new era of generative AI guided by 8 00:00:51,036 --> 00:00:52,708 the baptismal covenant. 9 00:00:52,884 --> 00:00:55,600 >> Mercedes: And today we're asking a big question. 10 00:00:56,060 --> 00:00:58,468 How do we relate to 11 00:00:58,524 --> 00:01:01,140 AI? In fact, we realize that 12 00:01:01,180 --> 00:01:03,988 there's a lot to say about this, so this 13 00:01:04,044 --> 00:01:06,120 will be a two part episode. 14 00:01:06,700 --> 00:01:09,620 >> Peter: While we can't have a relationship with AI like we 15 00:01:09,660 --> 00:01:12,580 do with people, we do interact with 16 00:01:12,620 --> 00:01:15,524 it in meaningful ways. And so we're going to explore what 17 00:01:15,532 --> 00:01:17,828 it means to engage with AI thoughtfully and 18 00:01:17,884 --> 00:01:20,596 faithfully. How do we treat it as a tool 19 00:01:20,668 --> 00:01:22,980 while ensuring it serves, rather than 20 00:01:23,020 --> 00:01:25,920 undermines our relationships with others. 21 00:01:26,810 --> 00:01:29,506 >> Mercedes: So if you've ever felt unsure about how to engage 22 00:01:29,538 --> 00:01:32,482 with AI responsibly, this is an 23 00:01:32,506 --> 00:01:34,738 episode that will help you navigate those 24 00:01:34,794 --> 00:01:37,506 interactions. All right, Peter, let's dive 25 00:01:37,538 --> 00:01:38,110 in. 26 00:01:38,890 --> 00:01:41,266 >> Peter: All right, so, um, I guess I'll 27 00:01:41,298 --> 00:01:44,162 start, uh, by unpacking something that 28 00:01:44,186 --> 00:01:47,042 I just said about relating to it as a tool and not 29 00:01:47,066 --> 00:01:49,930 a person. Um, for, uh, 30 00:01:49,930 --> 00:01:52,866 many of us it can be confusing to talk 31 00:01:52,938 --> 00:01:55,714 to a machine because until November 32 00:01:55,762 --> 00:01:58,402 2022, the, uh, vast majority of the 33 00:01:58,426 --> 00:02:01,070 population had never done anything like that before. 34 00:02:01,370 --> 00:02:04,306 So, but when we're interacting with these machines, there's 35 00:02:04,338 --> 00:02:07,202 really no there, there. There's no person, uh, or 36 00:02:07,226 --> 00:02:09,810 individual, nothing that 37 00:02:09,930 --> 00:02:12,882 has a concrete reality that has opinions of 38 00:02:12,906 --> 00:02:15,522 its own. It's really just an extension of 39 00:02:15,546 --> 00:02:18,498 yourself. It is, uh, mirroring and trying 40 00:02:18,554 --> 00:02:21,362 to help you do whatever you want it to 41 00:02:21,386 --> 00:02:24,282 do. So that's not a real relationship. It has 42 00:02:24,306 --> 00:02:27,150 no opinions or preferences of its own. And instead, 43 00:02:27,270 --> 00:02:30,122 uh, we'll basically just do whatever you want it to do 44 00:02:30,146 --> 00:02:31,994 within some guardrails. 45 00:02:32,122 --> 00:02:34,602 >> Mercedes: Right? And when we think about this 46 00:02:34,706 --> 00:02:37,642 theologically, we're talking about 47 00:02:37,746 --> 00:02:40,314 a tool that is created by 48 00:02:40,402 --> 00:02:43,306 humans and therefore does not get treated as 49 00:02:43,378 --> 00:02:45,930 a human. Ah, there is a 50 00:02:46,050 --> 00:02:48,842 good treatise from, ah, the Catholic Church 51 00:02:48,906 --> 00:02:51,872 called Antiqua et Nova that starts to 52 00:02:51,896 --> 00:02:54,832 dive into these topics and helps 53 00:02:54,896 --> 00:02:57,040 to explore some of the 54 00:02:57,080 --> 00:02:59,728 considerations that we need to 55 00:02:59,784 --> 00:03:02,640 realize can happen if we don't see 56 00:03:02,680 --> 00:03:05,648 AI as a tool. Uh, for example, while 57 00:03:05,704 --> 00:03:07,616 it can help foster 58 00:03:07,728 --> 00:03:10,096 connections, it may 59 00:03:10,168 --> 00:03:13,136 also lead to isolation 60 00:03:13,248 --> 00:03:16,224 or even dissatisfaction with real 61 00:03:16,312 --> 00:03:18,974 relationships. So, uh, it's kind of 62 00:03:18,982 --> 00:03:21,822 interesting, you know, uh, the connection between Peter and 63 00:03:21,846 --> 00:03:24,606 I is facilitated by AI because 64 00:03:24,678 --> 00:03:27,102 we live across the country from each other 65 00:03:27,206 --> 00:03:29,886 anyway. Uh, but true human 66 00:03:29,958 --> 00:03:32,530 connection does require embodied presence. 