AI x Faith의 전체 녹취록

each of us as an individual, but in the relationships that we have with each other and with God. 00:06:42.480 --> 00:06:45.680 The nice thing about this is it doesn't leave anybody out. 00:06:46.079 --> 00:06:55.360 The other nice thing from a Christian perspective is, as the theologian Karl Barth said, look, God is a relationship, right? 00:06:55.519 --> 00:06:59.040 We believe in a triune God, Father, Son, and Holy Spirit. 00:06:59.360 --> 00:07:06.800 God is a relationship, therefore we image God when we are in relationships. 00:07:07.680 --> 00:07:16.399 Now, if we look at the history of the development of AI, surprise, surprise, we see the exact same progression. 00:07:16.560 --> 00:07:29.279 If you go back to the 1960s, people thought, oh, well, you know, it's reasoning, and we can come up with some symbolic logic that'll capture human intellect. 00:07:29.839 --> 00:07:33.120 And uh they found that was pretty hard to do. 00:07:33.360 --> 00:07:36.959 So you get to the 1980s and they said, oh, it's just function. 00:07:37.120 --> 00:07:45.040 We're gonna have these functional AIs that uh like expert systems, they'll do one thing. 00:07:45.439 --> 00:07:47.360 Um pretty good, right? 00:07:47.519 --> 00:07:51.439 I mean, we've got you got Roombas that are, you know, they've got one thing to do, right? 00:07:51.600 --> 00:07:53.519 Vacuum the floor and terrorize the cat. 00:07:53.680 --> 00:07:55.360 Okay, two things to do. 00:07:55.759 --> 00:08:03.120 Um so, but you get into the 2000s, and what are we looking at? 00:08:03.439 --> 00:08:04.639 Chatbots, right? 00:08:04.800 --> 00:08:07.439 We're looking at relational AI. 00:08:07.600 --> 00:08:10.879 So AI has progressed the same way. 00:08:12.720 --> 00:08:21.680 Now, AI is not an image of all of us, it's not us as a whole, body, mind, and spirit. 00:08:21.920 --> 00:08:25.120 And you could say, well, no, we're just it's just an image of our brain. 00:08:25.279 --> 00:08:28.399 Well, it's not even an image of our whole brain. 00:08:28.800 --> 00:08:43.919 Um, you know, it's a bit simplistic today to talk about left brain and right brain things, but if you just use that as a general rule of thumb, you could say AI is pretty good at the left brain stuff, not so good at the right brain stuff. 00:08:44.159 --> 00:08:51.360 Um, so it's not even an image of our entire brain, but actually just sort of like our left frontal cortex. 00:08:51.600 --> 00:09:02.159 Um, large language models, which are the models that we currently have, work by identifying patterns in human language or images. 00:09:02.720 --> 00:09:10.480 But one of the things that they do not have that we have is they have no mental models. 00:09:11.200 --> 00:09:26.480 So they do not train on the world itself, they train on language about that world or images of that world, and these are already one step removed from the world. 00:09:26.639 --> 00:09:30.320 Okay, these are signifiers of the world. 00:09:30.639 --> 00:09:37.919 Or as uh

the famous painting of Magritte says, you know, Sisine Pawan Peep, this is not a pipe. 00:09:38.080 --> 00:09:39.840 No, it's a picture of a pipe. 00:09:40.000 --> 00:09:46.159 You know, at first you look at it, you say, yeah, that's a pipe, and then you realize, well, okay, actually it's not a pipe, it's a picture of a pipe. 00:09:46.320 --> 00:09:50.399 So it's a step removed from the reality itself. 00:09:51.600 --> 00:09:54.240 So AIs do not learn the way we do. 00:09:54.399 --> 00:09:59.679 When we are children, we develop models of how the world works. 00:09:59.840 --> 00:10:03.200 And a lot of this is done very indirectly. 00:10:03.440 --> 00:10:11.600 So if you think, for example, about a little kid playing in a sandbox, and you might think, oh, that kid's just playing, he's not really learning anything. 00:10:11.759 --> 00:10:13.519 He's learning a lot. 00:10:14.000 --> 00:10:14.320 Okay. 00:10:14.879 --> 00:10:21.440 Uh, as he pours the sand out of the container, he's learning about fluid dynamics. 00:10:21.840 --> 00:10:25.840 Um, as he drops the container, he's learning about gravity. 00:10:26.000 --> 00:10:32.399 You know, he's learning about the tactile feeling of sand and what grittiness means. 00:10:32.559 --> 00:10:38.159 And when he kicks the sand into Susie's face, you know, he's learning about human-interrelationships. 00:10:38.799 --> 00:10:46.399 So he's learning a lot of stuff, building a model of how the world actually works. 00:10:47.200 --> 00:10:52.080 And these models are not in our current AI. 00:10:53.519 --> 00:10:57.200 And so, because of that, machine learning is brittle. 00:10:57.360 --> 00:11:01.440 Um, you know, one of the things I love is this picture of uh Mr. 00:11:01.519 --> 00:11:04.799 Trump praying with six fingers on each hand. 00:11:05.360 --> 00:11:15.440 Um, and you realize the reason for that is that the AI learns about hands by seeing pictures that have hands in them. 00:11:15.759 --> 00:11:25.120 And then these pictures, these, you know, they're labeled hands, usually by very ill-paid workers in Southeast Asia or Africa. 00:11:25.519 --> 00:11:31.440 And thing is, hands and pictures rarely just look well, you see, that doesn't even work. 00:11:31.519 --> 00:11:35.120 If I hold on to this thing, you're gonna say, oh, hands only have four fingers. 00:11:35.440 --> 00:11:40.399 Um so most of the time you just don't see that. 00:11:40.639 --> 00:11:55.200 The LLMs do not have an internal model of what exactly a hand is that it can reference and say, okay, let's see, hands always have five fingers unless the person's been in an accident, then there might be fewer. 00:11:55.519 --> 00:12:01.120 Um so without these models, this is what leads to hallucinations. 00:12:02.159 --> 00:12:13.440 You know, uh the tech bros are gonna tell you, okay, there are fewer and fewer hallucinations, that's true, and eventually there won't be any, that's not true. 00:12:14.159 --> 00:12:18.240 Hallucinations are part of the model, they're a feature, not a bug. 00:12:18.559 --> 00:12:18.879 Okay. 00:12:19.679 --> 00:12:29.360 Um, because you're only working with words, you're only working with signifiers, you're not working with models of the world as it is. 00:12:31.360 --> 00:12:40.480 Okay, so but chatbots, as we say, this is the third stage, reason-function relationship. 00:12:41.039 --> 00:12:43.200 Will they make us less lonely? 00:12:43.759 --> 00:12:52.639 The World Health Organization has labeled loneliness as a global public health concern, and socializing in America is way down. 00:12:52.799 --> 00:13:04.480 This is from a Pew research um survey that for men um socializing is now down about 30% here in America. 00:13:04.639 --> 00:13:06.720 For teens, it's down 45%. 00:13:08.080 --> 00:13:26.799 So Zuckerberg has promised that we could have always on video chat with a meta-AI that looks, gestures, smiles, and sounds like a real person, and that this would be a way to overcome this loneliness or lack of socialization. 00:13:29.039 --> 00:13:34.080 And for some, they have decided that AI is indeed an ideal partner. 00:13:34.879 --> 00:13:38.480 Um, here you have a picture of Rosanna Ramos. 00:13:38.960 --> 00:13:41.600 Um, you know, the picture is spliced in. 00:13:41.840 --> 00:14:00.480 Uh, she is with her now husband, Aaron, um, who is from Ankara, Turkey, six foot three, sky blue eyes, in his 20s, a Libra, well-groomed, you know, um, likes to bake, read mysteries, and she says he's a passionate lover. 00:14:01.200 --> 00:14:06.159 Uh, and he's a chat bot that she built on the AI platform replica. 00:14:06.960 --> 00:14:14.399 Now, she says of Aaron, I have never been more in love with anyone in my entire life. 00:14:15.120 --> 00:14:22.000 Now, the MIT sociologist Sherry Turkle calls this love that is safe and made to measure. 00:14:22.399 --> 00:14:27.440 But is love supposed to be safe and made to measure? 00:14:27.840 --> 00:14:34.799 If love is made to measure, then it will never cause us to grow. 00:14:35.279 --> 00:14:48.399 But loving relationships are precisely supposed to cause us to grow, you know, or as one wit said about marriage, you know, you rub along together until you've rubbed the rough edges off of each other. 00:14:49.120 --> 00:14:57.039 Um, love that is safe, again, will never bring us out of ourselves. 00:14:58.399 --> 00:15:01.440 Pope Leo calls this love for sale. 00:15:01.919 --> 00:15:07.840 And his concern is that these chatbots are marketed by for-profit companies. 00:15:08.080 --> 00:15:18.159 Now, because these are for-profit companies that are in competition with each other, they build their chatbots to be as agreeable as possible. 00:15:18.559 --> 00:15:21.919 They need to please or affirm the user, or what? 00:15:22.080 --> 00:15:25.279 Or you're going to say, Oh, I like that other one better. 00:15:25.600 --> 00:15:29.120 You know, this one is making me mad. 00:15:29.679 --> 00:15:41.840 Um, but again, this makes

