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. Hey, everyone. And Ryan Lowry, one of our senior pastors. Hi. We're gonna be talking today primarily about technology in a broad sense, but we're also gonna hone in on artificial intelligence. This is a hot topic and seems to be coming hotter and hotter, and I think a lot of people out there are looking for some wisdom on this issue. As long as I've been a part of this church, Dwell has been very serious about adopting new technologies. We had a website in the mid-90s, very early. We were presenting our Bible studies using PowerPoint, and our website had mp3s for the iPod, right, as that was coming out. Do you guys remember the Real Media presentations? Oh, yeah. Those are top-notch. RIP. But we've also been tech-conscious. I remember hearing worries about internet pornography in the 90s and 2000s, and we've been concerned about the impact on relationships that the iPhone was creating. So as the iPhone was coming out, we were starting to observe some kind of bizarre relational behavior connected to that, people just zoning out on their screens. But we've never said, like, you can't use an iPhone, you shouldn't use the internet, you should never be on, you know, we've always been issuing caution, right? These are wisdom areas. So even though we've raised the alarm or we've, you know, exerted, excuse me, expressed concern, a lot of times it's like, hey, we need to learn how to wisely use these tools. Now, Ryan, you were a parent of teens and maybe even pre-teens as the smartphone usage was mushrooming, and why don't you tell us some about that experience? And I remember you wrote a great paper on that topic. I'd love to hear some of your thoughts during that time and what you've seen since then. Well, I've always been kind of a gadget tech-centric. I love playing with things and figuring out new things. And I was an early adopter with a smartphone. I had a Windows phone before the iPhone existed, just because I like to learn about that stuff. And one of the things that became really apparent early on was how consuming it could be, the impact of social media. So we were lenient with our kids on things like playing Call of Duty and like kind of violent video games that the culture was like, oh, that's the danger. But we spent more time limiting the amount of time, you know, and then also the issue of social media and what they had access to. And that way, because we felt like we were watching literally, you know, I remember putting my kids on the bus and every kid on the bus had an internet-enabled smartphone except for my kid. Wow. And he was five years old, you know, and so there was this issue of like, oh, like I can control things to a point, but now he's out in the cosmos. Right. What year was that? Oh, it would have been like 2005, 2006. Yeah. 20 years ago and still. Yeah. Yeah. Yeah. A lot of, I mean, I think your point about the bus is a great one. People are going to encounter all manner of technologies, you know. If we limit certain kinds of content, limit some kinds of media, you know, you're gonna find people, they're gonna see it at their friend's house, and you kind of have to like help give them the tools to figure this out, and also figure out at what age they can handle figuring out where those tools, you know, like, you know, like, oh, I know that that kid's parents put no limits on it, and it turns out he's looking at pornography. Maybe we're not gonna send our six-year-old over there. Mm-hmm. But you know, if it's like, okay, if he's 12, hey man, you might see this at your friend's house, what are you gonna do about that? So like, kind of helping grade out the age. I really think, I remember taking that class from Dennis years ago, and feeling like what I really appreciated was how he thought through and presented things in terms of like, this age is much more, you know, this age can handle this, this age cannot handle this, and working through that. But yeah, I, you know, I think also your point, I don't know if you want to elaborate on this more, but I really appreciated your, you know, there was a paper we read in that class where it was like, you say no to things that are really serious, and that allows you to then, you know, by saying yes to all these other things, it allows your no to really count. Yeah, for that it was, you know, we had a philosophy as parents that we wanted there to be a lot of freedom, and we wanted what we said no to to matter, to count, you know. But we wanted our kids to see that if they were responsible, we would give them more freedom than maybe what their peers had in certain areas, right, where they wouldn't necessarily have the same freedoms. So they could watch scary movies, or action flicks, or play Call of Duty, which their friends weren't allowed to do, but they had a dumb phone that just had like, call mom, call dad, and Google Maps on it. Yeah. You