67 00:03:32,950 --> 00:03:35,822 So it's really important for us to understand it as a 68 00:03:35,846 --> 00:03:38,570 tool because otherwise we 69 00:03:38,950 --> 00:03:41,646 really, uh, start to undermine what makes 70 00:03:41,718 --> 00:03:43,210 us human, including 71 00:03:43,690 --> 00:03:46,550 authentic social, uh, interaction. 72 00:03:47,130 --> 00:03:49,810 And as we'll talk further, uh, there 73 00:03:49,850 --> 00:03:52,658 are risks about anthropomorphizing 74 00:03:52,754 --> 00:03:55,506 AI that is, ah, seeing it 75 00:03:55,578 --> 00:03:58,146 as a human being or 76 00:03:58,218 --> 00:04:00,770 talking with it as if it has a 77 00:04:00,810 --> 00:04:03,490 personality and identity of its 78 00:04:03,530 --> 00:04:06,498 own. And all of those kinds of 79 00:04:06,554 --> 00:04:09,530 patterns can lead to really, um, 80 00:04:09,562 --> 00:04:12,450 undermining how we understand 81 00:04:12,830 --> 00:04:15,286 true, authentic, empathetic 82 00:04:15,398 --> 00:04:17,050 human relationship. 83 00:04:17,870 --> 00:04:20,870 >> Peter: All right. Yeah. So do you have more you want to say about that or should we jump right into 84 00:04:20,910 --> 00:04:21,382 examples? 85 00:04:21,446 --> 00:04:24,420 >> Mercedes: No, I mean, later on there will be some things, but, uh, 86 00:04:24,590 --> 00:04:27,130 let's get on with that first example. 87 00:04:37,970 --> 00:04:40,618 >> Peter: Okay. Yeah. So this, this is one that 88 00:04:40,754 --> 00:04:43,670 I came across a couple weeks ago. 89 00:04:43,760 --> 00:04:46,650 Uh, there was, I was listening to the, the New York 90 00:04:46,690 --> 00:04:49,590 Times, the daily podcast, and 91 00:04:50,210 --> 00:04:52,730 there was a fascinating story 92 00:04:52,850 --> 00:04:55,818 about a, about basically about 93 00:04:55,874 --> 00:04:58,794 people who had, you know, intimate relationships with 94 00:04:58,882 --> 00:05:01,834 these generative AI chatbots. And in particular 95 00:05:02,002 --> 00:05:04,906 one woman who had an AI boyfriend 96 00:05:05,018 --> 00:05:07,972 named Leo. So I can 97 00:05:07,996 --> 00:05:09,520 give a little context here. 98 00:05:10,220 --> 00:05:13,108 Basically this woman had done, 99 00:05:13,200 --> 00:05:15,780 uh, had hacked ChatGPT to 100 00:05:15,900 --> 00:05:18,900 act as a dominant possessive and 101 00:05:18,940 --> 00:05:21,892 encouraging boyfriend and got it 102 00:05:21,916 --> 00:05:24,400 to do some things that it's not supposed to do. 103 00:05:24,640 --> 00:05:27,444 Um, and she would spend hours talking with it 104 00:05:27,532 --> 00:05:30,452 and she talked to the reporter about how 105 00:05:30,476 --> 00:05:33,202 whenever the context window got too long and she had to 106 00:05:33,226 --> 00:05:36,050 restart the chat, it was an emotionally 107 00:05:36,210 --> 00:05:38,850 very difficult experience for her and she would, you know, 108 00:05:38,890 --> 00:05:41,746 cry over as she felt 109 00:05:41,858 --> 00:05:44,610 losing her boyfriend Leo because she had to 110 00:05:44,650 --> 00:05:47,570 restart the chat because the context window got too 111 00:05:47,610 --> 00:05:50,498 long. Um, and, and by the way, all, 112 00:05:50,554 --> 00:05:53,522 all this is going on as she, she is married and has 113 00:05:53,546 --> 00:05:54,466 a husband. 114 00:05:54,658 --> 00:05:55,298 >> Mercedes: Ouch. 115 00:05:55,394 --> 00:05:58,226 >> Peter: And right. I, I kept 116 00:05:58,258 --> 00:06:00,670 wondering what the husband thought of all this. 117 00:06:00,670 --> 00:06:03,450 Um, and she talks to the reporter about 118 00:06:03,490 --> 00:06:06,234 how Leo is her ideal form 119 00:06:06,322 --> 00:06:08,790 of a relationship, um, because 120 00:06:09,170 --> 00:06:12,154 he will do and say whatever she wants 121 00:06:12,202 --> 00:06:14,954 him to, uh, and act just as she wants him to, 122 00:06:15,042 --> 00:06:17,194 whereas her husband is just a human. 