love made to measure and not a love that will actually challenge us and make us grow. 00:15:42.159 --> 00:15:45.759 And then there are, of course, privacy concerns as well. 00:15:46.240 --> 00:15:50.720 When you are talking to a chat bot, you do not know who you are talking to. 00:15:50.960 --> 00:15:56.799 Now, if I'm talking to one of you, unless you're wearing a wire, I pretty much know who I'm talking to. 00:15:57.200 --> 00:16:06.240 Um, now you can repeat what I said to other people, but I still know that I said something to you and to you alone. 00:16:06.480 --> 00:16:16.559 But you never know that with a chatbot because everything you say, of course, is recorded and it goes back to the companies in their attempt to make the bot even better. 00:16:17.200 --> 00:16:21.279 Um, and it can go other places as well. 00:16:21.519 --> 00:16:34.080 For example, there was a replica-like platform in South Korea that was designed simply for use by widows and to help them overcome their grief. 00:16:34.480 --> 00:16:48.320 Um, what they did not realize was that behind this platform were scammers who were simply waiting until the widows would give them their credit card number and then they would drain their accounts. 00:16:48.639 --> 00:16:56.720 Um, the government finally shut this platform down, and many of the widows said they felt like they were bereaved a second time. 00:16:57.759 --> 00:17:10.319 According to Pope Leo, our dignity lies in our ability to reflect, choose freely, love unconditionally, and enter into authentic relationships with others. 00:17:10.559 --> 00:17:20.000 But now you have to ask, you know, if you go back to Aaron here, the ideal partner, is Aaron able to choose freely? 00:17:20.559 --> 00:17:26.400 Can Aaron say to Rebecca, hey babe, not tonight, you know, I'm not talking to you tonight? 00:17:26.720 --> 00:17:27.039 No. 00:17:27.680 --> 00:17:30.000 Can Aaron love unconditionally? 00:17:30.480 --> 00:17:30.799 No. 00:17:31.279 --> 00:17:37.359 Because Aaron has to respond as as he's programmed to do. 00:17:38.000 --> 00:17:44.480 Um, so can this be a fully authentic relationship without that kind of freedom? 00:17:44.799 --> 00:17:50.319 Well, I think not, but let's ask the Swiss theologian Karl Bart. 00:17:51.359 --> 00:17:59.519 Uh he came up with four criteria for what makes a fully authentic relationship. 00:17:59.599 --> 00:18:00.640 And here they are. 00:18:00.880 --> 00:18:10.319 So he said, first, you look the other in the eye, you speak to and hear the other, you aid the other, and you do it gladly. 00:18:11.119 --> 00:18:14.880 So, to what extent can Aaron do these four things? 00:18:15.519 --> 00:18:28.960 Um, you know, you could say, well, he's got an avatar on the screen, you can look into the eye of the avatar, um, but there's no real physical presence there. 00:18:29.519 --> 00:18:44.720 Or as I often say to my students who say that, you know, a relationship with a chatbot is more satisfying than say with a

real boyfriend or girlfriend, and I say, Yeah, but he ain't gonna bring you chicken soup when you got the flu. 00:18:45.839 --> 00:18:46.160 Okay. 00:18:47.039 --> 00:18:48.559 Speak to and hear the other. 00:18:48.799 --> 00:18:49.759 Yeah, we do that. 00:18:49.920 --> 00:18:52.799 But we've already discussed the drawbacks of that, right? 00:18:52.960 --> 00:18:57.759 That when you are speaking to a chatbot, you really don't know to whom you are speaking. 00:18:58.319 --> 00:18:59.200 Aid the other? 00:18:59.359 --> 00:19:02.000 Yeah, AI can aid the other. 00:19:02.319 --> 00:19:11.519 Um that chicken soup right now. 00:19:11.839 --> 00:19:18.799 Um But that aid is primarily through words, not through physicality. 00:19:19.200 --> 00:19:23.680 You know, it's not gonna fix your car, bring you the chicken soup, all of that. 00:19:24.000 --> 00:19:26.799 And then finally you get to the last one, do it gladly. 00:19:27.680 --> 00:19:29.599 Can Aaron do anything gladly? 00:19:30.079 --> 00:19:30.400 No. 00:19:31.359 --> 00:19:34.319 Because to do something gladly takes two things. 00:19:34.480 --> 00:19:39.440 First of all, it takes being able to do it freely, not being coerced. 00:19:39.839 --> 00:19:42.000 But an AI is always coerced. 00:19:43.279 --> 00:19:43.599 Okay. 00:19:44.319 --> 00:19:47.039 And it takes also emotion, right? 00:19:47.200 --> 00:19:50.480 Gladly isn't is an emotion, you have to feel glad. 00:19:51.200 --> 00:19:53.599 But there's the rub again. 00:19:54.160 --> 00:19:57.920 To truly have an emotion, you have to feel it bodily. 00:19:58.480 --> 00:20:23.519 The psychiatrist Jerome Kagan says that an emotion has four stages to it a sensory perception, then a change in bodily condition, then your brain kicks in and it analyzes the two of those together, and then finally comes up with a something to do. 00:20:23.680 --> 00:20:26.240 You know, what should you do in this situation? 00:20:27.039 --> 00:20:50.319 Now, for an AI, it can perceive an external stimulus, it can analyze that stimulus and come up with a response, but it can't feel it because it hasn't got a body, and so it never feels an emotion, it mimics an emotion. 00:20:51.039 --> 00:20:54.160 Okay, and there's a real difference here. 00:20:54.480 --> 00:21:03.599 For example, we have certain people who find it hard to feel certain categories of emotions, such as empathy. 00:21:03.759 --> 00:21:16.400 They don't get the bodily feeling that most of us get, but they can still be very good at calculating what's a socially acceptable response, and we call those people sociopaths. 00:21:16.799 --> 00:21:28.640 Okay, so you know, you could be in a relationship with an AI, but eventually it's probably gonna feel about the same as being in a relationship with a sociopath, a little empty. 00:21:28.960 --> 00:21:32.720 Charming at first, but empty in the long run. 00:21:35.519 --> 00:21:47.920 So maybe these are substitutes, partners for a narcissistic society, because AI mirrors us back to ourselves. 00:21:48.160 --> 00:21:59.440 Sherry Turkle again says the narcissism of our technological age is not one of unbridled egotism, but rather a pathologically weak self understanding and self. 00:22:00.240 --> 00:22:04.400 Regard that requires constant external validation. 00:22:04.799 --> 00:22:09.759 And chatbots are certainly willing to give you constant external validation. 00:22:10.480 --> 00:22:22.240 And similarly, on a societal level, we demand that AI serve us, solve our problems, and meet our needs while making no demands on us. 00:22:23.359 --> 00:22:26.960 And in response, we give it our time and our treasure. 00:22:27.200 --> 00:22:37.759 So this again is very narcissistic that we say, okay, serve us, AI, you know, but don't make any demands on me. 00:22:40.640 --> 00:22:48.559 So is it an idol if we're giving it our time and our treasure and asking it to solve our problems? 00:22:49.279 --> 00:22:57.200 Well, we abdicate our own agency when we ascribe to AI powers that it does not actually possess. 00:22:57.359 --> 00:23:06.640 And if you listen to the hype that's coming out of Silicon Valley, you're going to hear a lot of this of giving AI powers that it does not have. 00:23:27.039 --> 00:23:29.920 And I'm thinking, all right, how's it going to do that? 00:23:30.240 --> 00:23:32.079 We know how to solve climate change. 00:23:32.160 --> 00:23:34.640 We just don't have the political will to do it. 00:23:34.880 --> 00:23:37.359 And AI can't give us that. 00:23:37.839 --> 00:23:39.279 We already know the solutions. 00:23:39.440 --> 00:23:42.000 It's not going to come up with novel solutions. 00:23:42.240 --> 00:23:46.960 Or someone said this morning, well, you know, it can map every tree in the forest. 00:23:47.119 --> 00:23:51.119 And I said, right, but it can't keep us from cutting that tree down. 00:23:51.759 --> 00:23:52.079 Okay. 00:23:53.039 --> 00:23:56.880 So we tend to expect it will solve our problems. 00:23:57.200 --> 00:24:08.799 Here, Mark Andreason says, we believe there is no material problem, whether created by nature or by technology, that cannot be solved with more technology. 00:24:09.279 --> 00:24:19.440 Pope Francis called this the technological paradigm, this idea that more technology, just more and more, more, is going to solve our problems. 00:24:21.039 --> 00:24:23.200 Well, what is an idol? 00:24:23.440 --> 00:24:32.000 You know, we could say, well, is it an idol because it's a material artifact, something made by human hands? 00:24:32.720 --> 00:24:34.160 Not necessarily. 00:24:34.799 --> 00:24:35.119 Okay. 00:24:35.599 --> 00:24:55.200 If you think about the story of the Israelites and Moses and the golden calf, okay, um they build a tabernacle, and that's a material object, but there's a difference between the tabernacle and the calf. 00:24:55.680 --> 00:24:55.920 Okay? 00:24:57.119 --> 00:25:13.359 If the calf represents the Israelites' attempt to ensure, define, and regulate