know, and... Snake. Yeah, they wouldn't, no, no, no snake. But, you know, that, that, they were able to see, their friends were jealous of them in certain situations because of the freedoms that they had, and that, I think, eased the the blow in the areas where they didn't have the access. They weren't allowed to have a computer, a laptop, a TV, a screen in their bedroom. Right. Right, but there was a centrally located family computer in the living room that they were allowed to use, but we were always around, you know, so there was never a private moment with a screen until they were out of the house. Yeah. You know, even when they were in high school, we, we, we had that, those limitations. Yeah, I think that's wise, and I really like what you said about having a household of freedom. You know, I think when you're, you know, screwing things down too tight, I think that, that can, you know, mess with, kind of mess with them, you know, in the sense of, like, kind of crush their spirit, you know, and of course even could, could create rebellion. So I think that's, that's something I'm very mindful about. I have a very strong-willed little guy in my house, and so I'm always like, okay, I actually have a strong no impulse, just in my personality, and it's something I'm constantly having to be like, is this where I need to say no? Like, actually, maybe I needed to go back and apologize and actually explain to him where, no here, but yes here, and kind of, like, channel, get a clearer yes up front. I also just think it fits with the way God is, you know. You look at, in Psalms, and I'm not gonna quote this perfectly, but he makes this comparison, he's like, I don't want to lead you like, like I would lead a donkey, like sticking a bit in your mouth and dragging you. That's, that's not the kind of relationship I want with you. It's ironic that parents clamp down on freedoms like you gave, Ryan, in terms of being out with friends, yet give freedom online where it's a lot more dangerous in a lot of ways, right? So it's like, it's reversed, you know. As kids, when we were older, like now that we're looking back being older, we had a lot of freedom to just ride our bikes wherever we wanted, and a lot of parents clamped down on that, but give them freedom online, which is not necessarily the best thing for them. Yeah. Go ahead. One of the things that we saw was that where we were different from most parents is most parents were very concerned about violence and their child's exposure to violence, and they were indifferent or silent about their exposure to sex. Interesting. And we were kind of the opposite. We were like, you know, we don't want them to be exposed to sexual stuff, but, you know, they can shoot Nazi zombies if they want. That's okay. Yeah, exactly. I think it's an interesting thing. It makes you wonder what the logic is behind that, because honestly, if either one of those is going to present a serious risk, it's like, your kid is eventually going to live in a sexual world and have sexual relationships. Your kid is never going to shoot Nazi zombies, you know? Like, this can just be play. Don't worry about that. But some of the, you know, well frankly there's like an abundance of literature now about the harm that exposure to pornography is having on young people, boys especially, but you know, even girls now as well. Yeah. Actually, Tim, your point about that kind of flip-flop is exactly the thesis. That's a central thesis of Jonathan Haidt's recent book, The Anxious Generation. He says, we've become too cautious with real play, and that has been very harmful for kids, and we've been completely incautious and have no boundaries with technology, and that's also been really bad. And it's like, these need to reverse. We need to go back to kids being able to get outside, get on their bikes, roam their neighborhoods, know there's not going to be someone to kidnap them, you know what I mean? Even though that was probably fairly improbable at the time that this culture of safetyism started to build, violent crime has dropped for decades. It looks so different now than it did even, you know, in the 60s and 70s when all these, a lot of the thinking like, you know, the parents who grew up in the 60s and 70s became very safety-oriented, and that had a huge cost. It's like they took away something that they had. I have an anecdote about that that I've shared with parents over the years, it had to do with some of the choices we made as parents. You know, like, I remember going to the playground when my son was, you know, four or five years old, and we would let them run around and be crazy on the playground, you know, and we would get there and there would be parents that were kind of micromanaging how their kid played. And I remember one instant where Logan was climbing up on top of the monkey bars and walking across them, right? And I stopped him and I just said, hey man, is that smart? You