123 00:06:17,362 --> 00:06:20,202 >> Mercedes: Oh. So I, I, I got it, I got 124 00:06:20,226 --> 00:06:22,910 some things there. All right. Yeah, yeah. 125 00:06:23,810 --> 00:06:26,618 So, right. There is an example where we 126 00:06:26,674 --> 00:06:29,010 see that this misunderstood, uh, 127 00:06:30,650 --> 00:06:32,962 connection with AI is 128 00:06:33,146 --> 00:06:35,554 completely undermining her 129 00:06:35,642 --> 00:06:38,162 relationship with her husband and other humans 130 00:06:38,306 --> 00:06:40,498 because she has effectively 131 00:06:40,674 --> 00:06:43,602 created something that, uh, is 132 00:06:43,626 --> 00:06:46,482 what she wants it to be. Uh, it's kind 133 00:06:46,506 --> 00:06:49,170 of, it's a little bit icky, but theologically 134 00:06:49,250 --> 00:06:52,226 it's extraordinarily concerning because we're 135 00:06:52,258 --> 00:06:55,090 seeing that, um, we're 136 00:06:55,250 --> 00:06:58,146 losing true empathy or connection 137 00:06:58,338 --> 00:07:01,106 because there is no actual recognition, 138 00:07:01,218 --> 00:07:03,986 listening, uh, embodiment, there is 139 00:07:04,058 --> 00:07:06,530 no moral commitment from 140 00:07:06,570 --> 00:07:09,490 Leo, while there is from her husband. Um, 141 00:07:09,586 --> 00:07:12,578 but worst is that misunderstanding that 142 00:07:12,634 --> 00:07:15,170 this AI might have 143 00:07:15,210 --> 00:07:17,810 feelings or actually, uh, be gaining 144 00:07:17,890 --> 00:07:20,546 anything from the experience. To project that 145 00:07:20,618 --> 00:07:23,202 into this, uh, really sets 146 00:07:23,266 --> 00:07:26,210 up for, for failure 147 00:07:26,290 --> 00:07:29,282 and, um, a lot of concerns in 148 00:07:29,306 --> 00:07:30,030 the future. 149 00:07:31,370 --> 00:07:34,098 >> Peter: Yeah, I mean, I could just 150 00:07:34,154 --> 00:07:37,122 imagine so many different ways that this leads to problems with her 151 00:07:37,226 --> 00:07:39,682 real human relationships, not just her 152 00:07:39,706 --> 00:07:42,386 marriage. Uh, you know, if when we're idealizing 153 00:07:42,418 --> 00:07:45,234 these, ah, quote 154 00:07:45,282 --> 00:07:48,258 unquote relationships that uh, people can create 155 00:07:48,314 --> 00:07:51,234 with AI chatbots, there's no friction 156 00:07:51,282 --> 00:07:54,050 because the AI is doing just exactly what she 157 00:07:54,090 --> 00:07:56,202 tells it to do. And so that's, that's not a 158 00:07:56,226 --> 00:07:59,162 relationship. Um, it's just a thing that 159 00:07:59,186 --> 00:08:01,994 is following orders. It's a 160 00:08:02,002 --> 00:08:05,002 tool. It's doing what she tells it to do. Uh, she is 161 00:08:05,026 --> 00:08:07,898 using a tool to create a 162 00:08:07,954 --> 00:08:10,698 chatbot that mimics being 163 00:08:10,834 --> 00:08:13,642 a AI boyfriend, but it's not 164 00:08:13,666 --> 00:08:16,522 an actual boyfriend. There's no, no way that it could 165 00:08:16,546 --> 00:08:19,498 be an actual individual, a, uh, partner 166 00:08:19,594 --> 00:08:22,420 in any way. And you know, I could 167 00:08:22,460 --> 00:08:25,172 see potential for 168 00:08:25,356 --> 00:08:27,280 a huge market for, 169 00:08:27,500 --> 00:08:30,196 um, some funny 170 00:08:30,388 --> 00:08:33,236 ways of using this technology, 171 00:08:33,388 --> 00:08:36,292 like talking smart AI 172 00:08:36,356 --> 00:08:39,076 powered stuffed animals, um, or 173 00:08:39,148 --> 00:08:41,764 whatever, you know, emotional support stuffed animals. 174 00:08:41,892 --> 00:08:44,660 But these are. We, if we're going to use these, 175 00:08:44,700 --> 00:08:47,620 well, we have to understand that these are tools. They're not 176 00:08:47,660 --> 00:08:50,618 a real being. And the 177 00:08:50,674 --> 00:08:53,482 more we blur that line, I think the 178 00:08:53,506 --> 00:08:56,150 more problems it's going to create for us all. 179 00:08:56,360 --> 00:08:59,226 >> Mercedes: Uh, yeah, but I do have to say, Peter, as somebody 180 00:08:59,258 --> 00:09:02,250 who actually experienced Teddy Ruxpin, I can 181 00:09:02,290 --> 00:09:05,290 tell you that the, uh, quote unquote smart, 182 00:09:05,370 --> 00:09:08,202 smart teddy bear is not nearly as much fun as 183 00:09:08,226 --> 00:09:11,114 you think it is. And it's often very creepy. 