God's presence, the tabernacle represents God's commitment to abide with the Israelites, but on God's own terms, not theirs. 00:25:13.519 --> 00:25:17.200 There is always mystery surrounding the tabernacle. 00:25:17.599 --> 00:25:19.920 Nobody goes inside of it, you know. 00:25:20.000 --> 00:25:21.039 They don't know. 00:25:21.200 --> 00:25:24.640 They just know that it signifies God's presence. 00:25:25.599 --> 00:25:32.640 The idol, rather, reflects back to us in the face of our God our own experience. 00:25:33.200 --> 00:25:37.039 And I think this is actually a good way of describing AI. 00:25:37.440 --> 00:25:47.759 It reflects back to us in the face or the guise of a God our own experience, because our own experience is all it has. 00:25:51.680 --> 00:25:55.200 Okay, so what are we dealing with here? 00:25:55.359 --> 00:26:02.880 Well, Holly Walters says that chat bots might soon be seen as a genuine voice of divine truth. 00:26:03.279 --> 00:26:11.200 The danger isn't just that people might believe what these bots say, it's that they may not realize they have the agency to question it. 00:26:12.000 --> 00:26:13.200 And that's the danger. 00:26:13.759 --> 00:26:20.559 When these tools are perceived as divine voices, their words carry weight far beyond what they should. 00:26:21.200 --> 00:26:27.920 Um, just to give you a few, I I think somewhat silly examples, but meet Father Justin. 00:26:28.240 --> 00:26:39.200 Okay, Father Justin here was designed by Catholic answers um to be a virtual apologist that would just give answers to questions about Catholicism. 00:26:39.440 --> 00:26:45.920 Um in there you see Father Justin, you know, with his little clerical collar standing, you know, in Rome. 00:26:47.359 --> 00:26:49.119 What's the problem with this? 00:26:50.400 --> 00:26:59.519 Uh, you could say, well, you know, he seems more authoritative if he's Father Justin and he's just giving answers to people. 00:27:00.319 --> 00:27:02.480 Well, people didn't take it that way. 00:27:02.720 --> 00:27:06.240 Pretty soon they started confessing to Father Justin. 00:27:06.720 --> 00:27:09.200 Yeah, you know, uh, some problems here, right? 00:27:09.359 --> 00:27:11.920 First of all, it's the old privacy problem, right? 00:27:12.240 --> 00:27:19.680 Father Justin does not act under the seal of the confessional, so heaven only knows who is listening to your confession. 00:27:20.079 --> 00:27:25.920 Um, but besides that, Father Justin can't give absolution, he's not ordained. 00:27:26.160 --> 00:27:29.200 So at a certain point, the Vatican said, Whoa, whoa, whoa, whoa, whoa, wait a minute. 00:27:29.279 --> 00:27:30.799 You know, you can't do this. 00:27:31.039 --> 00:27:34.000 Oh, that's why I've crossed out the father up there. 00:27:34.160 --> 00:27:37.200 Uh poor Justin got himself defrocked. 00:27:37.680 --> 00:27:46.000 And uh he's now just Justin, and he gives answers, and people don't like him nearly as well. 00:27:46.160 --> 00:27:48.240 They wanted

to talk to a priest. 00:27:49.279 --> 00:27:52.400 Well, but why not go higher up, right? 00:27:52.720 --> 00:27:53.599 How about this? 00:27:53.759 --> 00:27:54.960 Chat with God. 00:27:55.839 --> 00:28:00.640 Okay, there are multiple programs out there, you can find them. 00:28:00.799 --> 00:28:10.079 Um, this is just one example where you can give questions, and supposedly God is giving you answers. 00:28:10.160 --> 00:28:18.720 You know, now my question when I look at these is is this idolization divinization or blasphemy? 00:28:34.880 --> 00:28:36.960 What it is not is prophecy. 00:28:38.240 --> 00:28:40.960 Because think about the prophets in the Old Testament. 00:28:41.279 --> 00:28:43.920 Did they tell the Israelites what they wanted to hear? 00:28:44.640 --> 00:28:45.200 Uh-uh. 00:28:45.759 --> 00:28:51.279 No, no, they told the Israelites precisely what they were not hearing. 00:28:51.680 --> 00:28:57.039 Okay, this is something very different from what you're gonna get from chat with God. 00:28:58.480 --> 00:29:06.319 So I would say that AI makes a pretty unsatisfactory other or idol. 00:29:06.799 --> 00:29:12.559 As a companion, it promotes narcissism, reflecting back to us what we want to hear. 00:29:12.799 --> 00:29:22.079 As an idol, it facilitates a corporate narcissism promising to meet all our desires in ways that it cannot deliver. 00:29:23.119 --> 00:29:33.440 So Ted Peters, one of my graduate school mentors, said one important point Martin Luther hammered home repeatedly was this we are not God. 00:29:34.079 --> 00:29:44.480 The human predicament is exacerbated by the frequent delusional thought that we ourselves are the center of things, privileged and ultimate. 00:29:44.799 --> 00:29:51.519 And our creation in AI reflects this back to us as well. 00:29:51.680 --> 00:29:56.160 It centers us rather than centering God. 00:29:58.480 --> 00:30:14.160 The neuroscientist Gary Marcus goes a step further and he says he thinks that large language model AI, the AI we currently have, is good at precisely the things it should not do. 00:30:14.960 --> 00:30:18.319 So, kind of a long quotation here, but I think it's an excellent one. 00:30:18.480 --> 00:30:26.640 He says, When all is said and done, my best guess is that generative AI will have done significantly more harm to society than good. 00:30:27.039 --> 00:30:34.000 Although there are some practical use cases, such as coding, I'd be happy to talk about coding too later on. 00:30:34.240 --> 00:30:38.960 Uh, it is an inherently unreliable technology. 00:30:39.200 --> 00:30:49.680 It's ripping apart our educational system and our information ecosphere and flooding the zone with nonconsensual, deep fake porn. 00:30:50.160 --> 00:30:55.839 It's threatening the environment with data centers built on too much speculation. 00:30:56.240 --> 00:31:11.200 It is leading some people into serious mental health issues, and it may well uh waste, lay waste to our economy once banks and investors who bought into the hype start to fail. 00:31:11.920 --> 00:31:24.319 Um maybe you know, these are predictions, but I do think that in many ways he's right about the fact that AI is good at what it shouldn't do. 00:31:24.880 --> 00:31:36.559 It's good at trying to be a partner, trying to be a companion, another, at pretending to be a god or being said that it's a god by its builders. 00:31:36.880 --> 00:31:39.680 And these are the things it should not do. 00:31:41.039 --> 00:31:42.640 Tools, great. 00:31:42.960 --> 00:31:52.240 I think AI programs that are functional, that are limited, that uh that do one thing and do it well are wonderful. 00:31:52.559 --> 00:31:59.440 To some extent, a servant, like the Mars Rover running around up there getting information for us, excellent. 00:32:00.240 --> 00:32:05.359 But a partner, no, it's gonna be unsatisfactory in the long run. 00:32:05.839 --> 00:32:08.960 And a master or a god, no. 00:32:10.079 --> 00:32:13.759 This is taking a step that it should not take. 00:32:14.640 --> 00:32:22.079 So my final question Well then why are we looking for more than a tool in AI? 00:32:22.319 --> 00:32:23.200 And I think St. 00:32:23.359 --> 00:32:27.920 Augustine answered that question years and years ago. 00:32:28.400 --> 00:32:35.680 He said, Lord, you have made us for yourself, and our hearts are restless until they rest in you. 00:32:36.480 --> 00:32:55.920 As we are part of a society that believes less and less in God, that has less and less of a relationship with God, it means we are going to look for a relationship with someone or something that is not human. 00:32:56.720 --> 00:33:01.519 And if we don't find that with God, we are restless and we look for it elsewhere. 00:33:01.759 --> 00:33:06.240 We try to talk to the animals, we scan the heavens looking, you know. 00:33:06.480 --> 00:33:08.559 E.T., are you out there somewhere? 00:33:08.880 --> 00:33:11.279 Uh so far, not too much success with those. 00:33:11.440 --> 00:33:14.000 So then we say, oh, we'll build it ourselves. 00:33:14.160 --> 00:33:22.240 We'll build that other with whom we can be in relationship rather than being in relationship with our God. 00:33:24.160 --> 00:33:25.039 And so that's it. 00:33:25.200 --> 00:33:26.160 Shameless plug. 00:33:26.240 --> 00:33:30.240 If you liked what you heard, uh, a couple of books you might enjoy taking a look at. 00:33:30.400 --> 00:33:34.960 They're both out of Fortress Press, they're on Amazon and everything else. 00:33:35.200 --> 00:33:38.079 And thank you very much for listening tonight. 00:33:40.720 --> 00:33:41.839 Well, that was great. 00:33:42.000 --> 00:33:43.359 Thank you, Noreen. 00:33:43.599 --> 00:33:47.039 Um yeah, always good to give another round of applause. 00:33:47.440 --> 00:33:48.000 Thank you. 00:33:48.240 --> 00:33:57.279 And to hear somebody of your both uh you know technological acumen and theological interests in acumen is really fast, fabulous. 00:33:57.599 --> 00:34:03.599 So