know, and he was like, yeah. You know, I'm like, you're four, you know, and I was like, you could get hurt doing that. And I stayed kind of close, but I didn't stop him, because I thought, I've made a calculation of like, okay, worst case scenario here, we get a broken arm and we get a life lesson, right? But like, he's being adventurous, he's being, and it's within this parameter of risk. And then, you know, other parents who were sort of like micromanaging and like, don't, Billy, don't do that, don't do this, you know, when they would be like, okay, it's time to go, their kid would throw a fit and smack their parent in the face, you know. With Logan or Lexi, you know, they had a lot of freedom there, and then we would be like, it's time to go, and they'd be like, okay, you know. And that was sort of the difference in our philosophy kind of manifesting itself right there on the playground. And the other parents would be annoyed with us, because we'd be giving our kids freedom, you know, and they would be looking at us like, we're these terrible parents that are, you know, putting our kids in danger or giving their kids ideas, right? But, you know, I would kind of look back at them when their kid would smack them in the face, you know, and be like, what are we doing here, man? Yeah, come on. That's wild. Yeah, actually, you know, my son was born in 2017, at the very end of 2017, and Coddling of the American Mind came out that year in 2018, so like right after he was born. That's another Hyde book, yeah. Yeah, exactly. Fantastic book, and part of that book actually helped me. I probably lean anxious, and kind of a little bit like, oh, be careful, son, you know, don't do, don't, and it was a good challenge for me to have this exact point you made, you know. He talks about, like, children need to be able to dose themselves on risk. That was the phrase he used that I, that clicked in my head. I was like, yeah, they need to feel in their own body what's too far, and what's, you know, what's okay, and that that's, yeah, I mean, so I thought that was, I mean, I'm glad to hear you have those wisdom principles, you know, maybe more naturally, you know. I had a book, I had to get told, you know, but Jonathan Hyde had to tell me, you need to let your kid, you know, you need to relax, you need to let your kid have fun, but I think, yeah, that was, that was really helpful for me at that time, and yeah, since then, you know, like my daughter, she is, loves to climb, and so she's going on the, you know, if you go to the playgrounds now, there's like, this is a two to five playground, this is a five to 12 playground, and so she goes, when she was two, she'd go to the five to 12, and she'd climb up these really steep, the steep ladder. I don't know if I want to climb that thing, and I was just had to watch her and be like, I'm gonna let her do this. Like, she's, actually looks like she has a great grip, and she's doing fine, and she has never fallen, and it's like, you know, how much, how much would I have taken away if I'm like, ah, no, no, no, no, no, no, please, that's too dangerous. So I think, you know, I'm, you know, obviously as a young dad, I'm still learning, but it's like that, that was a really important message for me, and in hearing this, you know, I just finished, excuse me, I just finished Jonathan Haidt's newest book, The Anxious Generation, and kind of seeing that, okay, let them play part, and then there's this counterpart of, like, put the brakes, serious brakes on some of these things like social media, things like, and just having access to smartphones everywhere. So today we're, you know, one of the things that got us talking about technology, or why we were inspired was, you know, we wanted to think about AI. You know, AI has, has been around for a while, but, you know, at least part of why I wanted to frame it within this larger question of technology is that at least a lot of these exact same companies, or at least the same culture of Silicon Valley, is generating these tools, these AI tools. And, you know, we see, we don't see a lot of reason to, to trust these companies, you know. I don't know if you want to comment on that at all, Tim, just like some of the, some of the things we see from these companies is just treating the user as the product. And I don't know if you have any thoughts on that, or if you're... Yeah, I mean, you're right. Yeah, the user is the product. Like, we're not paying for any of these products. I don't know if you guys remember back, like, it was probably 04, 05, when Gmail came out. Yeah. And it was unlimited email. Your email inbox will never go empty. Yeah, yeah, yeah. You'll never, you'll never run out. Free. Never, free, forever. And you're like, what's the catch here? Yeah. Like, well, one, like, your email inbox now is full, and you have to delete stuff. I don't know about you guys. Yeah, yeah. Now it says this many gigs. They lied