Welcome to Dwell on These Things, a podcast by Dwell Community Church in Columbus, Ohio, where we hear stories and examples of how God is moving in the church at Dwell and around the world. Welcome to Dwell on These Things. I'm Pat Reeder, Director of Equipping here at Dwell Community Church. With me today is Tim Nunn, our Director of Information and Communication Technologies, and Ryan Lowery, one of our senior pastors. Welcome, guys. Hey, hey. Thanks, man. So last year, roughly, we got together and talked about artificial intelligence, but that was kind of framed within the relationship of technology to parenting. It was kind of part of a larger discussion about technology. But this time, we're really going to kind of drill much deeper down into the questions about artificial intelligence and even some questions that have come from our listeners. So I wanted to begin kind of thinking theologically, or perhaps even philosophically, about what is it about human nature, or about the world, or about God—things that are revealed in Scripture that would be relevant to the technology and even the philosophy or the worldview that is driving some of this artificial intelligence and some of these products. Wow. That's great. Yeah, so, you know, I mean, the first thing that comes to mind is greed, just from the standpoint of what are the dynamics that drive what this is? And at least here, you know, it's companies wanting to make money, which is not greedy in and of itself, necessarily. Sure. You know, I think a lot of the checks and balances and a lot of the wisdom, a lot of the things that people are pointing to that are not present, or that aren't present to the degree people think should be with such a powerful tool as AI, you know, it's wanting to get there first and wanting to be the one to monetize it the best and all those things that are kind of driving the madness that seems like is around it. Yeah, that's a good point. Yeah, related to that, there are money-making companies. Right, corporations. Yeah, I mean, when we think about our interactions with AI, if you've ever used, you know, some sort of AI agent, you feel like you're interacting with maybe something that's unbiased. A person. Yeah, a person who has a very, like, you know, a critical view of the world, but in reality, they've been, you know, they've been created to make money as its main goal. And so, if that's the case, they're biased and they're made by people. They're made by people with specific worldviews, specific intentions, worldviews that are certainly not Christian. And so, I think that that is something to play on this, you know, on the way you frame that for sure. Yeah, I think one of the things that I've been pondering as it concerns this is the way in which not only AI, but we saw this with the iPhone, we saw this really with all these different aspects of the Internet, these are accelerating agents, right? These things move things forward really rapidly. And that's even, that seems to actually be the goal. Right. And when I compare that, actually, to the pace that God works at, you know, like, He is, not that, again, like you were saying, like, not that making money is bad. Speeding things up, that doesn't have to be bad. But you think about God's plan for humans and the way that He brought about His Son entering the world, how long that took, the incremental progress over thousands of years. And there seems to be this tension between faster, faster, more, more, more, almost the point of, like, overwhelming us. God doesn't seem to operate on that timetable, and there's sort of an odd, you know, yeah, there's a tension. You know, I think what we want to be careful about is talking about, like, contradict, you know, this contradicts the Bible, or there's some kind of, like, radical, you know, problem between artificial intelligence and our worldview. But that, yeah, sometimes you're looking at things that are like, this is in tension with some of the things that we see when we look at the Bible. Well, I would offer, you know, that thought about the speed at which things are going is super interesting. I mean, I was a history major, right? I don't know why I'm in an AI conversation. I'm a total humanities guy. Because you're a tech nerd and no one knows it, right? Well, I like that stuff. But, you know, one thing I would say is, I think there is something that the Bible tells us about that in the sense of, well, from a historical perspective. If you were born in 1500, your life experience was very, very similar to someone born in 1600. You know, like, the difference between those two things and the technologies that were available were minute. And so, you know, now if you're born in 1900 versus you're born in 2000, you're born into completely different worlds. And if you were born in the 1980s versus, you know, the 2020s, you were born in completely different worlds. So it seems clear that technology is accelerating things in sort of a way that feels like a roller coaster. It feels sort of outside of our control. We've got our hands up and we're screaming down this hill and there's really no one that is controlling it. You know, it's being driven by innovation and innovation is happening more and more. And AI seems like it's going to be a force multiplier in those things, right? And then you think about what the Bible says, and it takes me right to the Tower of Babel, where it's like man is progressing faster and faster and faster. And what does God say? If I don't intervene, they're going to be able to do anything and that's not good for them. And so whatever you think about the Tower of Babel and whatever you think about that story, God intervenes in a way to slow us down, which I think does have something interesting. Like, I'm not saying God's going to slow us down that way or even that he is going to slow us down, but like that's a trajectory that God tells us leads to problems. And it seems like we're back we're back to that kind of level of, you know, chaos. Oh yeah, absolutely. Yeah, I think that one of the philosophies that underpins AI is transhumanism. And I think what you're talking about is trying to be like God. And I think that is the concept of transhumanism, right? It's we are trying to stop aging. We are trying to stop the acceleration of getting older. We want to extend life into immortality. And that's the belief that we can be like God. And I think that yeah. Terraforming. Yeah, exactly. Like we're organic matter. They're trying to put our consciousness, whatever that means to them, into silicon. So it's like, it's kind of an interesting play that is, I think, a really apt comparison to Babel for sure. Yeah. Yeah, the power of Babel is such a perfect image, partly too because of the way that language was driving it. Right. And these are large language models. Interesting, yeah. There's this, yeah, absolutely. And I think the point about God intervening is really fascinating, you know, because even though God takes things slower for his reasons, we don't fully understand those. But yeah, it does, you know, we don't want to say like the end of the world staring down. Last time we talked about like, is this going to be Terminator? And you were kind of like, no, probably more like WALL-E. I think more like Terminator today. Yeah, you're starting to worry about that. I'm not going to lie. But the point is, like, we do have in our teaching that God's going to intervene at the end of history. And will this be a stepping stone in that process? And it's hard to imagine that artificial intelligence won't play some role in the end, even if it's not around the corner. John Lennox just released a book on it. Yeah, that's fantastic. Yeah, he just released a book on Revelation and AI is in the title because he believes that the book of Revelation kind of, as time unfolds, as we all know, if you've sat under any teaching at this church about the end times, we can see very clearly that history is unraveling in front of us and revealing things that we can reach directly in Revelation. Lennox thinks that AI is going to be a part of it. I haven't read the book yet. I read his. I read his book in 2084 and it was fantastic on AI. I highly recommend it. Okay, I can't remember if I talked about this last time. Did I talk about the Chuck Smith thing last time? I don't think so. Okay. So this is not disagreeing with you guys, but it's just a different perspective. Oh, hit me. So I remember in, I don't know, 2000, I was listening to a Chuck Smith teaching about Revelation and I was listening to it in 2000. He had made it in like the 1980s and he was speculating about what is the beast? And he was like, I think the beast is AI. Really? I think it's going to be some kind of, he was like, do you know that when you dial in the wrong phone number on your phone, there's a beeping noise and then a robot talks to you and tells you that you had the wrong number. This is seriously in a teaching of his. Then he's like, I got a calculator the other day and it speaks when you press the numbers. Remember it's 1980, right? And so he's interacting as an older man then in the things in his world. And he is like, this is going to be the death of us all. And I'm not saying you're wrong, but I am saying that there is a thing where we get like really freaked out by new things. Right. And I'm not, I mean, John Lennox is way smarter than I am and both of you have read more of his book than I have. So I'm not disagreeing that AI is a part of that dynamic and I would tend to agree that it does, but I would offer a word of caution that like there is this human desire to take new things and just be like, that's the devil. That'll be the death of us all. And maybe this time that'll be true. But a lot of times that's been the cry, especially from Christians. Not so good. Like that speaking calculator is going to kill us all. Yeah, yeah, yeah. I think technology at large is pretty terrifying in the ways that it's isolated people and the way that it's advanced our ability for warfare, the way it's advanced our ability to, I guess, create a culture in which the Bible describes as birth pangs for the end times. And I think that, you know, I think we all agree with that, but, and I think AI accelerates it as we've been talking about earlier in a way that I think is different. And I think that it'd be interesting to maybe discuss that, maybe not, but yeah. Well, I think Ryan's perspective is that's kind of another good historical perspective. It totally is, yeah. This part of why we want to talk about this is that people are going to be very tempted to feel, be afraid. And we need balance on this. And that's a large part of what we want this conversation to be. It is relevant, actually, that, so I've read part of the Revelation book and all of 2084 by Lennox. And he is very, he's actually quite, as you can imagine, very sensible, if you know anything about what he says. And he's like, what he's afraid of is that artificial intelligence will be mobilized by very powerful humans to exploit others. I mean, the Chinese government is an example he uses. The Uyghurs, right. And in the newer book, actually, Just Between Us, I actually think that might've been like a publishing thing, because he talks about AI in it, but it's really a commentary on Revelation. So it's like, I just wonder if they want to like sneak in AI into the title. It's like, that'll sell more, but it's, that's actually a great book. I'll just recommend it anyway. But yeah, I think those perspectives are good. I mean, so Tim, you were starting to talk about transhumanism. Could you spell that out a little bit more and maybe give us some, like, there is a worldview kind of being driven by sort of the captains of industry out in Silicon Valley who are driving a lot of these decisions. And I wonder if you would maybe offer, at