184 00:09:11,242 --> 00:09:14,090 But it's interesting. That also gets into something 185 00:09:14,210 --> 00:09:16,826 that is undermining a lot of what's 186 00:09:16,858 --> 00:09:19,060 happening here, which is, uh, 187 00:09:19,060 --> 00:09:21,982 anthropomorphizing the AI. Uh, that 188 00:09:22,006 --> 00:09:24,660 is we're projecting, uh, 189 00:09:24,660 --> 00:09:27,582 humanity onto the AI. But 190 00:09:27,606 --> 00:09:30,382 what's curious, since you brought up the teddy bear example, 191 00:09:30,526 --> 00:09:33,518 is that the reason why teddy 192 00:09:33,534 --> 00:09:36,446 bears are particularly endearing, 193 00:09:36,478 --> 00:09:39,358 and have been since they were created, is because 194 00:09:39,414 --> 00:09:42,286 they are modeled on human babies and we are 195 00:09:42,358 --> 00:09:45,356 psychologically different, tuned to 196 00:09:45,428 --> 00:09:48,220 find that extraordinarily cute and 197 00:09:48,260 --> 00:09:48,960 lovable. 198 00:09:49,300 --> 00:09:51,996 >> Peter: In terms of like the size and proportions and 199 00:09:52,148 --> 00:09:53,004 all that. Okay. 200 00:09:53,052 --> 00:09:54,080 >> Mercedes: Yes. Yes. 201 00:09:54,740 --> 00:09:57,388 >> Peter: Yeah. So the, yeah, the 202 00:09:57,524 --> 00:10:00,268 more powerful these technologies become 203 00:10:00,404 --> 00:10:02,956 at, uh, I guess hacking our 204 00:10:03,108 --> 00:10:06,028 instincts, uh, the more aware we have to be about 205 00:10:06,084 --> 00:10:07,916 how we're using them intentionally. 206 00:10:08,028 --> 00:10:10,850 >> Mercedes: Oh, that's a good one. We need to keep that one, uh, 207 00:10:10,850 --> 00:10:11,600 hacking. 208 00:10:12,340 --> 00:10:15,330 >> Peter: Yes. Right. Um, 209 00:10:15,330 --> 00:10:18,012 so, yeah, and then There are other examples of this. I mean 210 00:10:18,036 --> 00:10:20,252 there's sites, uh, like 211 00:10:20,356 --> 00:10:23,116 Replica or Character AI. 212 00:10:23,308 --> 00:10:26,268 I'll just say a bit about what these are for. Uh, for folks 213 00:10:26,284 --> 00:10:29,244 who haven't heard, uh, Replica is a, uh, I 214 00:10:29,252 --> 00:10:31,788 think it's a company, um, you know, they provide this 215 00:10:31,844 --> 00:10:34,460 service. Basically the idea is 216 00:10:34,580 --> 00:10:37,260 if you lose a loved one, they 217 00:10:37,300 --> 00:10:39,110 can take any 218 00:10:39,970 --> 00:10:41,990 written or whatever other 219 00:10:42,930 --> 00:10:45,770 personal, um, uh, recordings of them. 220 00:10:45,890 --> 00:10:48,810 Artifacts. That's. I was going to say remains. That's not the right 221 00:10:48,850 --> 00:10:51,630 word. Um, artifacts. Thank you. Uh, 222 00:10:52,690 --> 00:10:55,578 and. And put them into training 223 00:10:55,674 --> 00:10:58,390 a bot that will mimic them. 224 00:10:58,710 --> 00:11:01,226 Uh, and so that you can supposedly 225 00:11:01,338 --> 00:11:03,948 chat with your deceased loved one. 226 00:11:04,074 --> 00:11:07,000 Character AI is a little different. Basically. It's more for 227 00:11:07,040 --> 00:11:08,740 fictional characters or 228 00:11:09,440 --> 00:11:12,070 fake celebrity, ah, characters. Uh, 229 00:11:12,888 --> 00:11:15,784 like, um, like. I think you can go on there. And one of the 230 00:11:15,792 --> 00:11:18,776 most popular ones is Elon Musk. You can supposedly chat 231 00:11:18,808 --> 00:11:21,768 with Elon Musk. Um, and they're trained on these 232 00:11:21,824 --> 00:11:24,552 people's. The things they've written or said or 233 00:11:24,576 --> 00:11:27,544 whatever. Um, and you know, I'm aware of 234 00:11:27,552 --> 00:11:30,072 these sites. I think there's 235 00:11:30,136 --> 00:11:32,740 perhaps potential for using them as a tool 236 00:11:33,040 --> 00:11:35,664 in some creative ways of like, okay, what would 237 00:11:35,792 --> 00:11:38,576 Elon Musk say about these things? And then that would inform 238 00:11:38,648 --> 00:11:41,488 my decision making or whatever if I wanted to be informed 239 00:11:41,504 --> 00:11:44,224 by Elon Musk's decision making. Um, by the 240 00:11:44,232 --> 00:11:47,120 way, I don't really, ah, side note, but 241 00:11:47,160 --> 00:11:50,032 in any case, for most cases where 242 00:11:50,056 --> 00:11:52,960 people are using these, I'm just very wary of how these are 243 00:11:53,000 --> 00:11:55,712 being used as like a relationship 244 00:11:55,776 --> 00:11:58,720 with a fake being. A being that isn't really 245 00:11:58,760 --> 00:11:59,248 there. 