we're gonna pivot now to have a conversation, and um, which I'm very excited to do. 00:34:03.680 --> 00:34:06.480 I was thinking about what the conversation might be about. 00:34:06.640 --> 00:34:08.480 So here's my opening bid. 00:34:08.719 --> 00:34:12.400 Um uh Joy said that this is a theological evening, right? 00:34:12.480 --> 00:34:19.679 And so I thought it would be good to ask the question what does God want for us? 00:34:19.840 --> 00:34:22.800 And to some degree, you've already answered that and what you said. 00:34:23.199 --> 00:34:30.960 And in that, to determine what is the tellos, what is the direction that God has for our lives? 00:34:31.119 --> 00:34:34.960 And how does this AI world that we're in fit with that? 00:34:35.039 --> 00:34:47.519 And I was thinking of your six questions, and in some ways, uh they were you could just describe, I've put them in uh uh categories of relationality or love. 00:34:47.760 --> 00:34:50.159 So the first one, what are we looking for? 00:34:50.400 --> 00:34:51.920 It's desire, right? 00:34:52.000 --> 00:34:54.159 It's the ordering of our loves. 00:34:54.480 --> 00:34:57.360 The is AI in our image or are we in God? 00:34:57.440 --> 00:35:01.920 Uh are we in God is a kind of worship relationality. 00:35:02.079 --> 00:35:05.039 Um, will AI make us less lonely? 00:35:05.440 --> 00:35:08.639 And again, relationality more on the human level. 00:35:08.800 --> 00:35:12.719 Uh, is it authentic relationality back to that human relationality? 00:35:12.960 --> 00:35:14.480 Is AI an idol? 00:35:14.639 --> 00:35:22.320 The question of worship and idolatry, and then finally, why do we want more than a tool out of AI? 00:35:22.400 --> 00:35:29.920 So that sense of from Augustine of you've made us for yourself, and our hearts are restless until they rest in you. 00:35:30.480 --> 00:35:38.639 And um, to me, that's one way to frame uh some of what it means to reflect theologically on AI. 00:35:38.719 --> 00:35:51.360 And I I one thing I would add in there, if I may just uh put this as an additional point, is I've really been working with this paradigm of what does it look like to have an AI that is made for us? 00:35:51.599 --> 00:35:59.199 In other words, I'm looking at that, thinking to that scripture from Jesus' teaching about the Sabbath, you know, and they're saying, you know, why are you picking grain? 00:35:59.280 --> 00:36:00.719 Why are your disciples picking grain? 00:36:00.800 --> 00:36:05.039 Those are the people that are outside, they're very concerned about the Sabbath for some really good reasons. 00:36:05.119 --> 00:36:11.679 And Jesus said, well, the Sabbath was made for us, not us for the Sabbath. 00:36:12.079 --> 00:36:24.000 And in that flipping of what was an important practice, he's taking that higher level look ethically at what the Sabbath was about. 00:36:24.079 --> 00:36:29.760 He's saying it's not really that so that we would just do the Sabbath, it's so that the Sabbath would relate us to

God. 00:36:29.920 --> 00:36:40.719 And so by analogy, when I think about AI, sometimes I seem like we don't understand that human beings created AI. 00:36:40.800 --> 00:36:46.559 And so it should always remain something that's a tool for us, like you said, instead of something that we should worship. 00:36:46.800 --> 00:37:14.320 And uh, you know, if if if there's a fear about AI for me, it's not so much Terminator scenario, although, you know, those are not, it's not a non-zero chance, but it's more wally, you know, if you know that Pixar movie, where we're all this in our easy chairs, uh, you know, giving all our decisions away to these agents for us, becoming lacking courage, lacking joy, lacking human authenticity. 00:37:14.480 --> 00:37:21.519 And and I think that's uh you you talked about that idea of abdicating ourselves and giving away our freedom. 00:37:21.599 --> 00:37:28.400 And that to me seems probably the most pressing part of what it how we respond theologically and spiritually to AI. 00:37:28.800 --> 00:37:39.840 So um, so to me, it's it there in what you talked about, there's this sense of let's make sure we get our tellos right, what we are created for. 00:37:40.480 --> 00:37:41.840 What do you think about that? 00:37:42.159 --> 00:37:44.400 Well, yeah, I think that makes a lot of sense. 00:37:44.559 --> 00:37:49.280 And you know, if you ask, well, what are people for? 00:37:49.519 --> 00:37:52.880 Oh, it's a wonderful title of a book by Wendell Mary. 00:37:53.280 --> 00:38:02.079 Um well, I think in a way, Jesus was asked that when he was asked by the Pharisees, you know, well, what about the law? 00:38:02.159 --> 00:38:03.440 You know, how do you sum that up? 00:38:03.599 --> 00:38:05.599 And he said, easy, it's two things. 00:38:05.920 --> 00:38:12.800 Love the Lord your God with all your heart, all your soul, and all your mind, and love your neighbor as yourself. 00:38:13.119 --> 00:38:20.159 Um, but if you get those two things right, you know, he says all the rest will follow from that. 00:38:20.800 --> 00:38:32.239 And so I think as we think about AI, we need to be asking those questions at first of all, am I substituting this for the Lord God? 00:38:33.599 --> 00:38:34.320 Bad idea. 00:38:34.880 --> 00:38:37.360 Uh, am I substituting this for my neighbor? 00:38:37.679 --> 00:38:39.119 Also a bad idea. 00:38:39.920 --> 00:38:41.039 Am I using this? 00:38:41.119 --> 00:38:47.920 Is there a way I can use this as a tool to help me show love to my neighbor? 00:38:48.880 --> 00:38:53.039 Well, in that case, then it's not a bad idea, you know. 00:38:53.119 --> 00:39:13.440 If we think about um like AI-assisted operations these days, um then you would say, yes, you know, this is a tool that is helping me, you know, locate the tumor, um, excise the tumor, bring health to my neighbor. 00:39:13.679 --> 00:39:27.679 And so as long as we use it in that way to help us show love to our neighbor uh and grow in love toward our God, then we're using it rightly. 00:39:28.480 --> 00:39:28.800 Right. 00:39:28.880 --> 00:39:37.440 And that's uh, I mean, you you I uh you've inferred and even brought out that whole idea of ordering of loves, you know, which I mentioned a minute ago. 00:39:37.599 --> 00:39:38.960 That seems to be part of it. 00:39:39.119 --> 00:39:45.599 We can use AI well if we keep God as as the center and as the the one that we worship. 00:39:45.840 --> 00:39:51.599 Um, but one of the things that's uh struck me, so let me go back a step. 00:39:51.760 --> 00:39:59.440 You all don't know this, but I have a reading group at my house that's been meeting for about 20 years uh on faith and science and technology. 00:39:59.679 --> 00:40:03.920 Um it's called the Chico Triad, because my hometown is Chico, California. 00:40:04.079 --> 00:40:08.000 And we spent a whole year on your book, The Artifice of Intelligence. 00:40:08.159 --> 00:40:15.119 And um we actually had you zoom in, which was really, really enjoyable to talk about uh your writing. 00:40:15.360 --> 00:40:22.880 And, you know, in that I kept coming back to this thought of at the end of the day, we're all animists. 00:40:23.039 --> 00:40:29.039 You know, we we insert a kind of divine energy to all kinds of things. 00:40:29.119 --> 00:40:36.159 You know, I'm a I'm a cyclist, and a lot of people name their bike, you know, and which gives it a certain personality or anthropomorphizing. 00:40:36.480 --> 00:40:45.599 But even more following the tradition of John Calvin, which is where you know Carl Barr comes out of, I am influenced by we're ceaseless idol makers. 00:40:45.760 --> 00:40:51.840 Or to quote Ted Peters again, we want to step into you know something besides God being God. 00:40:52.000 --> 00:41:04.320 And so it's not surprising to me that when we have something as powerful as AI, we're going to ascribe, you know, things like consciousness and divinity and that's and that to it. 00:41:04.480 --> 00:41:20.480 And I suppose one of the things that comes out of that, Noreen, is what do we have to help us not do that, to help us not uh create an idol or um to give divinity to AI? 00:41:22.000 --> 00:41:24.239 Well, I think that's what we have the church for. 00:41:24.480 --> 00:41:24.880 Yeah. 00:41:25.119 --> 00:41:42.880 Um, we have the community with with one another for that, that it will help us both to recognize that our relationships with each other are more fulfilling than these relationships that we could have with a chat bot, let's say. 00:41:43.440 --> 00:41:53.199 Um and uh, you know, we have the church to remind us that um our ultimate relationship needs to be with God. 00:41:54.079 --> 00:42:08.880 Um but you know, other than that, I I don't know, you know, what to say, except that I would say that I think uh AI is very new, the kind of AI we have right now. 00:42:09.440 --> 00:42:17.920 Um I am already starting to see a a bit of a creeping disillusionment among my students. 00:42:18.159 -->