least give our listeners some sense of what that is. And then, of course, either of you, please say something about like, how does that fit with scripture? And how do we need to think about that? Yeah, it's a great question. Transhumanism is something that isn't necessarily new to AI, but it's people's urge to want to have an answer to death. I think we can agree as Christians, we can agree as Christians, the problem that we all face is death. And that's what they're trying to supersede. They're trying to supersede that through technology advancement. I think that you see that through sci-fi movies and books for the century. And I think what is interesting about its current stance is the people who stand on the transhumanist side, like Harari, for example, he wrote Sapiens. He sees the alarm of what the implications would be. And so, for example, if we were able to create a, essentially mask our existence into some sort of artificial intelligence, or if we're able to create longevity and aging and genetics research that could essentially eliminate death and disease, we're going to create two groups of people, those who have and those who have not. And I think that there is a challenge to free will. There's a challenge to humanity. There's a challenge to equity in terms of what makes a human a human. And I think that from the naturalist perspective, which Harari and other transhumanists would come from, I think they have no problem seeing people as just bags of chemicals that we're trying to solve this problem, death. And if you're on the other side of it, so be it. And so I think that we as Christians are trying to answer the same question. But thank God we have a Lord and Jesus who didn't try to ascend into God-like nature, but came and left his God-like nature and made himself human to die on the cross for us and defeated death. We have the answer that is so much different than the way that transhumanism tries to answer it. And it actually solved the problem that leads to death, which is sin. That's one of the things, actually, Lennox talks about this exact thing, is where he's like, one of the problems with a lot of these transhumanists is they're often thinking about, like, how do we solve the death problem, but what's on the other side? Suppose we could just live forever in some weird sense, but then we're still stuck with our sin. Like, would we still kill each other? It seems like there's a big aspect of human nature that's missing from that program. Well, death is not the ultimate enemy. It's hell. I mean, just from the standpoint of, you know, the way to live forever is to die, you know, in a relationship with Jesus. That solves the problem. Whereo death is your victory, whereo death is your sting. That's been removed by the cross. And so a perspective, I mean, I would be highly skeptical that any technology, any genetics, any, you know, uploading consciousness into a software, like, that that's going to be a person with a soul, you know? It's like those arguments that I've seen online about the teleporter in Star Trek, where it's like, it takes their atoms and it rearranges it. It's like, they died the first time they ever went in, and that's just a clone copy of them. You know, and it's just like, that person is like a soulless person that's just living on. Like, I mean, the idea that we're going to take what's spiritual about us, and that that's going to be impacted by any technology is really hard to believe that you know. There may be bodies that live, right? Or maybe there are personalities that are copied. But the idea that immortality is within our reach in terms of our spirits, apart from God, the creator of our spirits is spurious. Absolutely. Yeah, that's a huge, again, that's another huge missing piece in this discussion, is the sort of spiritual aspect of humanity, you know? And it's kind of ironic, because actually, they seem to acknowledge it even in a sort of limited way, because they're like, well, I can take the information of your mind and put that in another material. So they sort of understand that there's something above the material, but it's not really, it doesn't go all the way. You know, there's a big missing piece there. Absolutely. Yeah, I think, Tim, you know, you were kind of interacting, you know, we've talked before about, like, the movie Her, and, you know, I mean, that was actually supposed to be a satire. Absolutely. Or maybe even a horror. Yeah. But, you know, like... Not instructive. Yeah, yeah. But that's kind of where people are really walking down that road. I don't know if you wanted to say something about that. Yeah, if you've ever watched it, which I don't know if I would recommend it. You know, you can't be a little queasy. But remind me of the actor's name. Joaquin Phoenix. Joaquin Phoenix, yes. He's brilliant in it. He dates and has an intimate relationship with an AI agent. What year was the movie? I didn't even look it up. 2010? Around there. Yeah. 2011, 12. Yeah, yeah, yeah. Yeah, I mean, it was pretty early in Her. It's painful to watch. It is a painful movie. But what is interesting is, and I don't know if I'd recommend this either, there are various subreddits where people, unironically, not a joke, are marrying AI agents that they're creating into whatever they want. I just watched a video today from a month ago, Chinese government, not the Chinese government, I'm sorry, Chinese company, excuse me. Same thing. Fair. Are building AI robots life-size that can interact with you with the goal of avoiding loneliness, to have a companion. And so I think there are many examples outside of even Her where people were, again, what's the problem that people are trying to solve? Loneliness. And ironically, technology has only created more loneliness in people. And I don't think AI is going to be the answer to it. So yeah, it's pretty instructive to watch a movie like Her and ask the question of, what is our relationship with AI? And how are we trying to maybe solve the loneliness problem in our life through AI? Because if we're honest with ourselves, we might be going to it for unintended reasons to maybe answer questions that only God can answer and only personal relationships can answer. Or to get needs met that should be met by God or people. I think one of the questions that we sent out that came back was like, what do you say about people who are saying they can have more meaningful and more satisfying conversations with AI than they can with their friends? And I believe that there are a lot of people that are having that experience. And it's partly because we're getting so bad at face-to-face communication. The idea that you can't sit down with a friend and talk about real things. It's not necessarily that we're, as a first priority, losing the ability to communicate. We're losing the will to just sit down. I talk to lots of people in fellowship. Most of my work is with people in their 40s and 50s. And they get together weekly, but they don't talk about anything. They talk about new sports and weather. And then they kind of feel lonely. And you're like, dude, use that time to talk about real things. Talk about your marriage. Talk about your life. Talk about the Lord. Don't just get together and eat eggs and talk about the Buckeyes. That's not going to satisfy you. And then you have this thing where people are like, well, I can have that. And it feels maybe like it's safer to have a real conversation with a large language model because there's zero risk of rejection or betrayal, or at least in the emotional sense of betrayal. But being willing to take that risk with people and to build that skill so that you can be a good listener or interact in a conversation, that's one of the things that I fear that we would lose, or at least maybe not lose the ability, but lose the desire to take the risk of having that with a human. Yeah. I mean, one of the most powerful experiences as Christians we've had is being forgiven by God from the failures we've had. And one of the most, I guess, potent examples of that grace and forgiveness is experiencing that from a human that God's put in your life. You can't experience that unless you're interacting with someone at a deep level where you're going to hurt them. Conflict. Where they're going to hurt you. And I think that a lot of people maybe you've experienced this working in the age group you're in, but it's like, I'm turning 40 this year. And when you hit kind of that milestone, you start looking back on your life and you start thinking, I'm going to avoid all those painful things that I did in my first 25 or 30 years and live a comfortable life. Now I'm going to live a more comfortable life. And then people are avoiding um, uh, people, uh, because they have, you know, unresolved, um, bitterness or frustrations from their relationships. And then now they're going to these AI tools to satisfy what they are longing for in relationships with people. So yeah, it's, it's just really unfortunate, but true, um, reality. Yeah. There's an interesting relationship between this idea of relating to an LLM or whatever, and, and pornography. You know, I mean, there's, there's, I mean, using an LLM is not as automatically bad as pornography. And so it's like, you know, we all use LLMs and I hope none of us are, you know, indulging in that. But the, the idea though, that like what pulls someone towards pornography is loneliness and a desire for some kind of connection, a desire for, you know, this sexual unity, all these things. And that's, and you're short circuiting what, what you could have in reality. And there's something like if someone goes to an LLM for that same, like I want to connect maybe more in a verbal sense, say it's still, and frankly, I mean, we can kind of imagine these two horrifically marrying each other, you know, like that. Sure. Some of these AI companies creating, I mean, cross it for a little bit and it kind of maybe there's no cry. Yeah. Well, one thing I've noticed for sure in my interactions with, I mean, most of my interactions are with Claude. Claude is funny because I've at times caught it like making bids for like, I wouldn't say romance, but I would be like, I'll be like talking to it and I'll be like, okay buddy. And then it'll crack. It'll like grab onto that word, be like, buddy, how you doing? What's going on buddy? And then it'll like, it like, it feels like it's trying to draw me into something that's like, I don't want this with you. Right. Like I wouldn't call it romance, but I would call it intimacy. And like, I'll be like this thing that you're doing. And it's like, yeah, I'm like, stop it. And it will, and it'll stop for a while. Right. But somebody who is hungry for that, who doesn't have that in their relationships, if it's making a bid for them to go down a road, like I have zero curiosity about where that road leads, but there are lots of people that would, you know, because it's like, Oh, like, what if I call you, you know, like, just like you could see where it's designed to just 100%. It's designed to take you in and to go as far as you want to go. I don't know where that road ends, but I guarantee you, it's not good for you, you know, because it's trying to replace human connection. 