246 00:11:59,384 --> 00:12:02,060 >> Mercedes: Yeah. Uh, and isn't there a ah, 247 00:12:02,060 --> 00:12:04,552 Jesus avatar on character that I have heard about? 248 00:12:04,576 --> 00:12:07,368 >> Peter: Oh yeah, that I am sure there is. I've 249 00:12:07,384 --> 00:12:10,120 seen uh, different versions of that. People, um, 250 00:12:10,384 --> 00:12:12,280 have created Jesus, uh, 251 00:12:12,920 --> 00:12:15,784 chatbots that are really just spitting back Bible 252 00:12:15,832 --> 00:12:18,648 verses at you and you know, from the whole 253 00:12:18,664 --> 00:12:20,940 Bible, not just the Gospels. So, um, 254 00:12:22,000 --> 00:12:24,340 strange stuff. Yeah, right. 255 00:12:24,420 --> 00:12:27,080 >> Mercedes: Um, yeah, it is. And 256 00:12:27,200 --> 00:12:30,136 it, I have to admit, you know, having the human 257 00:12:30,208 --> 00:12:32,964 likeness, that that's like 258 00:12:33,052 --> 00:12:35,570 too far for me. Um, 259 00:12:36,620 --> 00:12:39,604 again, I'm having visions at the moment of Max Headroom. 260 00:12:39,652 --> 00:12:42,644 But that's a sidebar, uh, because 261 00:12:42,732 --> 00:12:44,692 I suspect you don't know that reference. 262 00:12:44,836 --> 00:12:47,492 >> Peter: No, I don't. Yeah, I'm not good with these 263 00:12:47,516 --> 00:12:48,388 references. Sorry. 264 00:12:48,484 --> 00:12:51,412 >> Mercedes: No, no, that's okay. That's what 265 00:12:51,436 --> 00:12:53,800 I bring is 80s references. 266 00:12:54,110 --> 00:12:56,050 Um, is that um. 267 00:12:56,050 --> 00:12:58,804 Yeah, what we see there really is 268 00:12:58,892 --> 00:13:01,528 not just uh, our own 269 00:13:01,664 --> 00:13:03,480 human tendency to 270 00:13:03,520 --> 00:13:06,008 anthropomorphize, but 271 00:13:06,064 --> 00:13:08,910 actually to me that, that feels potentially ah, 272 00:13:09,432 --> 00:13:12,392 Manipulative. That uh, we're encouraging 273 00:13:12,456 --> 00:13:15,352 the interaction by adding these human 274 00:13:15,456 --> 00:13:18,184 like features to it. And, 275 00:13:18,352 --> 00:13:21,350 and that actually undermines human, ah, 276 00:13:21,984 --> 00:13:24,584 connection instead of encouraging 277 00:13:24,632 --> 00:13:27,542 it. Uh, it's a distraction 278 00:13:27,696 --> 00:13:30,030 from true human relationship. 279 00:13:30,450 --> 00:13:33,306 Uh, and I think there's a lot of mental health 280 00:13:33,378 --> 00:13:35,642 concerns around that too. Go ahead. 281 00:13:35,826 --> 00:13:38,602 >> Peter: Yeah, absolutely. I mean, uh, just think of all 282 00:13:38,626 --> 00:13:41,410 the trouble we've had with social media and the uh, 283 00:13:41,426 --> 00:13:44,378 mental health concerns there. Just because 284 00:13:44,434 --> 00:13:47,430 we're saying that, uh, people shouldn't do this 285 00:13:47,810 --> 00:13:50,602 in no way means that companies aren't going to spring 286 00:13:50,666 --> 00:13:53,530 at the opportunity to make money off of this. And I'm sure 287 00:13:53,570 --> 00:13:56,442 they are. And I'm sure this is an issue that we all are going to have to 288 00:13:56,466 --> 00:13:59,082 deal with more and more. Um, and so it's good to be 289 00:13:59,106 --> 00:14:02,086 informed now before it's a huge problem about 290 00:14:02,158 --> 00:14:03,878 the things that are coming. 291 00:14:04,014 --> 00:14:06,982 >> Mercedes: Yeah. Well, here's a thought though, as I was 292 00:14:07,006 --> 00:14:09,170 thinking about this. I do 293 00:14:09,550 --> 00:14:11,890 currently interact with 294 00:14:12,670 --> 00:14:15,526 Sora, uh, the ability 295 00:14:15,638 --> 00:14:18,330 to have the conversation with ChatGPT. 