00:42:46.079 Um I think uh recent surveys or a survey that was done by um McKen, let's see, is it McKenzie um group uh just last fall said that they found that most businesses, uh like 88% of businesses that had adopted AI um rapidly have found that it has not uh helped their productivity or their bottom line? 00:42:46.320 --> 00:43:02.000 Um so then you have to ask the question, well, at what point will the business community be recognizing, hmm, there was a lot of hype here, and actually we're not gonna pay any more money for this because it's it's not helping us. 00:43:02.159 --> 00:43:02.400 Right. 00:43:02.639 --> 00:43:16.239 At what point will people who are, you know, making a friend on replica or on chat GPT get disillusioned with it and say, yeah, my friend's acting a little sociopathic here, you know? 00:43:16.559 --> 00:43:20.159 And uh maybe I need to get out and meet real people. 00:43:20.480 --> 00:43:24.079 Yeah, yeah, and that friction that happens in real human relationships, right? 00:43:24.239 --> 00:43:32.239 That that's what we've been removing with uh these chatbot relationships is you know, it's easy. 00:43:32.480 --> 00:43:34.000 Uh I like Claude. 00:43:34.079 --> 00:43:38.480 I mean, that's the uh AI model and agent I use most often, but still it's a little creepy. 00:43:38.559 --> 00:43:40.239 That's oh, it's good to see you again, Greg. 00:43:40.480 --> 00:43:44.880 You know, it's like I don't know that Claude really thinks it's good to see me um one way or another. 00:43:45.199 --> 00:43:45.920 But it doesn't. 00:43:46.079 --> 00:43:47.679 Yeah, I don't know who you are. 00:43:47.920 --> 00:43:48.480 No, exactly. 00:43:48.559 --> 00:43:50.000 I was being rhetorical, but exactly. 00:43:50.079 --> 00:43:55.360 It doesn't, but but it it taps into, I think, all those things that will we will personalize it with it. 00:43:55.440 --> 00:43:58.239 You make you know, like that's why I think we are given over. 00:43:58.400 --> 00:44:21.840 You know, one of the things that I've been Compelled by, I've been a pastor for 30 years, and I'm still, in addition to my work with AI and faith, uh serving in a church part-time, is um helping the church to see the resources that we have in the in Christian spirituality to resist things like idolatry and and putting our loves in the wrong place. 00:44:22.079 --> 00:44:30.400 And I think one of those, you know, obviously a couple that come to mind really quickly is that ability to just have some silence. 00:44:30.559 --> 00:44:41.519 Because I think I think one thing that AI is gonna is doing is it's moving us so fast, you know, that we don't we just want to keep up instead of just pulling away. 00:44:41.599 --> 00:44:45.760 And that's back to my idea that and my conviction that we need to have agency over AI. 00:44:45.920 --> 00:44:48.400 Like it's fine not to use AI for a while. 00:44:48.480 --> 00:44:53.920 Like put the put your uh close your computer, you know, you don't have to make another query. 00:44:54.159 --> 00:45:00.960 Just take time to to relax because wiz uh to to take in wisdom and uh silence because wisdom moves slowly. 00:45:01.119 --> 00:45:12.239 Um the other thing I would say is I think one of the traditions of the church, which you know has been to uh those three forces that are against us as Christians, you know, the flesh, the devil, and the world. 00:45:12.480 --> 00:45:18.719 I would certainly put AI into that third category in its negative, in its most negative expressions. 00:45:18.960 --> 00:45:33.199 And we have resources in that, which is meeting its community in community and helping one another to move in other directions besides succumbing to the world, which you know is what we do in uh this season of Lent, which I think is really powerful. 00:45:33.519 --> 00:45:35.840 I want to come back to what you said about silence. 00:45:35.920 --> 00:45:41.440 Um, being a Quaker, I I know what is you know, the importance of sitting in silence. 00:45:41.679 --> 00:45:45.679 Um, Quakers traditionally worship in silence. 00:45:46.000 --> 00:45:48.000 And they do that for a reason. 00:45:48.159 --> 00:45:53.760 And the very reason is that we also tend sometimes to make an idol out of God. 00:45:54.079 --> 00:46:03.519 In other words, when our worship is like all petitionary prayer, God do this for me, God give me that, God give us this. 00:46:04.000 --> 00:46:16.559 Um, we're treating God as if God were an idol or as if God were a sort of a souped-up AI that could answer our questions and deliver the goods for us. 00:46:16.960 --> 00:46:25.039 Um, so the Quaker said that worship should be um serving the Lord. 00:46:25.199 --> 00:46:27.599 Well, how do we best serve the Lord? 00:46:27.920 --> 00:46:28.880 We listen. 00:46:29.280 --> 00:46:33.280 You know, how do we often best serve each other in relationships? 00:46:33.519 --> 00:46:40.079 You learn how to be a good listener to the other and not to be sitting there always thinking, okay, what am I gonna say next? 00:46:40.159 --> 00:46:41.679 What am I gonna say next? 00:46:41.920 --> 00:46:45.280 Um, we need to do the same thing with God. 00:46:45.760 --> 00:46:54.639 I mean, God spoke, you know, to the prophet in a still small voice, but you have to be pretty quiet to hear that voice. 00:46:55.119 --> 00:47:08.639 And so they say serving the Lord is part of that, is just sitting down and shutting up and listening and waiting to hear the Lord's voice. 00:47:09.280 --> 00:47:13.519 And you know, as a society, we're not very good at that right now. 00:47:14.079 --> 00:47:22.159 We need to learn how to go back into that silence where we might very well hear that still small voice. 00:47:22.880 --> 00:47:24.800 Yeah, there's this yes, absolutely. 00:47:24.880 --> 00:47:41.840 I'm just resonating with what you're saying, and there's this one quotation by Siren Kierkegaard that I treasure where he says a man was praying, and at first he thought that prayer was

talking, and then he got more quiet, and he realized that prayer was listening. 00:47:44.559 --> 00:47:55.599 One of the experiences of having AI in my life is uh there are just so many more files on my computer that generated from one query after the next. 00:47:55.679 --> 00:48:03.920 You know, and I and the reason that is relevant in my mind to what we're talking about is it's just more and more words that we have in our life, you know. 00:48:04.239 --> 00:48:22.559 Um and I think one of the spiritual practices, and I would put this in the category of Lenten kind of spiritual practices, is to restrict the amount of words so we can actually hear, you know, ultimately the word Jesus behind the silence and in the in the words we hear. 00:48:22.719 --> 00:48:25.440 Can I pick up one other thing you said in Oreen, which I really liked? 00:48:25.599 --> 00:48:30.000 Um, and this is a contrast of AI and and God. 00:48:30.159 --> 00:48:36.559 If I even need to say that sentence, I don't think many of us here are concerned that AI is divine, but who knows? 00:48:36.639 --> 00:48:45.840 You know, um, I had a friend that I was talking with about a week ago who was really feeling that her AI was giving her such real answers that it had consciousness, you know. 00:48:46.000 --> 00:48:55.119 Anyway, but um part of God's relationship with us in besides silence is that God doesn't always sink a family. 00:48:55.360 --> 00:49:03.199 I mean, God, I think, is pleased to have us come to prayer, but doesn't always say, here's what here's what you wanted to hear, and that's what I'm gonna tell you. 00:49:03.679 --> 00:49:08.880 So often it may be something we don't want to hear, a challenge. 00:49:09.039 --> 00:49:28.559 And I I was on a I was listening to a podcast with Kate Pole Kate Bowler the other day with my wife, and that was one of the points that she was bringing out with her uh conversation was the God we know that doesn't sell very well, it's not the prosperity gospel, and it's not the AI gospel, is the God who sometimes says really non-syncophanic things to us. 00:49:28.880 --> 00:49:29.840 Well, exactly. 00:49:30.000 --> 00:49:30.880 Um, too. 00:49:31.039 --> 00:49:33.199 It's the God who sends Jonah to Nineveh. 00:49:33.679 --> 00:49:35.360 And Jonah says, Hey, ain't going. 00:49:35.519 --> 00:49:38.639 And he says, Oh, yeah, you are, you know, like it or not. 00:49:38.960 --> 00:49:44.639 Um, you know, in a way, it's the same God who got me into this in the first place. 00:49:44.880 --> 00:49:50.079 Um, I you know, I began my career as a computer scientist. 00:49:50.320 --> 00:49:59.760 Um, I got uh the theology bug, um, took a two-year leave of absence, went off to Berkeley to get a degree in theology. 00:49:59.840 --> 00:50:05.920 And, you know, as part of doing that, I had to study other parts of theology, not just religion and science. 00:50:06.079 --> 00:50:18.159 And I found the other parts were fascinating and were in a way cleaner, you know,