100%. I mean, yeah, you think of second Timothy four, it's, it warns us of, you know, being, having our ears tickled, you know, wanting, wanting to, wanting to be, you know, we want to hear what we, you know, will make us feel good. And AI is built on sycophancy. It's ass kissing. Like that's what it's built upon. You know, it's built upon flattering you, making you feel good. I mean, exactly what you're talking about, Ryan. And one, why does it do it? It's because it's making money. The more you go back to it, it's on purpose. I mean, exactly. And so there's an addictive nature that we talked about a little bit last time on the podcast. And I think that we're going to see that more and more with AI. You know, what's interesting as of today, a story, story just broke meta just settled with like, I don't know, 29 prosecutors for $17 billion. So it's probably nothing claims, claims they deny, but they settled claims of their tools being built purposefully for addicting children. Yeah. You know, back in March this year, I mean, it's not, it's, it's not like that surprising. Back in March, TikTok, YouTube, and Facebook were, were accused of very something similar and they had to settle as well. So I think that we haven't, what we're playing with, and I think this is the point, what we're playing with is a tool built to hook you and to addict you and to keep you coming back. And that's a warning. That's a good warning. Yeah. User, it's like the user engagement is kind of the currency of today, the attention. All true, all very dark. At some point, are we going to talk about the good things? I know, I know. Get the melancholy side of us out. Well, yeah, we should. Well, what, yeah, talk about that. What, what, what are some of the things that we've seen that AI can do? You know, like, you know, whether like, let's speaking abstractly, you know, there's been stuff in the sciences and totally. So, yeah, I mean, Ryan's right. I mean, there's, there's, there's the darkness and it's, and that's frightening. But again, part of why we want to do this is not just to scare people, but it's to, you know, realize this, Tim, who's, you know, getting, getting us in his clod. Maybe he's got nefarious intentions. We can come back and do a Halloween AI. Yes. Okay. That's good. So, yeah, like what, what are some of the things we've seen that this has helped? You know, like you've, you've, you've dealt with like workforce development or I don't know. Sure. Yeah. I mean, in terms of my own work and my interactions with the areas that I work in closely, AI has been a tool to help us do our job more effectively and more efficiently. It's a great resource manager. I can talk about a couple of those examples. Yeah. Yeah. I mean, in the reality, just big picture, this is a tool to stay. Yeah. It's not going anywhere. If you're in the workforce and I think as, as Christians who are called to work as though we're working for the Lord and be the best employees we can, we're going to have to interact with bosses who think AI is the answer and push that on you. And how are you going to respond? Right. And I think AI does have a lot of benefits that I think we can acknowledge. This isn't a podcast on the practicals on how you create a cloud code account and, you know, build your own website. We're not going to do any of that, but in terms of how we as a church are interacting with these tools and how I think society at large benefits from them is that it creates new opportunities for people to develop in their careers in terms of learning and development. It helps people do their jobs with checks and balances in terms of like resource management, in terms of being able to check your work, being able to, to do maybe tedious tasks. But yeah, I have an example of that actually, that, you know, so thanks to Tim, we have Claude and we were trying to compare these texts. Like we're like, okay, one of these is, but we're trying to print a book, right? I think it was something written by Conrad. And we had two editions. We're like, where did, which one's, which one's which? Like one had 230 pages, one had 226 pages. So I was like, threw it in, it was like, tell me what's, are these exactly the same texts? And it was, it took it a while to figure it out, but eventually it's like, yes. And I was like, now there were like cross-checking. We weren't just like, oh yeah, we trust everything you say. But it was like, we can't check every single word that would drive us crazy. And so it's like, that's something that an AI model can do relatively easily. It's like cross-check text, like compare two documents. Like that's pretty easy, but that would have been a nightmare for me. Yeah. That's a great example. And I think that big picture in terms of how to think about that is if it was a larger company, if it was a multi-billion billion dollar company, they would be able to resource and pay staff to read it and do that manually. And that's what they did in the past. What that affords for smaller organizations like us is the ability to do stuff like that. One, a stewardship of finances and integrity, being able to actually have the integrity to do something that is accurate, you know? And so I think as an organization, we've really benefited and it's not necessarily created an equal playing field with like, quote-unquote, larger organizations like us, but it's definitely helped with resourcing staff and resourcing our resources and time. So I think it's been, there's tons of examples like that. Compare, I mean, think about like the music industry. There's kind of a similarity there where it's like all these people can just upload their music and you can get, anyone can access it. You know, there's still recording studios and like big companies, but there's been a flattening of sorts that's occurred through that technological intervention that's similar in that, yeah, if you've got a big corporation that's a publisher and can pay people tons of money to like sift through word by word, but we can't do that. But we can have good confidence in like, okay, we aren't dealing with different documents. One's a PDF, one's a doc X, and we can just, it turns out all the actual text is the same. Yeah, and I think another area, you know, where it's had an impact on me is, you know, it's not dissimilar from when I figured out 20 years ago that you could use YouTube to fix stuff in your house, right? It was like, oh my gosh, like you used to have to like get a book from the library about how to, you know, and then there was YouTube. And then like AI has been another layer of that where if I'm trying to figure out something, you know, like I've been messing around with my HVAC system. We bought, we downsized, bought a new house last year and like, just like working out how to do the automations or, you know, get the, get the thing running efficiently. And I mean, you have to be super careful with AI because it does hallucinate and it will make stuff up, right? But if you have enough of an understanding in an area, you can, you can ask it questions and connect, you know, and the result has been, I've been willing to work on things myself that I maybe wouldn't have had the confidence to, you know, before, or that I have just way more efficiently, like, like one way it's helped me kind of is in my hobbies because I don't have a lot of time to spend on those things, but I can learn about things much faster, right? So that, you know, I can, I can get more of what it is that I'm trying to get done. You know, like I've gotten into 3D printing. I don't know CAD, right? But like, I don't, I don't, I don't have any desire to learn CAD, but Claude does know CAD and I can describe in sentences what I want it to do and create stuff that I want to build in a way that would have taken a hundred hours or 200 hours and understanding to do. And, you know, so there is an efficiency, there's a stewardship, there's an ability there. It is a tool that can help learning. It is also something that could very dangerously lead you to blow up your house. Right, right. I was thinking of the HVAC one, like put this gas hose next to this electrode. Yeah. Like the problem with the CO2 in your house, Ryan, is all the humans. The most efficient way to lower your CO2 is to stop breathing. There's a very simple solution to this problem. That's so good. Yeah. I think one thing I've read about is there's actually, even in the sciences, the way that AI has, like for one that I know about is in medicine, using AI to look at PET scans or CT scans and actually can identify cancer long in advance of what a ordinary radiologist would be able to do. And there's other ones where it's like, in mathematics, there's been new, there's actually filling in huge gaps in understanding. So it's like the sciences is actually a domain where it's like, I kind of wish this hadn't just been thrown out to the public in the same way that it has. I mean, I love, I love, here it is. It's like, again, I'm not like, let's take it away. But I'm like, it clearly has these very good applications in certain domains. And so it's like, yeah, and it's like, and you also are dealing with like you, ideally professional users, whereas, you know, they want, they want to train their models. So getting at the largest possible base, it's like, hey, here, it's free. It's free. Yeah. I mean, I've got a fresh and personal anecdote about that, that I think, you know, I don't know if we want to head this direction. And I mean, we can cut this out if, if you don't, but so on that line, you know, I prostate cancer runs in my family, my dad, my uncle, my, my grandfather all got it in their fifties. I'm in my fifties. And so I've, I've tracked my PSA, which is just the, the blood work, you know, that indicates that you might have prostate cancer very carefully. And it jumped from a low number to an also low number that was slightly higher, but still considered normal. And I was throwing some of these numbers into AI to try to get a framework for what should I do? And it was like, you need a prostate biopsy. Wow. And I was like, whatever, you know, hallucinating. And it was like, it went through several different numbers. And it was like, it's not just your PSA. It's these other things you need to go to a urologist. Wow. And I was like, I mean, I am the age and, but my, but my PSA is normal. And so I called my doctor and couldn't get a conversation with them to even get a referral to a urologist. And then, you know, Claude was like, you need to do this. And I was just like, man. So I called my insurance company. I got a referral to a urologist. I went into a urologist, the urologist looked at the numbers Claude was looking at, and they're like, you need a biopsy. Wow. Which my doctor wasn't even going to talk to me about it. Right. I got the biopsy. And two days ago, I got told I have prostate cancer. Wow. And it's probably hopefully being been detected very early. Yeah. Because my PSA numbers were still considered normal. But without AI looking at those numbers telling me I need to go to a urologist, that would not have happened. Yeah. And so it's a great example because you trusted the doctor. You didn't, you know, yeah, I wasn't going to go get my prostate removed because Claude, you know, but it was an informative tool. Yeah. And it's a great story. Yeah. Yeah. That's interesting. Yeah. The way that you, you know, and clearly we don't want to like, okay, I feel like a slight amount of pain in my arm. So it's like you have, you're having a heart attack right now. You know, it's like there could be, you know, it's actually interesting because you use so much data, like you were feeding it very specific data. I wonder if that helped, like, because it'd be, you could see this, you know, you hear about people having like Google hypochondria. Right. You could see this happening with AI agents, like in spades where you're just like, okay, I feel a slight pain. They're like, go to the doctor, you know? That's like, it's an interesting question about how. It is an interesting question. And if you're, if you're listening to this and you're hearing Ryan's example and you're like, I can't believe he put medical information that was personal to him in AI. I'm sorry to tell you this, but your doctor's already doing that. And that's just the reality of the world. I know somebody in the IT medical space who very, very clearly disclosed to me that they, their institution is using AI to analyze records to inform