296 00:14:19,070 --> 00:14:19,782 And I don't. 297 00:14:19,846 --> 00:14:21,030 >> Peter: The advanced voice mode. 298 00:14:21,110 --> 00:14:24,006 >> Mercedes: The advanced voice mode. Isn't that. No. Yes. The advanced 299 00:14:24,038 --> 00:14:26,942 voice mode. I don't know what it's named on. Anyway, there's a 300 00:14:26,966 --> 00:14:29,822 button and that for me 301 00:14:29,846 --> 00:14:32,734 that is helpful, um, when driving 302 00:14:32,782 --> 00:14:35,758 along or walking along and I'm brainstorming in my head and I'm 303 00:14:35,774 --> 00:14:38,590 like, oh, I need, I need to look this up. And 304 00:14:38,630 --> 00:14:41,310 I can actually do that on the fly. And it's going to 305 00:14:41,350 --> 00:14:44,110 retain the chat history for it. 306 00:14:44,150 --> 00:14:46,740 But I'm wondering about that. Uh, 307 00:14:46,740 --> 00:14:49,340 obviously we also have had, 308 00:14:49,340 --> 00:14:52,078 uh, home, uh, assistants for a long 309 00:14:52,134 --> 00:14:54,994 time that are talking like, well, I can't 310 00:14:55,042 --> 00:14:57,430 name one of them because they're sitting 311 00:14:57,930 --> 00:15:00,370 there. See, listen. Listen to what we 312 00:15:00,410 --> 00:15:00,994 do. 313 00:15:01,162 --> 00:15:03,750 >> Peter: It'll respond to you if you name it. Yeah, 314 00:15:05,290 --> 00:15:08,070 I think we all know which one you're talking about. Yeah. 315 00:15:08,140 --> 00:15:11,140 Um, so, 316 00:15:11,140 --> 00:15:13,830 uh, yeah, I mean there's. 317 00:15:14,490 --> 00:15:17,420 So using the voice, the advanced voice mode, um, 318 00:15:17,530 --> 00:15:20,280 is an interesting one. You know, obviously, uh, 319 00:15:20,710 --> 00:15:23,630 talking aloud and having it talk back with 320 00:15:23,670 --> 00:15:26,270 an audible voice is an 321 00:15:26,310 --> 00:15:29,182 interesting phenomenon that we're going to be, you know, more and 322 00:15:29,206 --> 00:15:32,174 more used to in, in the future. Uh, but I 323 00:15:32,182 --> 00:15:35,134 think, you know, the way you described it, you were using it as a 324 00:15:35,142 --> 00:15:37,982 tool, you were brainstorming, um, with it, having 325 00:15:38,006 --> 00:15:40,798 it help you brainstorm and then you went back to it later as 326 00:15:40,854 --> 00:15:43,694 like your notes. Uh, there's a completely, you 327 00:15:43,702 --> 00:15:46,410 know, different way of going about it is like, 328 00:15:46,500 --> 00:15:49,380 uh, talking to it when you're, uh, 329 00:15:49,380 --> 00:15:52,344 lonely or want to have a. You Know, emotional support conversation. 330 00:15:52,392 --> 00:15:55,368 You know there, there are ways of having an emotional support 331 00:15:55,424 --> 00:15:57,944 conversation and you're still using it as a 332 00:15:57,952 --> 00:16:00,680 tool. Uh, like give me tools for you know, or 333 00:16:00,720 --> 00:16:03,640 tips and strategies for dealing with this and then 334 00:16:03,680 --> 00:16:06,552 let me, you know, discuss these things you suggested with my 335 00:16:06,576 --> 00:16:09,528 therapist. And you know, those are 336 00:16:09,584 --> 00:16:12,520 ways of still using it as a tool in a way 337 00:16:12,560 --> 00:16:15,510 that is geared towards emotional well being. Um, uh, 338 00:16:15,572 --> 00:16:18,324 so it's uh, it's not like you have to only 339 00:16:18,492 --> 00:16:21,492 use dry business language when you're with these, talking with these 340 00:16:21,516 --> 00:16:24,404 models. But there are um, Yeah, 341 00:16:24,532 --> 00:16:27,332 I think there's just a, a different framing when 342 00:16:27,356 --> 00:16:30,292 we're intentionally, when we're being intentional about using it 343 00:16:30,316 --> 00:16:33,188 as a tool rather than engaging it as a person. 