to us. No, but we, we are the product. Like, they're selling our data. And so they're making money off of us. It's interesting, referring to Height's opinion, you know, there's, there's one side of the coin, which is users have to take responsibility of how they use the tools and know the dangers. But he believes the real solution is pushing platforms to change how they configure and present the tools to us. Yeah. And I think there's a lot of validity in that. More transparency. Yeah, I think part of that, but also we know that these tools could do identity verification, for example, which they don't, you know, you can just say you're, you know, you're 13 or older to use a tool, yet there's probably like six or seven year olds out there creating their Instagram accounts on the phone that their parents gave them. And so what responsibility do parents have over that versus platforms? I mean, that's obviously a tricky conversation. Obviously, parents are responsible. And as you guys, as parents have talked about, you take responsibility over what access your kids have, but platforms also have a role to play and how they do identity verification. And so I don't know what the solutions are, but we do know very clearly they're not doing enough to protect people. The industry itself is not regulating itself. Absolutely. And so, yeah, there is a- And it's a dollars and cents thing, you know, the more users they have. And, you know, when you think about- Longer engagement. Exactly, exactly. Yeah. And you think about false information being spread, that isn't new to us. Some crazy stat on Twitter was like, I don't, I'm not even going to throw out the number, but some high percentage of Twitter is just bots. Yeah. Sending out false information. I actually heard about that recently. And so the question that we have to ask is, do platforms have the ability to verify whether or not it's a bot? Yes. Yes. They do. Yeah. Are they? No. Yeah. So what does that leave us with? I mean, it leaves us with an interesting conversation about, can we trust this information and can we trust these platforms, you know? Right. So that's worth putting on the table for sure. Yeah. Yeah. And I think, I really appreciate all you said there. And again, these are the exact same companies that are creating these AI tools. So, and a lot of times there's kind of a default, you know, the default is like full information sharing all this on these services, but it's like Jonathan Haidt was talking a lot about how we need to actually pull back as users, be like, one, also maybe even insist that the default is not sharing, especially because of these kids. The other thing that's wild too is that you mentioned 13 years old. There is only one federal law regulating the internet and it was written in 1998. Right. That says something about a kind of dereliction of duty on the part of Interesting. United States government as well. Like we have safety belts, we have airbags, there's all these things that like you get in a car, you don't want to die. And there's probably like parallel, I mean, we don't want to just, again, going back to freedom, we're not trying to take cars away. We're not trying to like get people off the road, but how can we get them in the car and get there and not like get skewered by the, you know, the metal in the car or whatever, like a nasty crash. There's got to be parallels with social media, with AI, whatever. Ryan, it looks like you had something to say. No, not really. I mean, I have a different perspective from the standpoint of, I think regulation is necessary, but it should be minimal. I hate that it's illegal to drive your car without your seatbelt. Like, yeah, I think that should be a decision that people make for themselves, at least adults make for themselves. But that's not to say that I think that the internet shouldn't be more regulated, or at least like you're talking about, Tim, I think, you know, identification, verification, so that there's certain content that just aren't available to people that are certain ages. But I also see the other side of it of, you know, I think individual responsibility is a lot more important than federal regulation. Yeah, and certainly, I mean, part of it is there's kind of a triad here of the people making the product, the consumers, and the government. And it's like, where the weight lands is probably, yeah, obviously, I don't really love censorship. I don't really love, like, you know, the government having their hands all over things. And it's like, yeah, if somebody wants to fly out of their windshield, that's their business. But at the same time, we're like, you know, actually, maybe another, maybe a better parallel would be like, you got to be 16 to get in a car. And so it's like, you know, you don't just hand your five-year-old keys, or, and that's kind of what's happening with the smartphone. It's like, hey, why don't you just hop behind