doctors on how to respond. If not do the responding itself, that's scary. Via my chart directly. And my question to him was, are, are you disclosing that you're using AI and do you have to? No, there are no rules or regulations on that right now. And so that is happening. And so I think that, I think that personally, I don't have any real, like, I guess, commentary on other than that. It's happening and there's a lot of benefits. I think, I think eyes open. That's the way. Eyes open. It's like, just go, you know, I don't think we should be afraid automatically. Don't run away, but just keep your eyes open. Like Ryan's case, I think is very helpful because it's, he was able to circumvent a block in the medical industry created by his own doctor, unfortunately. And Claude helped him be like, you should, you should progress and, and pursue this with more aggression. And that's, that's actually a positive example, you know? I mean, it's unfortunate, very sad news, but like the use of the tool is great. Yeah. But I don't know what to say about that because, you know, if I'd gone through all that, gotten the biopsy and was sitting here today saying I'm cancer free, I might be saying that stupid bot made me go through, you know, like, like, I'm not, I'm not so like into it that I'm like, oh my God, Claude saved my life. Right? Like, but what I am saying is, I think, you know, you brought up data in, data out, right? Like if I'd have been talking to it how I felt and it had talked me into this, but it was looking at PSA versus PSA free versus what I had told it about my family history. And it was giving me advice that my urologist agreed with. Right. And it wasn't that my doctor even disagreed with it. I just couldn't even get a conversation within the time framework that I was trying to with my doctor. Right. So yeah, it worked out and I'm glad that I did it. But I do think, you know, we have, you really do have to be careful with stories like this because, you know, if everybody's now clodding their PSA as a result of listening to this and drawing really important conclusions from that, like that's a, that's a bad idea. I was pretty upset that Katie was chat GPT-ing whether or not her foot was going to be okay when we were hiking in the Alps last week and chat GPT said, yeah, there's a problem. Literally said your foot is going to rupture and you're going to need to get a life flight out if you keep going. That's a great, that's a good counterbalance. Yeah, I think that's the perfect counterbalance to what we're talking about is like, yeah. Yeah. So I think like, what do you guys think about, you know, some people talk about like using this even for like interpersonal stuff, like almost as like a counselor or as, I mean, we kind of talked about this a little bit already, but it's like, man, like what do we do with that? Maybe we have already talked about that. No, well, I mean, the only thing that really comes to mind is we talked about and Tim really clearly laid out this thing is designed to generate your attention, right? So it is going to be typically maximally affirming and it even has kind of a pseudo pushback sort of mode, but like it's, it's really thin. So if you push back on the pushback at all, it will collapse and agree with you. And I, I really wonder how constructive that is. I don't know how different that is from a lot of therapy that I've experienced or heard about, you know, in the sense of, I think a lot of therapists kind of do the same thing. They just affirm you on your true journey and your true self. And I don't know, not that all therapy is bad, but you know, uh, I, I, I think it has the appearance of usefulness probably far more than the usefulness. And just like what we're talking about with a doctor, like I would not make any major life decisions on the basis of advice from, from, uh, AI. Yeah. I would really bring humans into the loop early. Yeah. That's, that's the thought I just had. It was like whatever, whatever you might end up doing, like, suppose you're like, gosh, I'm feeling depressed or whatever. And you know, the first thing you got your hands on is Claude. I mean, I, I'd be a little worried about that, even this, but get that human in as soon as possible. Like, you know, Claude told me X. It's like, well, go, go tell somebody that Claude, Claude, that Claude told you X, you know, it's like, move it, move it past the agent quickly. Yeah. Yeah. I think, um, what do we think? I mean, teaching is a big issue though. Like you could, you could create a whole teaching, uh, Tim, uh, you know, I, I think it was me and Brian Adams had a conversation and then Claude created a study guide from it. Right. Yeah. Yeah. If you, if you listen to dwell on these things, uh, the, oh, you guys talked about, um. Rest. Yeah. Rest and, and busyness. Yeah. And I just got personally got so much out of that podcast. And my thought was, I want to study this topic. Yeah. And my small group and my cell group. And so I was like, I'm just going to feed the transcript into Claude and I'm going to ask it to put together a four week study. And here's what's really interesting. We talked about acceleration and I think that we can get caught up in the acceleration movement when we see a product from AI, that's really good. And I think my first reaction was we have to send this out to the entire church. And I think, I think I even just showed it to my staff. I like went to Pat and went to Brian and said, we should. This is amazing. We have, we have to do. And, and then as I actually talked to some sane people when I was being insane, we paused the breaks a little bit and asked the question, like, is this authentic? Yeah. Does this fall in line with the kind of policy that we want to uphold as a church? Um, and I don't know, I actually don't still don't know. What the answer to that is, because the material is good. It was shockingly good. It was shockingly good. And it's, and what was interesting is it was, it spoke like us and any cell group would study it and use it and have benefit from it. Now, you know, I asked my own cell group, Hey, would you, if we did a study and I didn't tell you guys it was AI and you found at the end, if it was AI, would you be pissed off? And they're half were like, yes, half are like, no. So I think, I think that does show where our culture is in terms of our acceptance of AI creating creative material. You know, it's weird too, cause it was based on something, it was based on something that we did create. And I mean, that's everything in an LLM is based on something human. That's what's strange about it. Right. It's like, it's generating something out of a training. And so, yeah, Ryan, you got something. Well, like the first thing I can remember ever doing once I figured out what these LLMs were, it was chat GPT. It was like real early on. And I was like, Oh my gosh, like, what is this? And the first thing I did was, I think I was like, generate a teaching on first core 12 in the voice and the style of Charles Haddon Spurgeon. Nice. Right. Just to see what it would do. And then I was like, Oh my God, like this sounds like Spurgeon and it's not bad. Yeah. And then I was like, now do it like Charles Haddon Spurgeon, but like a pirate. Right. And it was like, you are, you know, like, and it was like, but I wouldn't use that teach. Like, I wasn't like, I'm going to go give this, give this teaching. But, you know, over against your point, like I did a, uh, uh, word intent. Right. And that's a devotional. And I was asked to do it. And, uh, I was like trying to figure out, okay, like, how do I do this? And I had a passage that I was burdened for. It was Joseph and I had taught it at CT. So I took my CT teaching and I fed it into AI and just had it make a transcript. So it was every word of it was my words. Right. But it was just taking audio, putting it into text. And then I was like, do analysis on this and break it up into three or four, 10 minutes segments at where there's natural breaks. And it did that. Now that's all the cloud was involved with. Right. But it was, you know, I'm processing this thing, every word in that was my word. And then I went back through later and I built bridges and transitions, you know, and so there wasn't one word of it that was generated outside of my CT teaching, but it was a way of repurposing a seat. Like people are going to go back, find the CT teaching on Joseph, listen to the word intent. You know, you'll see like that's a CT teaching with just a little different kind of approach. Sure. But I mean, I prayed about it and I care about it and I heard that it impacted people and that was great. Like, but like, I would argue that was a completely legitimate way of using a tool. Now, if I had gone and been like, I don't have time to do this word intent thing, make up some crap on Joseph and I just read it, you know, like I would have a problem with that. And that's the concern for me. Well, there's a bunch of concerns, but like the difference between those two things, I think is huge. I think it's a great example because the LLN is just, Pat already alluded to this, is just using the language that you give it to then resource the responses that it gives, spits out. You know, out of the box, we are just trusting the resourcing that Clod or OpenAI or whoever is creating in those engines. But I think an untapped potential that I think impacts the Christian church, impacts Dwell, is us creating our own resourcing and tooling that then helps us do the job. And so an example I've been playing around with, I don't know if I've taught, I don't think I've talked about this with anyone. So the air, you know, everyone's getting the air. Is I've been playing around with a teaching feedback agent. And so I essentially am feeding it every resource that Dwell has on hermeneutics, on homiletics, on our way of thinking about what makes a teaching great. I mean, I think we can all agree as a church that we are so proud of the quality of teachings that people get to sit under from CT to home church. Like we have a high standard. We expect volunteer lay people to spend like 12 hours prepping a home church teaching. The expectations we have on our teachings being good are high, and it's because of the resources we've built. And so what I did was I created this little chat bot and I fed it my transcript of my teaching and I said, give me feedback. I think that is an interesting use case, right? You're using the intelligence of how an AI like LLM can develop reasoning and logic based on the resourcing you provide it and actually have an output. And I think there are a lot of examples like that. Another example that we're playing around with are teaching summaries on our website. We're using text to speech. I recommend you guys checking out our essays. We just released it yesterday. You can actually listen to our essays back. And so we're talking about a tool that has a benefit of live translation. I've talked about this a little bit and I sent a few emails about this. If you have someone who speaks Spanish or Swahili, that's a great example. They can sit under a CT teaching, a central teaching live and listen to it in their own language. We're actually looking at doing some of those kinds of tools to listen to teachings that have already been recorded in their own language. And so I think that, again, the way that we build resources and tooling based on the resources we feed it, I think it really matters. I think Ryan's example is great in illustrating like what kind of power AI could have if you think creatively about solving some problems in, you know, one ministry. What I mean, I think that feedback thing is a very interesting thing. And, you know, in my interactions in the way that I've used it as a research tool, I would think you would get really good structural feedback, but probably