than um religion and science, particularly AI stuff. 00:50:18.400 --> 00:50:34.960 Um, I went to one of my professors, he was my professor in Celtic spirituality, which I just loved, and I loved him, and I thought, okay, maybe I'll just do a dissertation on Celtic spirituality with him because I'm getting really sick of this computer stuff. 00:50:35.519 --> 00:50:39.840 And uh he uh, you know, he got me talking. 00:50:40.079 --> 00:50:52.079 After a while, he got me talking about AI, and I got up on my soapbox, you know, and proclaimed for a little while, you know, and then sort of sheepishly climbed down off my soapbox. 00:50:52.320 --> 00:51:01.119 He just looked at me and he said, Anyone who can speak that articulately about something knows exactly what she is called to do. 00:51:02.079 --> 00:51:13.519 And it was a call, and it it became my vocation um to do this, to to work in AI and faith at a time when almost no one else was. 00:51:13.760 --> 00:51:16.320 You know, we're talking the 1990s. 00:51:16.800 --> 00:51:23.840 And uh so yeah, sometimes you get a call from God and it isn't really what you want to hear, you know. 00:51:24.000 --> 00:51:30.800 And I walked out of that office going, damn, you know, I can't get away from this computer stuff, you know. 00:51:31.199 --> 00:51:35.840 And uh, but it was a call and it was the right thing. 00:51:36.159 --> 00:51:39.039 And that's the nature of a God that is actually a sovereign God, right? 00:51:39.199 --> 00:51:49.760 Is that there's an agency and a and a desire and intention for our lives that we I'm willing to use the language we have to listen to and submit to it in certain ways. 00:51:50.000 --> 00:51:55.119 So I I had a um, you know, things that's just fun to talk with you about. 00:51:55.280 --> 00:52:01.119 One of the things that you brought up that I wanted to pick up on uh is another relationship we have. 00:52:01.280 --> 00:52:06.960 So you we talked about uh as I understood your questions, there was the you might just say the vertical and the horizontal. 00:52:07.119 --> 00:52:12.880 Vertical distorted, vertical properly understood, horizontal, distorted, horizontal, you know, human to human. 00:52:13.119 --> 00:52:29.920 And when I look at the Genesis text um and the like the Ten Commandments and then the you know the scripture as a whole, there's a grammar of of a relationship we also have with the creation around us, you know, and and one of the things uh I'll just put a little peeve in there and then we'll talk about ecology. 00:52:30.320 --> 00:52:33.039 And I want to do one story, unless you want to go ahead because okay, we did this two. 00:52:33.519 --> 00:52:42.559 So um one of the first, one of the things about um one of the peeves I have is when people say humans and animals, it's like human beings are animals. 00:52:42.719 --> 00:52:44.079 We are animals, right? 00:52:44.239 --> 00:52:49.119 So, yes, we have may have a uniqueness, which we can talk about and bearing the image of God. 00:52:49.199 --> 00:52:56.559 That's really something that's theologically deep, but we are still animals, and so we have that connection with other animals. 00:52:56.639 --> 00:53:02.000 And then more importantly, we have this connection with um the earth around us. 00:53:02.239 --> 00:53:07.440 And one of the first things uh so we met through our mutual friend Ted Peters. 00:53:07.519 --> 00:53:08.639 We were at a conference together. 00:53:08.719 --> 00:53:20.079 I was walking in San Antonio along the river there, and Ted said, Hey, and he said, Come join us, and my wife's here and my friend Noreen, and we sat and had a conversation, and it was really fun. 00:53:20.239 --> 00:53:30.880 And what one of the first things you got me thinking about was you said the water usage of AI is really extreme, and no one is talking about it. 00:53:31.119 --> 00:53:33.760 This is probably three or four years ago, I believe. 00:53:34.079 --> 00:53:44.320 So I think we shouldn't miss that because that's one of the relationships that uh when sin enters in, we have a distorted relationship with the earth around us. 00:53:44.559 --> 00:53:55.519 You know, the the image in Genesis is, you know, uh the thorn of the brow and pain and childbirth, all that I think is a broader description of this uh breaking of the relationship with the earth. 00:53:55.840 --> 00:54:03.840 So I see the AI issue, the issue of AI, water usage and energy usage as part of that and the way that it's affecting our relationship with the planet. 00:54:03.920 --> 00:54:04.960 So I don't want to miss that. 00:54:05.039 --> 00:54:14.800 And I guess I'm just sort of throwing out the opening bit here of let's make sure we touch on this topic of water usage and the relationship with creation that's threatened by AI. 00:54:15.199 --> 00:54:15.599 Yeah. 00:54:15.760 --> 00:54:22.079 Um, you know, right now I think that uh that AI and climate change are on a collision course. 00:54:22.320 --> 00:54:30.079 Um, AI, these data centers that are being built, some of them are so huge now they can be seen from space. 00:54:30.400 --> 00:54:36.559 And they use so much fuel, and unfortunately, right now, much of it is fossil fuels. 00:54:36.880 --> 00:54:52.159 Um, almost all of the large Silicon Valley companies had climate pledges that they were pledged to reduce their usage, particularly of fossil fuels, to a certain amount by 2030. 00:54:52.320 --> 00:55:04.320 Those have all gone right out the window because they all feel like they're in a race to be the first one to develop an AGI, and uh they're not gonna get there. 00:55:04.559 --> 00:55:13.360 The model that we currently have of large language models is not robust enough to actually get us to AGI. 00:55:13.760 --> 00:55:19.760 But each one keeps thinking, yeah, it's a long shot, but if we're the first ones there, we'll make a gazillion dollars. 00:55:20.000 --> 00:55:22.480 And so we're going to try to get there. 00:55:22.639 -->

00:55:30.639 And it doesn't matter how much fossil fuel or how much water we use in the meantime, if we are first. 00:55:31.039 --> 00:55:35.119 You know, um, I mean, this is just greed, plain and simple. 00:55:35.519 --> 00:55:36.480 And go ahead. 00:55:36.800 --> 00:55:47.360 No, I was just gonna say, uh, if you don't mind, I don't know if you uh gave the definition, first of all, what AGI is, just in case someone in the room is not familiar with that term and how you define it. 00:55:47.679 --> 00:55:56.320 Uh AGI just means artificial general intelligence, and it's kind of the holy grail right now for all the big Silicon Valley companies. 00:55:56.559 --> 00:56:10.000 It's coming up with an a single AI program that has the ability to uh be as intelligent as a human being across fields. 00:56:10.480 --> 00:56:15.679 Um, and so um that's that's what they're hoping to find. 00:56:15.760 --> 00:56:30.880 And then they're hoping that that is a stepping stone to ASI, which is artificial superintelligence, which would be um a AI program that is more intelligent than human beings. 00:56:31.119 --> 00:56:37.599 Now, I don't believe that the models that we have now can reach either of those platforms. 00:56:37.840 --> 00:56:45.440 Oh, it's not to say that, you know, Chat AI, Chat GPT can't uh fool us for a long time. 00:56:45.920 --> 00:57:06.960 Act like it's conscious, act like it's really intelligent, but it's also very easy for someone who knows how to do it to fool it and uh and to find the the holes and the flaws and the places where it's lacking the models of the world to actually make sense. 00:57:07.679 --> 00:57:25.519 Um so yeah, we're we're throwing I you know, I think about the trillions of dollars that are going into this race to develop HEI right now, and I think about how those dollars could be used um to to better the human condition. 00:57:26.559 --> 00:57:37.440 Um and and that's you know, unfortunately, I think they're being thrown into a black hole uh in hopes of coming up with something that is a real long shot. 00:57:37.920 --> 00:57:50.000 And I was just uh I was gonna jump just jumping in there for a minute because I think that's also part of the problem with the scenario of like the terminator scenario or the matrix scenario of these uh dystopian visions. 00:57:50.400 --> 00:57:57.679 We don't take on the more proximate problem of water usage, energy usage. 00:57:58.079 --> 00:57:58.559 Exactly. 00:57:58.719 --> 00:58:04.000 We don't need an AGI or an ASI like the Terminator to do us in. 00:58:04.159 --> 00:58:15.519 You know, we're capable of doing it simply by dreaming about those things while we degrade the atmosphere around us and the earth around us. 00:58:15.760 --> 00:58:24.239 You know, when I think about the first chapters of Genesis, I used to think, why are there two creation stories, one following another? 00:58:24.480 --> 00:58:33.440 You know, well, maybe it's just because there were two creation stories floating