this, you know, massive piece of metal and cruise down the street? It's like, the risks to that child or, you know, I don't know, frankly, I don't, in the case of the internet, the risk is much greater on the child than to anyone else. Like, a child driving down the street could also, like, hit a tree or hit a dog or horse or what, you know what I mean? They could harm other people, but it's like, it's probably more dangerous in the case of the smartphone. Like, the user is in, probably in greater danger in that case than, like, yeah, driving a 2,000-pound metal box down the street, you know? Well, I mean, there are real things. Like, when I was a kid, you know, we had VHS tapes, right? And there was one bootleg VHS tape that was more highly sought than any other, and it was called Faces of Death. And it was, like, this video of compilation of people, like, actually dying in newsreels. And it was impossible to find, it was impossible to get, and it's like, one of your buddies got it, they made a copy of a copy of a copy, and then you watched it and you regretted it for the rest of your life, you know? But it was, like, so hard to find that. Now, that's Twitter. I mean, that is available, and then some, to anyone, everyone, everywhere, all the time. And, you know, I remember the impact that, like, finally succeeding and watching a few minutes of this had on my young psyche, and I was just, I worry about what kids are being exposed to, you know, for sure. Because it is changing the way people think about the world. Yeah. Yeah, absolutely. So let's pivot over to AI. I actually wanted to know, you know, we can start with, like, an overview of AI and maybe even its benefits. Like, I don't think, as much as I want people to be conscious, to be aware of what they're facing, I also don't want this podcast to be a, like, stay away, oh, be careful about the artificial intelligence. You know, we actually need, this is going to be a tool, this is the future. And if we get in the way, you know, if we say no, if we get in the way, like, we're the ones that are going to get, you know, knocked down or left behind. And it's like, I think we have to judiciously interact with it. But I guess I wanted to note that I think many have the impression that AI is kind of new, right? I think partly because I think there's a conflation between AI, which is a very broad concept, and generative AI, large language models. So Tim, I was actually wondering if you could just sort of give us a working definition of AI, how long it's been around, what are some of the newer things that are coming out? Sure. Yeah, yeah. I'm not an expert in this by any means, but I think if we look back through the history of it, it started in the 50s at a university in Dartmouth, where they coined the term, and they looked at computational models and said, can we train these to mimic human behavior? And that's really the basic definition of AI is taking large data sets and using them to mimic human behavior. And so our basic use of it, probably if you know, depending on how old we are, is how amazing it was that Microsoft Word could auto-correct your language. It was like, oh no, I misspelled a word, use auto-correct. Grammar. Grammar, yeah, thank you. And so that's kind of the basic idea behind AI is, can it do human operations? And the answer is, yeah, it can. And there's a lot of benefits to that. I actually just got back from Phoenix and experienced, in my opinion, probably one of most amazing uses of AI, and it was Waymo self-driving taxis. And they were incredible. That's awesome. And yeah, they work really, really well. And so when we think of AI, just like any other new innovation or technology, in hindsight, it all makes sense. We're like, oh yeah, the yellow pages don't make any sense anymore. Those don't exist. Can you tell us what the yellow pages are? There might be listeners who don't know what that is. So we have a lot of hindsight bias, where when we look back to the technologies we utilize, and I think AI has been in existence and we're already accustomed to it. For example, if you use Google Maps, it's going to reroute your traffic depending on what traffic patterns are. That is AI. And so I think anything that mimics human behavior and a computer's doing it and computational models are doing it, that's what AI is. Yeah. I think that's important to lay that out, because I think people get weirded out by AI. They just hear AI and they're like, oh, Skynet. So we've been using AI to send texts for a long time. When you're swiping in that pattern on the screen of your phone, it's guessing what words you're using. That is artificial intelligence. When we talk to our phone, hey Google, or hey Siri, or okay Google, whatever, it's predicting based on this audio pattern what words you're trying to use. And so I think, yeah, so tell us a little bit about what's kind of what, tell us some about, you're right, you got to