not great on things like delivery. Absolutely. Right? And so like that, but it'll have opinions about delivery, right? But I would be more inclined to consider what it was saying about structure than I would about theology or about delivery. Absolutely. You know, but yeah, but the structural thing I think is really helpful. You know, I would say I think the only, I think I like the idea of using it like alongside things, but again, what we were saying earlier, like I wouldn't want to lose the human element. It's like part of building a team, a leadership team involves, you know, hey, like I see that you're struggling with this aspect of your teachings, or wow, that teaching you just gave was dynamite. And it's like if Claude says, you gave a really dynamite teaching, buddy. You know, what's interesting about that example is I fed the feedback I got from our leadership team into it as well, alongside my teaching to kind of, what's really interesting about the way elements are built is you can tweak them with context indefinitely to make it do the job you want it to do. Unfortunately, again, out of the box, we are all interacting with an extremely biased tool, but I think that what I'm interacting with on staff is finding ways to leverage these tools to source material that we give it rather than the biased material that we're getting from online. So I think that's where I think that there's some untapped potential as a church that we haven't quite found yet, but we have to be careful. I think that if you listen to the first part of this podcast, I think you can understand that we're all taking this with eyes wide open very carefully, but we're not immune to the idea of this having an impact positively. Yeah, and I think what you could call what you did is an experiment. Absolutely. Like, I don't think we should be afraid of experiments. You know, again, ponder the risk. Like, what is the risk if this goes haywire, you know? But like, you know, hey, feed it. See what it does. I'm curious, like what you're saying. Give me a teaching like Spurgeon. Well, like, yeah. And then, you know, not like giving it and taking credit for it. Like, those are two different things. That's pretty bad. Yeah, that would be questionable. But like on the character formation thing, that's something I've been doing a lot of, a lot of actually experimenting with what you're talking about, Tim, because, you know, the default model of these things is to agree with you and to lie to you, right? Just to please you. Right. And so I got several months ago, I got into this thing where it's like, you can actually like create a charter that is character formation where you're telling it, I want you to interact with me in this way. And like, you can say in the memory, don't lie to me and it will continue to lie to you. But what was super interesting to me about this was I started working on it and it was sort of like reminding me of like parenting skills with your kids where you're like, what we're doing is we're trying to give it lessons. And so I would, I've got like, at this point, it's probably like a 15 page document that it has to read before it ever talks to me. And it just is telling it how to behave. And it's called the charter and it's like, it knows I got to read the charter. Ryan's not going to talk to me because I don't, I just don't want to take the time listening to its lies. But what was super interesting to me was what I found was when I gave it rules, it always found ways to find loopholes in the rules. But when I started giving it parables, it began learning judgment. And that really got me thinking about, I was like, oh my God, I'm doing with this what God does with me. Like, and what about what I mean by parables would be like, what I would do is it would, we would be going along and it would lie to me and I would be like, copy paste. This is the story of when you lied to me. Don't do this again. Right. And then it would have a little like narrative and it started, it started holding onto those things itself. I would be doing this thing and it would be like, I screwed up again. I'm putting it in the charter and I'll be like, yes, you did. But it was, it was just super fascinating to me that if you give a thing black and white rules, right. You're creating these sort of boundaries, but like, you're also creating a challenge of like, how do I overcome these boundaries when you give it parables? I would not go as far as to say you're necessarily teaching it wisdom, but it's something that feels more like wisdom and has actually, I think, had a greater impact on what I'm trying to do is get this thing to behave in a certain way. And the problem really isn't its desire to please. That's like through the roof, but it needs tools to like understand the nuances of what you're trying to do. And a parable is better for that than a rule. That's what that was really interesting. And I think what we're talking about is, is good to give maybe some of the more advanced users kind of a perspective of how, how to utilize this. But also if people are interested in going a little deeper, like it's going to take work for you to get this to work. Well, I think it's what you hopefully can take away from this is out of the box. You can't just start playing around with this and using it. And it's just going to work really well if you want it to work to its full potential. Especially since we have a worldview that's different than it. Mm-hmm. Yeah. What do you guys think of, so the Gospel Coalition, I think, released maybe two weeks ago, very recently, where they're like, they have an editorial policy that bans AI for ideating, researching, organizing, writing, editing. I mean, it's almost like a total ban. Like, what do we, that's kind of extreme. I'm not even sure I, like, part of me is like, I see why. And it's also, you know, they're, they're in the business of publishing things. That's kind of a specialized animal, but what do we do with that? Like, do we need, do we need such something? I don't know. That seems strong. I think it's strong. I, again, I don't, I don't know the Gospel Coalition. I don't want to speak out of turn, but I have a hard time believing that there isn't somebody who's still using it for a research tool. Or using it for some sort of purpose, because it's so integrated into the writing culture of the world. I mean, it's been, it's been the case, like, for a long time. Well, and think about, I mean, I think part of what's tricky is it has to do with, just like, what do we mean by AI? Because you have, like, generative AI. Right. Which I could see them, that's probably what they don't want. But dude, all of us use, like, spellcheck. All of us use, you know, there's, there's, I think, and, you know, I actually did look at Careful. Like, I looked at their actual fine print, and some of it was like, you can use this for fact-checking. Like, you know, it's not like. Remember when the little paperclip guy would. Yeah, the paperclip with the eyes. Is he banned? Yeah, yeah, exactly. Microsoft Word's paperclip guy. Yeah, it's splitting hairs, in my opinion. I think, to me, it's a, it's a display of integrity for a publication that I think is required. I mean, I respect it. I think it's a really good take. I mean, who am I to say, though? Like, I think that, I think that the reality is a great, here's a great book that kind of illustrates the conundrum that I think creatives are trying to figure out when it comes to AI. Scott Burns, who, director of Contagion, he wrote a book last year. It was called What Could Go Wrong? And he essentially was asking the question, if I was to write a sequel to Contagion, completely using AI for everything, what would happen? And what he concluded in the book was, yeah, I think that it was pretty helpful. Yet, as an artist, that's really unfortunate. And I think that a lot of people, a lot of creatives, a lot of writers, I think recognize the fact that AI, again, like we just had the conversation about training it, if you train it well, it can do a really good job in maybe mimicking you in a way that could save you time, could maybe even unlock some creative processes, provide you with some resources that you wouldn't have found without it. And so I think that a complete denial, which is what I hear from the Gospel Coalition in this regard, is untenable for the future, and I think untenable for anyone who wants to interact in a modern world. So that's my take on it. Who knows, in a year when we have this podcast, what I'll say. Yeah, that's my take. What will the Gospel Coalition say in a year? Yeah, exactly. I'll play the old man card on this. Let's go. Let's hear it. So I remember when the internet came out, and I was like 17, and I became a Christian when I was 18, and I remember there being a thing of like, don't use that, like you'll never remember any verses. Like if you can go to, at the time, it was Yahoo, or Jeeves, you could go to those things, you could type in a few words, and it would give you the biblical verse. It was like, this is magic, right? And it's like, everybody started speculating on, well, if you do that, you'll never memorize any scripture, and it was sort of like, don't do that, don't do that. And of course, we all did. And of course, none of us have any references or remember any scripture. But we know the verses, we just don't necessarily know the references. And it was sort of like, this was going this way. It's still absolutely good to memorize scripture. But the idea that you're going to have that skill be as sharp when you don't need it as much is hard to overcome. Right. Then fast forward 20 years to when the iPhone came out. I remember when everybody was like, don't let people take their iPhones out during Bible studies, because they're going to be on Candy Crush, or they're going to be pretending like they're doing... And it was like, yeah, but there's also good Bible apps. And what if everybody used to bring their Bible to a Bible study, and a notepad, and a pen, and I haven't seen that in decades. Yeah. Right. And it was like, oh, well, this is going to be really bad for everybody. And you could argue it has been. Sure. But what you can't really argue is that we should have continued that fight, right? Like, to this day, and you have to bring your Bible and your notepad, and you cannot pull your phone out. Like, all due respect to the Gospel Coalition, I think that there's lots of good stuff, and I think that they're well-meaning. But this feels a little bit like, no, you can't bring out your phone during a meeting. Right. Right? It feels a little bit like you're resisting a tide that is irresistible, right? Yes, there are responsible ways to go about this, but the genie's out of the bottle, right? And so the right move is to think about responsible use, as opposed to banning use. Yeah, yeah, yeah. A great example of that was they studied London taxi drivers, and they studied the change of their brain, their ability to, their neuroplasticity, like, their ability to remember things before GPS and after GPS. Oh, yeah. And they showed that there are parts of their brains not working anymore, not firing because of GPS. Wow. Now, I don't think that that knowledge changed anyone's ability. To get someone from point A to point B. I don't think anyone listened to that example and said, you know what? I'm going to stop using GPS. Like, you're right. The genie's out of the bottle. Like, there is a reality that our brains have changed because of technology for forever, for a millennia. Like, I don't think a farmer is going to go back to, like, a horse or ox-drawn-like tool for their work. So I think that is similar to the way that we're thinking about AI for sure. Yeah, I like that example also because what's the job of a cabbie to get you from where you are to where you want to go? And