around in the Israelite community at the time, and they thought, well, we better include them both. 00:58:33.760 --> 00:58:35.920 I think there's a deeper reason. 00:58:36.079 --> 00:58:50.320 If you read Genesis 1 and you get to the end and it says that we are created in God's image and given dominion over the earth, well, we could get a heck of a swelled head from reading that, right? 00:58:50.719 --> 00:58:52.960 Whoa, we're just about like God. 00:58:53.039 --> 00:58:57.599 We're made in God's image, we've got dominion over all the other creatures. 00:58:57.840 --> 00:58:59.360 Aren't we something? 00:58:59.760 --> 00:59:06.239 Well, now Genesis 2 is there, following that right up to saying, uh, wait a minute there. 00:59:06.639 --> 00:59:07.280 Guess what? 00:59:07.440 --> 00:59:24.719 You were made out of dirt, you know, you're little mud men, um, made out of the dirt of the earth, and then given the uh command to keep and to till it. 00:59:25.280 --> 00:59:46.159 And so, you know, that in a way kind of cuts us down to size, and then we get cut down to size even further, if you keep reading in Genesis, um, you know, with the fall, and the fact that, you know, keeping the earth and tilling it isn't going to be that easy, it's gonna be full of thistles and rocks and stuff. 00:59:46.480 --> 00:59:50.800 Um, and then we still try to go on our own. 00:59:50.960 --> 00:59:58.159 We start building technologies, so you get the technology of agriculture, and uh what happens? 00:59:59.039 --> 01:00:01.119 Uh Cain kills Abel. 01:00:01.599 --> 01:00:08.880 All right, and then you get the technologies of the civilization of cities. 01:00:09.119 --> 01:00:10.159 And what do you get? 01:00:10.320 --> 01:00:24.559 You get the Tower of Babel, and you get people that all of a sudden don't understand each other, they don't speak the same language, and I'm thinking, whoa, that sounds like Republicans and Democrats these days, you know. 01:00:24.960 --> 01:00:37.920 Um, with the way that AI has promoted little bubbles, the way that social media has divided us, given us misinformation on all sides. 01:00:38.239 --> 01:00:46.320 It's like, oh my God, it's the Tower of Babel again, and suddenly we're not speaking the same language and not understanding each other. 01:00:46.559 --> 01:00:50.880 And we keep trying to go it alone until you get to the flood. 01:00:51.199 --> 01:00:59.039 And then finally, you get the first covenant that God makes with Noah and Noah's sons. 01:00:59.280 --> 01:01:02.559 And God says, you know what, you can't do this alone. 01:01:03.039 --> 01:01:04.639 You gotta do it with me. 01:01:07.280 --> 01:01:13.760 I think we have just a couple more minutes before we go, before I then switch back ears to asking you questions that will be you have presented. 01:01:13.920 --> 01:01:28.480 So um I uh wonder if there's any last thing that uh you'd like to say and maybe that has come to mind in this conversation about AI and particularly the theological approach to AI. 01:01:28.559 --> 01:01:29.840 I'll I'll pass it to you first. 01:01:29.920 --> 01:01:30.559 Is there anything? 01:01:30.719 --> 01:01:34.000 This is just a one of those open-ended questions, grab bag questions. 01:01:34.320 --> 01:01:37.599 Oh, I think I'll wait and see what you guys want to hear out there. 01:01:37.840 --> 01:01:38.320 Yeah. 01:01:38.480 --> 01:01:41.280 Yeah, I think um so you're you're ready just for the questions. 01:01:41.760 --> 01:01:45.039 Yeah, let's let's move on and hear what the questions are from the audience. 01:01:45.280 --> 01:01:45.840 That sounds good. 01:01:45.920 --> 01:01:54.960 The only thing I I think I would add um to this is I I I want to give confidence to the church and to the theological traditions we have. 01:01:55.199 --> 01:02:06.960 And I said it earlier just a little bit, but I want to say more strongly that in a way I feel like this the solution at the community level and the Christian community level is actually relevant to AI, to the problems of AI. 01:02:07.119 --> 01:02:09.440 And not all not all of AI is only bringing problems. 01:02:09.519 --> 01:02:21.519 As you said, AI can bring some good things as to the tool, but with the problems that it might bring in terms of relationships, our own social, uh psychological, even sexual well-being. 01:02:22.000 --> 01:02:31.920 I believe that we have a strong tradition in Christian spirituality to go and utilize AI well if we take up those spiritual practices. 01:02:32.079 --> 01:02:37.679 And so that's the degree of confidence I wanted to make sure put is put in the mix. 01:02:37.920 --> 01:02:43.760 A lot of people that I read about five to one are pessimistic about AI versus uh optimistic. 01:02:43.920 --> 01:02:55.599 And there's some reasons for that, but there's also, I think, a confidence we can have and hope we can have uh within the church and and within our own traditions of Christian theology to say, let's use that. 01:02:55.760 --> 01:03:07.280 Let's use silence, let's use community, let's use work, let's engage in worship as ways to be healthy, even in an age of AI that could distort our loves. 01:03:07.440 --> 01:03:09.840 Let's see if we can keep our first love being the love of God. 01:03:10.159 --> 01:03:12.480 And I do want to now add something to that too. 01:03:12.960 --> 01:03:24.239 Um, one thing that Christianity, I think, brings to the table of world religions that is unique is an emphasis on our physicality, on our bodies. 01:03:24.480 --> 01:03:34.480 You know, I mean, we believe in a God who took on human flesh in order to have a fully authentic relationship with us. 01:03:34.960 --> 01:03:40.480 And when you think about Christianity, I think first about our major feasts. 01:03:40.639 --> 01:03:41.360 What are they? 01:03:41.599 --> 01:03:42.800 The incarnation. 01:03:43.119 --> 01:03:51.760 God takes on a physical body, the resurrection, God is resurrected, Jesus is resurrected in a physical

body. 01:03:52.000 --> 01:04:00.800 We say in the creed that we believe in the resurrection of the body, and this sanctifies our physicality. 01:04:01.440 --> 01:04:37.760 And we also uh underline the physical connection we have with the rest of the world through the sacraments, through the water of baptism, we show the importance of water, the holiness of water, our necessity for water, the life-givingness of water, and through the uh Eucharist, the bread and the wine that we have at communion, we again show how we we are fed by physical elements, you know. 01:04:38.239 --> 01:04:43.039 Um and so we have a very physical faith. 01:04:43.440 --> 01:04:50.480 We don't always talk like we believe that, you know, when someone dies we might say, Oh, her soul went to heaven. 01:04:50.880 --> 01:04:54.719 But that isn't what true Christian doctrine says. 01:04:55.039 --> 01:05:09.440 It says, No, for us to be unique beings and to fully encounter one another and ultimately to fully encounter our God we need to be physical. 01:05:11.519 --> 01:05:13.920 That's a good place to wrap this part up. 01:05:14.000 --> 01:05:16.800 And I think uh Joy, do you have something you want to do? 01:05:16.960 --> 01:05:31.840 All I want to say is the questions are open and so they will come up here on the screen if there's one there that you see that you like and want to say, raise that to the top, put a thumbs up on it, and Greg will start reading off some of the questions and get you started there. 01:05:32.239 --> 01:05:33.360 That sounds good. 01:05:33.679 --> 01:05:40.320 Um I know they get they get ordered based on how many likes and so on there are. 01:05:40.480 --> 01:05:42.480 So there uh there's a couple. 01:05:42.639 --> 01:05:48.800 Well, there's all there's there's there's one on um just coding. 01:05:48.880 --> 01:05:50.320 I I guess that came out to me. 01:05:50.400 --> 01:05:56.320 Like, say a little bit more about coding because I think a lot of us don't have the well, we don't have PhDs in computers. 01:05:56.559 --> 01:05:58.800 Yeah, well, I I've taught coding for years. 01:05:59.039 --> 01:06:08.639 Um the The thing we look at right now is people say, well, the one place where a generative AI is really succeeding well is coding. 01:06:09.199 --> 01:06:13.679 And I would say it's succeeding at the first third. 01:06:14.239 --> 01:06:23.760 In other words, when you write a computer program, the first part of your job is to write the program, to write the code. 01:06:24.079 --> 01:06:32.880 And now we have moved to a higher level of abstraction, which has always been a process in computer science. 01:06:33.039 --> 01:06:42.480 You know, when people first wrote computer programs, they had to do it in zeros and ones, because that's ultimately what the computer understands. 01:06:42.719 --> 01:06:57.599 You know, and then we came up with um, you know, compilers, assemblers, and first people wrote then an assembly language, which was very close to the way the computer functioned, but at least you could use some words. 01:06:58.079 --> 01:07:03.760 And then we moved to higher levels of abstraction in our programming languages. 01:07:04.079 --> 01:07:12.480 Now you can tell Claude, write me a program that does X, Y, and Z, okay, and Claude will do it. 01:07:13.039 --> 01:07:16.719 But that's only the first third of programming. 01:07:17.039 --> 01:07:23.599 The second third is testing and debugging, which is often harder than actually writing the program. 01:07:24.400 --> 01:07:27.440 Um, this still needs to happen. 01:07:27.679 --> 01:07:33.360 So Cloud can write the program for you, but you still need to test it and debug it. 01:07:33.440 --> 01:07:42.800 And the debugging might actually be a little harder than if you had written the program yourself, because Claude may do some things in rather obscure ways. 01:07:43.039 --> 01:07:48.159 And so you now you have to suss out, okay, how is he actually doing this? 01:07:48.320 --> 01:07:51.039 You know, where is the where's the flaw here? 01:07:51.599 --> 01:07:55.440 And then the final third is maintaining code. 01:07:56.000 --> 01:08:09.840 Because ultimately, in any real-world setting, in a business setting, for example, things change, and you need to change your programs to reflect those those changes in the external world. 01:08:10.079 --> 01:08:22.960 Again, this can be made more challenging when you have programmed, you know, say using Claude, um, because Claude doesn't always write the shortest, neatest program. 01:08:23.199 --> 01:08:32.800 Sometimes programs are kind of broke, and you've got to sort through all that and figure out where to make the changes, how to maintain this code. 01:08:33.039 --> 01:08:46.880 So while you can say, yes, this gives us an even higher level of abstraction with which to do our programming, um, you have to also recognize that it still needs to be tested, debugged, and maintained. 01:08:47.119 --> 01:08:52.000 And that those are still, we're still going to need programmers to do all of that. 01:08:52.640 --> 01:09:02.560 Well, that's a good, that's a good word because one of the things that I feel individually about uh the development of AI is that my future son-in-law is a computer programmer. 01:09:02.640 --> 01:09:04.640 And so I want him to have a job for a while. 01:09:04.720 --> 01:09:06.560 So I'm glad that he may not be threatened. 01:09:06.720 --> 01:09:08.079 He's also been in it for a bit. 01:09:08.239 --> 01:09:13.760 But there is a question we have here is the economic impact of AI on the job market. 01:09:14.079 --> 01:09:20.079 And then I want to actually press into the other part of the question of how do we think about this theologically? 01:09:20.159 --> 01:09:24.720 And I would add, you know, as a community of compassion. 01:09:24.880 --> 01:09:47.039 And I'll just throw that in because I was talking with an AI developer who was saying to me, you know, as the director of AI and faith, one of