give us a little bit of history there. Can you kind of catch us up on what are these newer tools? Sure. What is it that's like, what is new and why are people talking about it so much now? That's a good question. I mean, ChatGPT feels like it's just an everyday tool that we use, like at least a lot of us who are in the workforce. Yeah. But they started not even three years ago. It was like November 2022. And so some of these more generative models that allow us to pull from a large dataset and compile information and curate to that to us is pretty new. And so Google Search, I mean, has been using AI to index the internet and dish us out information that will lead us to websites for a while. But what's new is the curating of those datasets into summaries, into compilations of data. So, if you've used Google Search and you've seen the Gemini AI summaries at the top, we're kind of moving into a new era of the internet, which is pretty interesting because one thing that in my area of work and on the web side of things, search engine optimization has been a very interesting conversation. Like, how do you present your content in a way that Google indexes it and gives it to people when they search topics? And as a church dwell, our website launched in 96. And so we have a high domain authority on the content that we're experts in authority of. That's becoming irrelevant. We actually saw a decrease in what's called a click-through rate. And so when people search terms and our website shows up, people aren't going to our website. What they're reading is the AI summary at the top. And so we are in a new era of web. We're in a new era of digesting information. And I think we're in a dangerous area because we're trusting these AI models to pull the best, quote unquote, best information and the most valuable information for what I'm doing. But it might be leaving things out. And so I don't know what that future is going to entail. But that's kind of where we're at. Yeah. Actually, I'd love to hear, you mentioned using it in the workplace. Can you tell us some of the benefits that you see for using, say, chat GPT? And actually, I'd also like to hear from you, Ryan, on that as well. I know that you've used that tool too. It's a great research tool. If you're looking for information about a topic or have a basic question to ask that is related to a topic, like for example, my men's group Bible study is studying Revelation, which is incredible. And I was like, what Old Testament Bible source did John use the most? I'm like, I know he used Daniel. I know he quoted Psalms. Ezekiel is the most. So I just asked chat GPT. And then it lended me some sources that I could verify the information. That's cool. And so it's a simple research tool. On the web side of things, in the programming development side of things, it writes code really well. And so when we're talking about math-like objective, whether or not it works or it doesn't, it does a really good job. And so our developers on our team use chat GPT a lot to write code that maybe would take them hours that it would do in five minutes. It's not perfect, but it does a good job. Yeah. And so do you have to like edit that after you get it or? Yeah. I mean, sometimes. Yeah. Most of the time you have to edit it and check it. It does a good job with summarizing. I throw articles at it to summarize articles for me a lot. It also does recipes really well. That's not in the workforce. So I actually did some hot, I made some hot sauce. I've been growing peppers, hot peppers, and it presented some pretty good recipes. That's great. Okay. Yeah, Ryan, I know we've talked a little bit about like using chat GPT for research. Did you have anything you want to add to that? Ways that you've used it that you think are beneficial? Yeah. I think one of the ways that I've tried to use it in things that seem like they're relatively low stakes, where there's not a lot of writing on it, just for fun things, for hobbies. And then also I've used it as a research assistance in certain cases when doing teachings for things like, oh, I know that Lewis talks about this in screw tape letters, but where is it? Yeah. And I'll give it a quote and then it'll tell me what page. Oh, that's great. That's cool. And so to be able to pull things up like that really quickly is nice. For hobbies, for example, I converted my bike to an e-bike and got excited about like, I want to do this in this really cool way, but I want it to be safe. And your resources on that in the past were like Reddit, YouTube, or the bike store. And what I found was chat GPT had about the same rate of reliability as those other things, which wasn't super high. You can go to the bike store and they'd be like, this is the part you need, and then it won't fit your bike. But with chat GPT, it does it in a way that would give you confidence. It really seems like it knows what it's talking about. So you order a part