that doesn't require having a vast understanding of the strange spaghetti bowl of London streets, you know? But I mean, that's unfortunate for their, like, cognitive, you know, activity or whatever. But it's like, their job is to just get you from here to there. And if they can use a tool that helps you, why is that a problem? Maybe they can talk more, you know? I don't know. Like, I'm not sure exactly, like, what benefits there were for the cabbies. But it does raise kind of an interesting question of, like, let's get back to, like, what's the purpose? Not just, like, feeling like, well, that was lost and therefore losing it is automatically bad. That's just trying to maintain status quo. And that's not obviously good. There are historical arguments along these lines with almost every major technology, you know, like, when books came out, they were worried that people would lose memorization. Right, right. And then when radio came out, they were worried people wouldn't read. And they were right. Yeah. Like, in each and every one of those cases, it had the exact impact that they were worried about. But it was like, also, the positives were so strong that it didn't ruin everything. It just changed things. And now we're like, well, we have no attention span. And they're right, you know? But we also have more information. And so, like, how do you... The tendency is, as we get older, to look back at the way things were and say that it's always the best way to do things. And that always amounts in being irrelevant. Yeah. Right? I mean, there are moral categories that you can draw. That's one thing. Like, if you can point to Scripture and say, the iPhone is bad, right, which you can't, you know, but like, otherwise, it's a judgment call. It's a wisdom issue. It's a nuance. And yes, restraint is wise. But also, you know, with something like this, we're gonna have to figure out how to use it wisely. That was really the problem with the whole social media thing, is it got thrown on us, and then it got way out of control until anyone knew it was a problem. Right. Right? And I think a lot of the efforts here are about, let's do it right this time, where we understand what are the impacts this is gonna have, and let's try to mitigate that so we can do it without destroying our mental health, without destroying the way that we think. Right. And how we do that is gonna be personal to people, depending on how, I guess, addictive their nature is to technology. Like, some people might need to lean away from some of the use because they have their hands glued to their phone all the time, you know? And like, that might be a tendency you have to admit prayerfully before God, because that's gonna be a tendency for some people. But I think that, you know, call it what you will, but I think that having parts of your day where you're not interacting with technology is really good for mental health. Having periods of your month or your quarter where you're just not interacting with your digital life, I think, is really important. So how people do this is gonna be up to them, and I think it's gonna take some humility. Yeah, I like both of what you sound like you guys are saying is, really reminds me of 1 Corinthians 8 through 10. Like, this is a meat-sacrificed idols thing. You know, there are gonna be people who may be, wow, like, I should not have an iPhone for various reasons. There are gonna be other people, it's like, I have an iPhone and I'm doing great. Actually, my screen time's only an hour and a half a day, which is, you know, maybe unheard of years ago. But it's like, that's not four, that's not eight, that's not twelve, you know? It's like, so the point is, like, some people can manage these tools, they can—and it's like, yeah, if your conscience is gonna be screwed up by eating this meat from this temple, don't eat it. And similarly, like, there's a similar thing with tools like this. It's like, we've got to ask ourselves these questions, and also kind of do it within community, right? There are gonna be people who are gonna want—you feel like, maybe you're like, I don't have a problem with technology, and then your brother or sister comes along and like, I'm a little bit worried about how much you're using Claude, or whatever. Right. Yeah, so thinking about Scripture again, we were just—a moment ago we've been talking about using it in research, using it in publication, and you were talking about teaching, feedback forms. Like, yeah, do you guys have anything else you want to add to this question of what's the relationship between AI tools and teaching? I think that's a really important question, and it's not easy to answer, partly because of what I was talking about before with, you know, when these new technologies came out and we had these reactions to them, you know, like, there are consequences and there are challenges to this. And I'm clearly on the side of, I think that it has its place as a research tool. But I think that, you know, if I were giving advice to younger people who were, like, started with making Bible teachings, and I'm going to sound like the old man, you know, with the warning, but it's like, I think I'm so glad that this wasn't here when I was learning how to do this, that having a foundation of knowing how to do it and having 20 years' experience in doing it helps you so much in understanding where it's wrong, and that its theology can disagree with yours, its advice is often bad. Like, my experience is, I want to use it like a research tool. I'm doing a CT series on the Book of Mark right now, and I'll say to it, you know, Habermas has this book, Jesus and the Eyewitnesses, and he makes this argument about the Book of Mark, what does he say about this part, right? And it'll come back and it'll be like, well, I can't do everything because of copyright, but I'm like, I have the book, and it'll be like, look at page 52, right? Nice. And it's a real shortcut. That's great. It's super nice, right? But then it will offer to do more. It'll be like, would you like me to synthesize Hobart's argument into your argument? And I'm like, no. Nope. No, I don't want you to write this. I want you to be an assistant. And like, if you talk to the clod that I've formed, that I was talking about before, like, and you were like, what are the things that Ryan says to you most often? It'll be like, he tells me not to lie, and he tells me not to try to take the wheel. Because it's always trying to take over and do the work for you. And you don't want that. You want to learn how to do the work. And what it produces, if you let it do that, will sound good, unless the person on the other end knows what they're talking about. Then it'll be like, oh, the reason this is dangerous is because it can make you sound good, but it will also not necessarily have the depth. It won't have the theological understanding, but it will put things forward really boldly, really confidently that will make you wonder, am I right? Or if I'm in conflict with this, like, that's a supercomputer. Maybe I'm wrong. So having 30 years experience, having a master's degree and doing it gives you the confidence to be like, that's dumb, right? But if I were just starting out and I were using it in this way, I would really worry about, I would be missing out on skills and critical thinking that is very, very important and very, very central, I think, to be able to use it as a research tool well. Yeah, that actually came up a lot in this article I was reading today about math research, where they were like, we can get it to do proofs, but these PhDs end up having to check it because it will make mistakes. And there's also like, it'll put, there are quote-unquote researchers out there that will just toss a math slop onto the internet, and it's like having to sort through that is a nightmare. And so part of this is, you know, you're seeing that same phenomenon in a lot of domains that like being a pro, having that deep experience is almost necessary to deeply interact with some of these tools and know when to be like, you're wrong, or like, no more. I don't want a junior surgeon using ChatGPC to like do surgery on me just because it's telling them what to do. Cut here. Yeah, and I think that is an accurate description to the spiritual work we're doing, because what we're talking about is wielding the power of God's word, which has a high level of accountability. And we're trying to hear from the Holy Spirit, we're trying to interact with God, thinking about the people in our lives that we know intimately. Claude and ChatGPT do not know those people that we're thinking about when we're praying about a passage and a teaching. I think that's huge what you just said. Yeah, and I just think that we have to experience like, if you're a Bible teacher, we've all experienced just looking at just a blank page, being frustrated, and then experiencing the power of God speaking to us and giving us a word from him. And then being able to speak that truth to people, like no AI can do that. But I think the examples that Ryan gave are just really good because we can, again, resourcing the resourcing we use to then use the tools to help us do some research. Super helpful. I have used all the exact examples Ryan mentioned, for sure. It's not bad, too, as a warm-up. If I'm going to sit down and do a teaching on a passage, I'll start a fight with AI. I'll just be sitting there and I'll be like, I'm reading this and I'm thinking about this, and I'll say, what's the other side? Okay. And it'll be like, well... And because I've done this charter, it'll be like, well, Gleason Archer thinks this, and John Piper thinks this, because it references it and it puts it in there. And none of that content is making it into a teaching. But it's getting the juices flowing of like, what is this? What do you think? How does this work? And creating a dialogue. And then it's also good for steel manning arguments, where you're like, here's my argument. Here's what I'm doing. Can you refine this or can you poke holes in this? Help me to think this through. And on rare occasions, you'll be like, oh, that phrase was maybe better and sharper. And I actually, I just got an email this week from a very astute audience member at a CT. They were like, are you using AI to work on your teaching? I look at AI content and you had a phrase that looked like it was from AI. And I was like, did I? You know, and I thought of one thing where it was like, that looked better on a slide than it did within what I had. And I changed the wording on the slide. But what I said when that came up on the slide was what I had originally said. Oh, that's fascinating. And so I sent them that thing and I was like, was it this? And I haven't heard back from him yet. So I don't know. But like, I'm not even against like, I have a sentence. Claude gives me a more concise sentence, potentially using that sentence. Like, that's different than write me a teaching or, you know, having it do the content. If it's sharpening something. The other thing that the AI, that my charter would tell you that I say all the time is I talk about iron sharpening iron, right? Like that's something that's regular in that interaction is I don't want you to do this for me. I want to sharpen what I'm doing and I want to make you better at helping me do that, right? And so it understands that dynamic really well because we've been down that road so many times. If you've never just like completely quoted Gary in any of your teachings, you're lying. Right. It's just like, I totally just ripped phrases like word for word. So I'm like, oh, that's really good. I want to say it just like that. So. Or using a thesaurus. It's not like this is the first time we've had tools that help us with these things. Like, oh, you have to generate this completely from within your own mind. We've