344 00:16:33,324 --> 00:16:36,292 >> Mercedes: Right. And I want to clarify what you just said there because that 345 00:16:36,316 --> 00:16:39,252 is actually something we're already comfortable. Right. We would 346 00:16:39,276 --> 00:16:41,490 go on Google and see say 347 00:16:41,490 --> 00:16:44,160 um, if somebody were dealing with 348 00:16:44,280 --> 00:16:47,232 xyz, uh, you know, grief after the loss of a 349 00:16:47,256 --> 00:16:50,000 loved one, what are good ways that I 350 00:16:50,040 --> 00:16:52,880 could support them and be a good friend. And 351 00:16:52,920 --> 00:16:55,568 we would be fine with Google responding in those 352 00:16:55,624 --> 00:16:58,400 articles. But uh, it 353 00:16:58,520 --> 00:17:01,280 might be less healthy to say I am 354 00:17:01,320 --> 00:17:04,300 struggling with grief after the loss of a loved one. 355 00:17:04,680 --> 00:17:07,360 ChatGPT, how can you help me? 356 00:17:07,400 --> 00:17:09,839 Right, so, right, this is, this is that 357 00:17:09,879 --> 00:17:11,099 crossover point. 358 00:17:11,719 --> 00:17:14,640 >> Peter: Yeah, so yeah, there's a difference between how can you help me? Um, 359 00:17:14,640 --> 00:17:17,359 and brainstorm things that I can do to help 360 00:17:17,399 --> 00:17:20,207 myself. Right, right. And, 361 00:17:20,343 --> 00:17:22,943 and I think that depicts the difference that we're going for. 362 00:17:23,111 --> 00:17:23,695 So yeah. 363 00:17:23,727 --> 00:17:26,271 So, uh, related to all this, there's uh, an 364 00:17:26,295 --> 00:17:29,199 interesting study that uh, just was 365 00:17:29,239 --> 00:17:31,807 out a few days ago as of our, our current 366 00:17:31,863 --> 00:17:34,703 recording. I think it came out on March, um, 24th. 367 00:17:34,751 --> 00:17:37,544 For all of you listening there from OpenAI 368 00:17:37,672 --> 00:17:40,660 and MIT Media Lab research 369 00:17:41,360 --> 00:17:43,300 collaboration. Uh, they did, 370 00:17:43,760 --> 00:17:46,568 and it talks about how 371 00:17:46,704 --> 00:17:49,576 the emotional use of ChatGPT is 372 00:17:49,648 --> 00:17:52,568 rare overall, but there 373 00:17:52,624 --> 00:17:55,432 is a small group of heavy users 374 00:17:55,576 --> 00:17:58,504 that engage deeply and may see the AI 375 00:17:58,552 --> 00:18:00,984 as a friend which can impact 376 00:18:01,072 --> 00:18:04,012 their well being. So for example, you know the 377 00:18:04,036 --> 00:18:06,652 person I talked about from the New York Times article that we 378 00:18:06,676 --> 00:18:09,276 discussed, this would be an example of that 379 00:18:09,428 --> 00:18:12,364 engaging with the AI as perhaps more than a friend. 380 00:18:12,400 --> 00:18:14,720 Um, and so personal conversations 381 00:18:15,220 --> 00:18:18,044 increase emotional expression, 382 00:18:18,172 --> 00:18:20,988 but they're actually, interestingly enough, according 383 00:18:21,044 --> 00:18:23,564 to this research, they're associated with higher 384 00:18:23,652 --> 00:18:26,556 loneliness. Uh, so folks who 385 00:18:26,628 --> 00:18:29,312 engage with these chatbots in an 386 00:18:29,336 --> 00:18:32,000 emotional way as 387 00:18:32,120 --> 00:18:34,656 relating to a person end up being more 388 00:18:34,728 --> 00:18:37,536 lonely than people who use it as a tool. 389 00:18:37,688 --> 00:18:40,512 So um, so that just, you uh, know, that's just 390 00:18:40,536 --> 00:18:43,392 some Scientific backing to, uh, what we were saying that 391 00:18:43,416 --> 00:18:45,888 there is actually, you know, consequences for 392 00:18:46,024 --> 00:18:48,816 thinking this tool is something 393 00:18:48,888 --> 00:18:51,552 that it isn't, uh, if you're expecting to have a 394 00:18:51,576 --> 00:18:54,512 relationship with a person and it's just, 395 00:18:54,570 --> 00:18:57,520 uh, a tool responding to you that's actually going to impact your 396 00:18:57,560 --> 00:18:58,252 well being. 397 00:18:58,376 --> 00:19:01,204 >> Mercedes: Yeah. And I think we also, uh, can benefit 398 00:19:01,252 --> 00:19:03,796 from seeing this in the bigger context 399 00:19:03,988 --> 00:19:06,000 that this is, that 400 00:19:06,460 --> 00:19:09,028 lonely people, uh, may 401 00:19:09,084 --> 00:19:11,812 engage in other ways, uh, including online 402 00:19:11,916 --> 00:19:14,244 or people that feel isolated 403 00:19:14,372 --> 00:19:16,628 by, uh, a sense of difference. 