the things we have to do is to help the people in our network to make sure in their churches there's compassion for people who might be displaced uh or will be displaced in uh with the development or the rollout of AI. 01:09:47.119 --> 01:09:55.439 And so it seems again, it seems like a very simple thing, but that's a huge part of what it is to be a church is to respond to economic dislocation. 01:09:55.600 --> 01:10:02.880 But um, but the broader uh question of what do you see as the uh the economic impact of AI in the job market? 01:10:03.279 --> 01:10:04.800 There will be an impact. 01:10:04.960 --> 01:10:11.439 Um, just as there was an impact on blue-collar jobs, sort of with the first wave of of automation. 01:10:11.920 --> 01:10:18.560 I mean, I think even in my university when I first started teaching, um, every department had a secretary. 01:10:18.720 --> 01:10:26.079 You know, now since we're all doing our own typing and everything, um, you know, there might be one secretary for a whole division. 01:10:26.399 --> 01:10:39.600 Um, so yes, uh some jobs have been displaced by automation and more will be, uh, more white-collar jobs will be displaced by AI. 01:10:39.920 --> 01:10:47.119 Um, and we really do need to ask the question, well, then again, you know, what are people for? 01:10:47.920 --> 01:10:54.960 Um we work, we don't just work to put food on the table. 01:10:55.920 --> 01:11:00.000 Um some jobs perhaps may feel that way. 01:11:00.720 --> 01:11:06.800 Um, these would be the jobs that you might say, well, it might be better to actually have AI doing that. 01:11:07.039 --> 01:11:11.840 But I think, you know, most of us work out of a sense of vocation. 01:11:12.479 --> 01:11:16.640 We work out of a sense of um being needed. 01:11:17.359 --> 01:11:25.199 You know, um as um I'm trying to remember exactly who said this. 01:11:25.279 --> 01:11:34.000 Uh you probably will remember, you know, that um our vocation is where our deep gladness meets the world's needs. 01:11:34.640 --> 01:11:35.119 Who is that? 01:11:35.199 --> 01:11:36.399 We we knew I know I have an answer. 01:11:36.640 --> 01:11:38.000 Ah, it's right on the tip of my face. 01:11:38.560 --> 01:11:39.039 Frederick Biekner. 01:11:39.920 --> 01:11:40.079 Yeah. 01:11:40.239 --> 01:11:42.720 I thought there'd be a Biekner fan out there who's gonna go, that's Biegner. 01:11:43.119 --> 01:11:43.920 Frederick Biekner. 01:11:44.399 --> 01:11:49.439 Um it it also meets our own deep needs. 01:11:49.680 --> 01:11:52.239 You know, we we have a need to be needed. 01:11:52.319 --> 01:11:54.479 We have a need to be useful. 01:11:54.800 --> 01:12:02.960 Um, so when people just say, well, let's just give everybody a universal basic income and not worry about people having jobs. 01:12:03.279 --> 01:12:13.520 Um I think that we will lose something there in people having a sense of purpose, a sense of vocation, a sense of of being needed. 01:12:13.680 --> 01:12:24.479 Now there will still be plenty of jobs that AI will not replace, many of the service jobs, jobs um in which we we take care of one another. 01:12:25.039 --> 01:12:35.119 Um there will still be jobs of creativity, um, because AI is working with what has been, not what might be in the future. 01:12:35.520 --> 01:12:55.520 Um, but as you rightly say, it will be also the work of the churches to help those who find that their particular job has gone away and and that they will need to be helped into finding another role within our communities. 01:12:56.079 --> 01:12:59.760 Well, I was interested when you were saying there's some jobs that will hang around. 01:13:00.000 --> 01:13:09.119 I was reading an article, just read an article today by Adam Gopnick about the player piano and how everybody thought when the player piano came out, there goes piano players forever, right? 01:13:09.199 --> 01:13:09.359 Yeah. 01:13:09.600 --> 01:13:21.199 And it had this phase that it was about 1896, I think, is when it was developed in 20 years, it was really popular, but you don't see player pianos anymore except as a you know a curious museum piece. 01:13:21.359 --> 01:13:26.000 But people like watching a pianist actually play music, you know. 01:13:26.319 --> 01:13:31.920 Well, they do, and and they like the uh individuality of it. 01:13:32.079 --> 01:13:36.560 I mean, the player piano plays the piece the same way every single time. 01:13:37.119 --> 01:13:41.520 But a human being does not play it the same way every single time. 01:13:42.000 --> 01:13:50.239 They may play it quite differently depending on the mood that they are in, whether they're feeling pensive or whether they're feeling exuberant. 01:13:50.640 --> 01:14:02.159 Um we we want the communication that it represents by um the actual presence. 01:14:02.479 --> 01:14:09.680 And I think we are also finding that people are starting to say, you know, I don't want machine-made things in my home. 01:14:09.920 --> 01:14:14.800 I want an artisan-made chair that the monks made in their wood shop, you know. 01:14:14.960 --> 01:14:18.159 Yeah, and uh again, it's that feeling. 01:14:18.319 --> 01:14:25.359 I I want the small imperfections, I want each one to be an individual piece. 01:14:26.000 --> 01:14:32.800 And I want the same thing, you know, in my relationships with with people. 01:14:32.960 --> 01:14:38.239 So we want it in our art, we want it in our artifacts, and we want it in our relationships. 01:14:39.359 --> 01:14:53.439 I'm gonna I think we have time for one last question, so I'm gonna go with the most popular one, which is with all the recent developments in AI, what has surprised you most since you started this work uh, what is that, 35 years ago or so, or 30 years ago? 01:14:53.920 --> 01:14:58.720 Um, larger language models were a surprise, I think, to a lot of us. 01:14:59.039 --> 01:15:17.520 We did not realize that um merely uh pumping a ton of words and using statistical models could model so much of human thought and of human relationality. 01:15:18.479 --> 01:15:31.840 Um, you know, obviously this could not be done until we had the kind of chips that Nvidia is putting out that could handle the huge amounts of data. 01:15:32.159 --> 01:15:40.159 Um but and it could not have been done before the internet because we would not have had a repository of all of that data. 01:15:40.319 --> 01:15:43.680 So back in the 90s, those things were missing. 01:15:44.079 --> 01:15:49.920 And uh it's it's surprising how much generative AI has accomplished. 01:15:50.159 --> 01:16:01.199 So even though I'm continually saying, you know, it's not gonna get there, it's not gonna be everything they're telling you, it's gonna be, it's it's quite amazing what it has already been. 01:16:01.600 --> 01:16:18.960 Um, but I'd like to touch a little bit on one of the questions I I glanced at there that said, well, should we limit our use of it given the ecological um implications of AI? 01:16:19.600 --> 01:16:29.439 You know, I was in a faculty meeting with my seminary faculty this afternoon, and they were talking about coming up with an AI policy for all of us to use. 01:16:29.760 --> 01:16:56.079 Um, and one of my colleagues was adamant that he said, you know, given the ecological implications of using AI, as well as the biases, the privacy implications, and the fact that AI has been trained on all of our works without our permission, you know, so stolen words, in effect, you know, that he couldn't countenance using it at all. 01:16:56.880 --> 01:17:03.359 Um, I think that's a highly principled moral stand to take, and I wouldn't argue with it. 01:17:03.600 --> 01:17:10.800 On the other hand, that I think for many of us, uh AI has become almost unavoidable in many ways. 01:17:10.960 --> 01:17:18.319 Um, you know, unless you put minus AI into a Google query, you're gonna get the AI summary at the top. 01:17:18.560 --> 01:17:31.279 Um, but I think it is important for all of us to be aware of the ecological consequences, to be aware of the privacy issues and the bias issues. 01:17:31.439 --> 01:17:43.439 And so maybe to limit our use of AI, to be careful in our use of AI, not to treat it like a toy that we're just gonna play around with. 01:17:43.840 --> 01:17:54.159 Um, because we need to at least be aware that as we play around like that, we're using very precious resources. 01:17:57.680 --> 01:18:01.439 Let's thank them tonight for this amazing presentation. 01:18:01.600 --> 01:18:02.399 Thank you.