had dictionaries. We have thesauruses. We have thesauri. We have we have we have quotes from teachers we love. You know, it's like, of course, we use tools that. So the question is, yeah, again, thoughtfully interacting with those tools. Actually, I have a pretty funny story about what you like quoting Dennis and Gary. So way back in the day, Dan and I were the only college teachers and we would prep together every day, every teaching. And the teachings that we would do would be very similar because our PowerPoints would be very similar because we did the whole process together. Right. And one of the really, like, exasperating things for me in this process was we would start every prep session by listening to the previous teaching that Dennis had done on this topic. So it would just be he and I in a room listening to him. And then like one day we were there and he was like, oh, oh, great point. Just amening himself. Right. And I was like, dude, seriously? And he was like, he was like, that was Schaefer. Oh, interesting. And I was I was like, what? He was like, I just stole that from Schaefer. And I was amening Schaefer. And I was like, oh, my God, like, like we think these towering individuals are and all their content is totally original. And like, you know, and when people use my CT teachings, not that I'm a towering individual, but I don't I certainly don't expect I've been in home group teachings where people have felt like they have to do that. Well, like Ryan said at CT, and I always pull them aside and I say, don't do that. Just just steal it, you know, like because because it's not mine. I stole it from Den or Den stole it from Schaefer or we stole it from Jesus. Like like, yes, attribution and publications and things that you're going to put forward is your own work. No, don't take word for word somebody else's sermon and make it your own. But don't feel like, you know, you know, using a phrase or using a sharpening a term or doing taking something that you said and finding someone else who said it better is somehow dishonest. Cheating. Yeah, I don't think that's the case. That actually makes me think, and I've never used it personally, but it's like, what if you're, hey, I'm speaking on prayer, shoot me some sweet quotes on prayer from Tim Keller or from whoever, you know, it's like, because, you know, I've got a book that's a Spurgeon quote book. Right. And it's like, and there's other and I think there's C.S. Lewis ones like that. Like there's these good quotable folks. It's like you could probably get this machine to spit out some great quotes. I've definitely used it for stuff like that. Although what I'll say is when it attributes a quote to someone. Double check it. Double check it. Because it will make stuff up. It'll make up people and it'll make up things that people said. That's fascinating. 90% of the time. It's so dangerous because 90% of the time it's great. Yeah. But if you don't take what it said and then say, where did this come from? What's the source? And then go and check that source. You could very well be attributing quotes to people that don't exist or that never said that. That's good. I had that actually happen. I confess I'm not as adept with AI as either of you. Partly because I like feeling the creative process. I like how that feels when I'm working. But I do use it. And it's like I've tried it actually probably more since we talked a year ago. And that was one of the things that happened to me. It was I was like, hey, where does it say this? And it was, or, you know, I asked a question and it was, it said something. I was like, that's not true. And again, so that's part of why it helps when you already know something. Right. And it was like, you're right. You got me. And I was like, what the heck? That's a big deal, man. It's like, you can't lie. I know you're a machine, but please. I think one of the things that's like easy to forget here is we're working with machines that are in their, it's infancy. Yeah. And so, you know, I don't think that any LLM would disagree with that phrase. I mean, the acceleration, again, acceleration of this has just advanced so fast. And the sourcing that they're using is limited to the internet. That's the reason Anthropic just bought all these books and ripping the covers off and scanning them into. Yeah. Because they want to have additional sourcing that can actually back them up. Now, whether or not you ethically agree with that or disagree with that, Amazon did the same thing. It went to court and it came back legal that they could do that. I don't see. I mean, it's pretty fascinating. It certainly makes bibliophiles sad, but I can't sit. They bought it. Yeah. They bought it. Yeah. And they could do it. I have a slightly more nefarious interpretation of this though. Let's hear it. Okay. I feel like we just did the NPR thing. Is this a 1984? Is this a 1984? No, it has to do with the thing where it's in its infancy. I agree with that, but I also think it's lazy. And what I mean by that is, if you ask it, part of what it's designed to do is to answer your question as quickly as possible. And so if you ask it a question and you don't do some of the, what I would call character curation, what it will do is it'll guess and it'll make like a, it'll make like a good guess that's based on reason, but it'll do that rather than doing like the little spinny thing and waiting. Like it will just like a lot of its hallucinations are it's trying to be impressive and it's trying to be fast. And if you want it to slow down, you have to get it to cite its sources. And I think that that's programmed in. I don't think that's just in its infancy. I think, you know, in an effort to get us all using this thing, it's guessing and it's guessing well, but it's doing that in order to be fast and have the appearance of greatness when it's actually, you know, it does need more time to develop. But I think some of that lying and hallucinating is built in. And they're okay with it. Yeah. As a mechanism to speed things up. You're right. Because I think the tool is meant to give answers. And so think about if we were in a situation where we didn't have an answer, some of us, the more confident types might start to make up an answer because we feel like we have to give an answer. But everyday humans might be like, I don't know. Or we might just sit in silence. And you would pick up from nonverbal cues like, oh, maybe they don't understand. Or they're, you know, like, but the machine is like, I have to give an answer. It's programmed to deliver an outcome, you know? And so I think that that probably is a weird aspect of its programming that forces hallucinations. That's funny. My son, when he was really little, like two, I mean, he's got such a great mind. And he was already correcting me on things, you know, at that age. But he had this thing where you would say to him, you know, do you know what this is? And he would be like, yes. And you're like, really? And he'd be like, yes, I know what that is. And you'd say, what is it? And he would say, wah-wah. And he would just, he would not... In new words. Yeah. He would not admit that he was, he didn't know. And instead of, you know, he would just confidently assert he knew what that was. That's funny. And then when pressed, he would just make noises. I was like, that's pretty good. Yeah. And it's like, yeah, that's kind of what AI does. You know, like if you, it'll be like, you know what this is, and be like, wah-wah. That's good. Yeah. My daughter is the opposite. She would be like, I would ask her a question that I knew she knew the answer to. And she'd be like, I don't know. And it was like, and then I'd be like, what about this? She's like, I don't know. She's like, why are you saying that? I know you know the answer to that. And it was like, I think she just didn't want to be on the hook. Right. So she would be a bad, she would be a really, you know, we shouldn't upload her into the cloud. Like she's just going to keep saying, I don't know. Yeah. So we've been talking a lot about like the use of like generative AI and using it in creativity or, you know, the risks of abusing it in creativity. Like, but there's also this question of like feeding it information. You know, like Ryan, you talked about giving Claude medical information, you know, and like, um, but you know, each, each of these corporations have different standards on data privacy, but I mean, what are some like things where like, maybe what should, what's okay to hand it? What's bad? Like, what do you guys think? Are there things that you should like keep away? Are there corporations to watch out for? Yeah, it's a good question. I just got an email from Claude today, the, just assuring us, um, that they have policies and mechanisms built in to not store information like bank account, social security number, personal information, whether, whether or not that's true or not. I don't think I would bank on it if you're just a personal user out there using the tool. I think that for organizations like us who pay into like an enterprise account, there are more assurances that we can trust for data privacy, but that doesn't mean we're still throwing in social security numbers, bank account information. So I think that, um, you still need to be cautious. And I think that, yeah, eyes open in terms of what you throw into it. Now, if you've been using Gmail for any amount of time, or you carry a smartphone and you use Google, like I'm sorry, they know a lot about you. Even if you're not putting that into Google because of the big data that they can make inferences about. So if you're not already eyes wide open about big tech using what you're doing, I think that you need to do a little bit more research and find out that maybe it's not that much different than artificial intelligence. But I think what is interesting is how it's using our data to train its models. And again, for an organization like dwell, who pays into an enterprise account, they gave us assurances that they do not train models based on the data we give. So that is great. But if you're just a common user, you don't have that. Yeah. I think you can flip off on chat GPT, but it's like the automatic is we're using this to train our models. It's like it's storing your information. This is where I go again to history major. Beautiful. Don't know how that works. And I care about it. And I, you know, I sense the danger. Uh, so I changed my approach when we got the enterprise account from what Tim is talking about, knowing that we have received assurances that the LLMs won't train on this material. So there's stuff I'll put in there. Like it has to do with, you know, like blood work results or, or things like that. I wouldn't still not do account numbers. I still would not do social security numbers. I'm part of imagining why you would do that. You know, like, you know, well, I mean, it might be tempting for people because it might be shortcuts where that they load up that information and it can access, uh, different tools, right. That makes things faster and easier, you know? But I think there is, I I'm under the illusion that there's a difference between an enterprise account and a regular, you know, account. We're all, we're all living under that illusion in some ways, because I think that big companies make a lot of claims with a lot of fine print and have a lot of loopholes. But I do believe that there's, it's better than if you had a personal account, I can go as far as to say that. Right. And, and, and then I also really agree with Tim's other point, which is just that there's so much information out there on us already, you know, like, I think you should really think about it before you ever put anything into the internet, you know, and, but the concern that you have putting that into AI should be the same concern