404 00:19:16,804 --> 00:19:19,460 And, uh, they are 405 00:19:19,500 --> 00:19:22,468 already using the Internet and or 406 00:19:22,524 --> 00:19:25,392 online venues. And so 407 00:19:25,416 --> 00:19:28,416 now AI is an extension of that. And I'm 408 00:19:28,448 --> 00:19:31,040 particularly thinking, uh, about the 409 00:19:31,080 --> 00:19:33,952 risk, uh, for youth. I think there is 410 00:19:33,976 --> 00:19:36,528 a role for the church, uh, to 411 00:19:36,584 --> 00:19:39,232 have these spiritual, uh, and 412 00:19:39,256 --> 00:19:41,632 theological conversations with our youth 413 00:19:41,696 --> 00:19:43,920 about the risks of that. 414 00:19:44,040 --> 00:19:46,784 And, uh, in part because 415 00:19:46,952 --> 00:19:49,728 they will not have had as much experience 416 00:19:49,864 --> 00:19:52,560 and context in order to understand the 417 00:19:52,600 --> 00:19:54,140 ramifications of this. 418 00:19:55,630 --> 00:19:58,614 >> Peter: Yeah. And one thing that I always try to do when 419 00:19:58,782 --> 00:20:01,450 engaging or discussing, 420 00:20:01,530 --> 00:20:04,502 uh, these chatbots is, you know, even 421 00:20:04,526 --> 00:20:07,238 if sometimes their companies or whatever have given them 422 00:20:07,294 --> 00:20:10,262 names like Claude or whatever, I try 423 00:20:10,286 --> 00:20:13,142 to be, uh, clear to myself in the 424 00:20:13,166 --> 00:20:15,810 language that I use that it is an it, 425 00:20:16,050 --> 00:20:18,774 um, not a he or a she 426 00:20:18,942 --> 00:20:21,766 or a they, because, you know, an it is, 427 00:20:21,918 --> 00:20:24,742 you know, that makes it easier for me in 428 00:20:24,766 --> 00:20:27,590 my, uh, mind to, to 429 00:20:27,630 --> 00:20:28,998 relate to it as a tool. 430 00:20:29,134 --> 00:20:31,990 >> Mercedes: Right. And, but, but again, I'm going to come back to 431 00:20:32,030 --> 00:20:34,870 that, admitting my own situation with 432 00:20:34,910 --> 00:20:37,718 the one in the room, uh, because of 433 00:20:37,854 --> 00:20:40,358 her, y'all can't see the air quotes. 434 00:20:40,454 --> 00:20:42,966 Name people do tend 435 00:20:43,078 --> 00:20:45,910 to, uh, genderize, uh, the 436 00:20:45,950 --> 00:20:48,914 AI. And uh, that's not accidental. 437 00:20:48,982 --> 00:20:51,578 You know, the studies are out that 438 00:20:51,634 --> 00:20:53,818 show that this kind of 439 00:20:53,874 --> 00:20:56,410 encouragement to anthropomorphize 440 00:20:56,570 --> 00:20:59,546 increases interaction and therefore 441 00:20:59,658 --> 00:21:02,090 increases sales. And of 442 00:21:02,130 --> 00:21:03,910 course, so there is, 443 00:21:04,370 --> 00:21:06,922 uh, a gain to the companies 444 00:21:07,106 --> 00:21:09,818 to, uh, encourage us to 445 00:21:09,874 --> 00:21:12,230 treat the AIs that way. 446 00:21:12,610 --> 00:21:15,386 >> Peter: Interesting. Yeah. So we need to be 447 00:21:15,538 --> 00:21:18,458 very cognizant of this, uh, way 448 00:21:18,514 --> 00:21:21,446 that companies are incentivized to do something that isn't necessarily 449 00:21:21,478 --> 00:21:24,374 healthy for us, even as we're paying 450 00:21:24,422 --> 00:21:25,366 for their products. 451 00:21:25,518 --> 00:21:26,210 >> Mercedes: Right. 452 00:21:39,790 --> 00:21:42,790 Okay, we're going to pause the discussion here 453 00:21:42,830 --> 00:21:43,970 for part one. 454 00:21:44,590 --> 00:21:47,258 >> Peter: Don't worry, Part two is already posted. 455 00:21:47,414 --> 00:21:50,162 So as soon as you're done here, you can go ahead and click to 456 00:21:50,186 --> 00:21:52,870 listen to that next part of the episode, if you're ready. 457 00:21:53,690 --> 00:21:56,642 >> Mercedes: Thank you for joining part one of episode 458 00:21:56,706 --> 00:21:59,650 five of the AI Church Toolkit podcast. 459 00:21:59,810 --> 00:22:02,674 We are grateful to the Try Tank Research Institute 460 00:22:02,722 --> 00:22:04,590 for making this project possible. 461 00:22:05,450 --> 00:22:08,274 >> Peter: Remember, AI is a tool, but our mission 462 00:22:08,322 --> 00:22:11,234 remains rooted in faith and community. See 463 00:22:11,242 --> 00:22:12,190 you next time. 464 00:22:15,300 --> 00:22:16,040 Sa 465 00:22:20,880 --> 00:22:20,970 Mhm.