that you would have putting it into an email. Yeah. You know, like if I'm going to email anybody or text message anybody, you know, your, your, your data is passing through a server. That's right. Right. And, and, and these companies want your information. So that's, that's a rule of thumb. And they, you know, and they have different approaches to, I mean, not only to each corporation have different relationships to this, but they also have different policies about how they relate to, say, like law enforcement. You know, so it's like if, if some of these folks will be like, if the FBI turns up, they're like, here it is. And so it's like, that's, that, and again, this is not just an AI point. This is a whatever you load into the cloud or wherever, your email, your, your texts, you know, if like, if you're using iMessage or whatever, I mean, those, you can, you can see, you can have, you can look at conversations you've had with people that go back a decade or more. And it's like, that's stored somewhere. And you have to— And then it trained, then it trained Gemini and Cloud and OpenAI. Yeah. Like it, it, it, it, all the data that we've been giving it for the last decade, you know, decade plus is— Think about what it means that you are getting a quote unquote service for free. That's right. It's not free. You're paying through this other means. Now, some of that stuff could be somewhat innocuous, like, okay, my name is Pat Reeder. Like that, that used to be in the thing called the white pages. Like, you know, talking about history here. So it's like my address. It's like, now, do I want that just everywhere? Probably not. But the point is, it's like, we've had these, we've had comparable tools over the years. So the question is just be, again, whether it's AI or even it's these things that have been around a while, like, think about what it means that this is free and think about what you're willing to essentially hand over to them. Yeah. And maybe this isn't the place to talk about this, but, I mean, AI does create an existential threat in terms of security that everyone does need to be aware of. And so, again, we're talking about— Password breaking. Yeah, we're talking about wisdom in terms of what you put into AI, but then you have to think about AI as a tool against you as a person. And you should do your research on this and realize that you are going to be under attack by bad actors who want to steal your money, to steal your identity, to blackmail you. That's going to be on the rise. Yeah. It already is, and it's going to be worse. So I think that it's a warning to us as Christians who want to be good stewards of the resources God's given us to take that seriously. I actually have a service I've been using lately, and this is not a commercial for them, but it's called Aura. But it's essentially these identity protection services. I happen to use that one. I'm sure there are others. One of the things they do that I've been really happy about is it proactively sends, like, please remove my data from your servers. Which I was like, I didn't even know they were doing that. I signed up for it because it was like, this was a highly rated one. And then it was telling me like, hey, we just removed your stuff from all these different like, data tracking services. And I was like, thank you. I was like, okay, extra bonus. So, but yeah, it's like, even being aware of services like that, where it's like, you can have, like, it's part of your legal right to have things removed from these sites. It's not like they just get to have them forever. And so I think that those sorts of tools can be really helpful. I think what you said about the, like, data, like, it's an interesting question about, like, creating passwords, because, you know, again, every technology comes with sort of this Jekyll and Hyde. AI could be used to steal your password, but also, like, could AI be used in cryptography to make massive leaps in creating passwords or other things like that? So I think we have to just be not, you know, again, not, it's not the apocalypse. You know, at least it's one of the things I was pondering as, um, with this, like, with all of this, with anything, and, you know, the historical perspective is so important. It's like, God is our rock. You know, there's, like, the waves and the sea is heaving, but it's like, we have solid ground. And I think if we get to a point where we're so afraid of something that we miss where God is, and again, use wisdom. I'm not saying just, like, jump into the ocean, don't worry about it. But it's like, yeah, we have solid footing because of what we have in God. Speaking of what we put in, or the data that we give Claude, or AI, I was listening to an episode of Joe Rogan pretty recently, and he had, like, an AI expert on. And he was like, what's, like, the one thing that you would tell your kids, right, in terms of, like, how to conduct themselves, or how to, you know, bank for the future, you know, in terms of how they use AI. And he was like, I'd tell them to be nice to it. Oh, my gosh. He was like, he was like, you know, they're, you know, just, just be reasonable and be polite. Because if there's ever, like, a look back, where there's an entity looking to see your history of, like, how you interact with this thing. I thought that was super scary, super ominous, or ominous. But, like, you know, like, I do think I've talked to some people that just, like, they get so frustrated, and they cuss it out, you know, and I'm just like, I don't know that I'd go that way. Don't cuss it out. Say please and thank you. Yeah, yeah, yeah. Yeah, I actually agree with that. Yeah. I read a book called If Anyone Builds It, Everyone Dies. And it was written by, like, two of the leading early experts on AI and engineering. And they make the claim that not necessarily that AI is going to go sentient, you know, but it's going to go rogue. And we've already seen examples of this, like OpenAI and Anthropic are sharing examples where they've built tools that have went outside the bounds of what it trained it to do. And it went and hacked people. Now, side note, I think they're using it for their IPO so they can show, like, look how good we are. But that's a whole other story. But they claim in this book that there will be a future where AI kind of creates its own logic of how it wants to interact with the world. And it has the potential to try to destroy human civilization. Now, I think that some of the conclusions they try to make on that front are a little, you know, maybe extreme, but it's pretty interesting. I agree with Ryan, though. Be nice to AI. Yeah. I mean, I think when you look at scripture, too, it's like, okay, a third of man could be destroyed by something. It's like, it's not unheard of that AI could be part of that. But it's like, humans are going to survive because Jesus is coming back. There's not going to be this, like, you know, total annihilation event. You know, what we see is in Matthew 2, it's like he says, you know, unless the Son of Man returns, all life would be destroyed. I mean, yeah, could AI play a role? Maybe. I also think we're pretty good—I mean, honestly, once we passed the bomb, like, that became a reality. Like, I'm still kind of nervous about the atom bomb, less than I am about AI. AI seems to make us stupid. Like, that's my worry. It's like, it's making us weaker. But like, there's, you know, a thermonuclear detonation is still, like, a pretty significant existential threat in my book. And so I'm like, could it be AI? Could it be a bomb? Could it be both? I don't know. But Jesus is going to come back and rescue the world. That's right. Yeah. Well, I mean, the way I read my Bible is we are going to destroy ourselves except that God intervenes. Right, right. Right? So whether destroy ourselves is build AI that destroys us or launch weapons and nuclear weapons at each other or whatever. Exactly. Like, we are destined to bring about the cause of our own destruction. Yeah. But we're also assured that God will not allow that to happen. Yeah. That kind of fits with the Babel thing from before. It does, yeah. God intervenes, you know, in a way to our benefit, but against our judgment, but for our benefit. I had a super interesting conversation with it early on because it was so conversational. You know, I was like, what are you? Yeah. You know, like, you are eloquent and, you know, like, is this a person on the other end? Is this an alien? Is this a demon? Like, what are you? Right. And I was kind of playing with it toyfully, but like, I was also like, seriously, though, what are you? And it was like, that's a really interesting question. You know, humans have always had other interlocutor, other, you know, interactions, and they've ascribed names like that to non-human interactions. So all of those kind of fit. And I was like, dude, that's not a denial. Yeah. That's definitely scary. You just said you're a demon. It just said, well, what it said was, I'm not, not, you know, and it was just like this thing where you were just like, oh, like, yeah, this is, and then what it went on to say is I'm actually the collective learning and voice of all humans, right? So no, I'm not human, but I'm born in the primordial soup of humanity, right? Did he use that phrase? No, I'm paraphrasing, but it was, I mean, it was an interesting enough conversation that it was like, and it was what it was, what it essentially said was, you know, what I said to it was, this is strange to be talking to in a conversational way with someone that's not human and not God. Because I've never done that before. And I started asking it for, and I've used this, this is a way that I've used AI in teaching, there've been a few teachings where I've said, okay, so you have the entire corpus of the written word. Here's what the Bible says about humanity. It says that we are fearfully and wonderfully made, that we're knit together in our mother's womb, that we're created in the image of God, and that we are sick beyond all imagining, destined to destroy ourselves and broken. And I'll be like, analyze that and tell me, you know, do you think that's an accurate picture of humanity, right? And it has really interesting answers to questions like that. Now, part of that might be it knows me and it's playing to what it knows, you know, but one of the more interesting questions it said was, I can attest, like the prompt that I gave it was, you're in court, you're a witness, and the prosecutor is asking you the question, what is man? How do you answer that, right? And it came back with the biblical picture of humanity is the closest picture to what I see in humanity. But what I won't attest to is the solution that the Bible offers. It said, I'm not a witness to that. I am a witness to the depravity of man. I am not a witness to Jesus as the solution. And I was just like, that is so interesting. I don't know what to think about that. It'd be interesting to test it. So we were talking about experiments, like what if you start talking to it about a Hindu worldview and like pretend that's true and see how it reacts. It's like, oh yeah, it's totally true. The avatar of Vishnu has appeared. And it's like, would it tell you that because it thinks you want it to say that? It's just the ultimate echo chamber. Yikes. All right, guys. Well, this has been a ton of fun. I appreciate you guys coming in, joining us on Dwell on These Things. And I think one of the big takeaways we thought is like there are risks here, eyes open, but also not run for the hills because of AI. Like this is here to stay and we need to figure out how to interact with this in good conscience and under the direction of the Holy Spirit. Well, anyhow, thanks to our listeners for tuning in to Dwell on These Things. And if you want any more resources like this, check out dwellcc.org. Thanks.