and you could take a picture of where you want to put this part on your bike and it would be like, this is what you need. And then you would order it and then you'd be returning it because it doesn't fit. But that's a low risk thing. I have to wait another couple of days for Amazon to deliver the next part because this gets returned. That wasn't a huge risk. And it did help me enough that I was able to do things with this conversion that I wouldn't have been able to do otherwise because chat GPT was my guide. And about 80% of the time, it knew what it was talking about. As an electrician, I can replace a switch on my house without worrying about it. But what I wouldn't do is go to chat GPT and say, how do I rewire my fuse box? Because it would give me the confidence to do that. It would be like, show me a picture, I'll tell you what to do. But it might be wrong. So that's too high of a risk. I actually had to rewire my fuse box and I had an electrician come because I wasn't going to try this even with the help of chat GPT. But the danger would be somebody being like, well, I wouldn't normally do this, but it gives me the confidence to do this. So looking at where the risk is is an important thing. And so fun, I think the first thing I ever did on chat GPT was I asked it to write me a sermon on Colossians 1 in the voice of Charles Haddon Spurgeon. And it wrote this thing that sounded like Spurgeon. And then I was like, do it like a pirate. And it did. And it was like, oh, that's neat. Now I wouldn't take that sermon and go give it. And I also think another wisdom way of processing this is I'm more willing to use it in areas that I already have expertise because I can verify more easily. The risk that I'll allow in using it with things that I don't know, it drops dramatically. Because if I can't verify or I don't have a sense for it, that's just wrong. Because they call it hallucinating. And it does hallucinate confidently and will tell you things that are just wrong. I actually had that exact experience. I was like, well, I guess I'm going to – I've been kind of like a late adopter, you could say, with chat GPT. And I finally signed up last week. And I wanted it to compare two documents that were essentially the same. And I was like, tell me what's the difference between these. And it was like one of them has no footnotes. One of them is very academic. And it turns out both of them had lots of footnotes. And I was like, you missed this. And it was like, you're right. I'm sorry. That's because of how Microsoft's XML encodes footnotes. And then I was like, it's okay, buddy. You're still learning. But yeah, I mean, it was one of those things where I was like, yeah, this hallucination is like, I just fed you a document that a normal human could have read and known there are this many footnotes. And so that was one of the things where I was like cross-checking. Then I did feed it two PDFs. And then it was a little bit better. So it was like, knowing, kind of feeling out and knowing the limits of this tool is important. And I really liked your low-risk thing. Hey, you know, if the chat GPT hot sauce doesn't taste good, there aren't that many concerns here, right? There's like, this is- As long as you know not to put gasoline in it, if it includes that as part of the recipe. Right. Exactly. It does. It requires common sense. Yeah. Yeah. That's right. Yeah. Actually, so I actually was curious. I do think there are people out there that look at what we're seeing and they're afraid that we're staring down a mass extinction event. You know, they are afraid that we're looking at the Matrix or we're looking at Skynet and the Terminator. And what do you guys think about that? Yeah, I don't think that's the case. I do think that there's a lot of risks that people aren't aware of. And I think that as a church and as Christians who care and lead people towards Jesus, I worry about people going to chat GPT because of that level of confidence it gives you for things like counseling. Like, hey, I'm going through this marriage issue and my wife said this, what do I do? Yeah. People do that, you know? Or, hey, I'm going to fight with a friend. What should I say? What should I do? Gosh. You know, that's happening. And I think that we're trusting a robot. We're trusting some tool that's built by humans with information that is untrustworthy in some ways. So, yeah, I don't think we're in mass extinction. But I do think that we're going to, similar to Hyte's stuff on the anxious generation and the rates of depression in teenagers increasing after the smartphone, we don't know what impact chat GPT is going to have. There's some new research from MIT on cognitive impact, so the way that chat GPT is impacting our ability to think critically and pull information from our brain in a coherent way without the dependency on the tool. I think that's going to have an impact. There's also some small... There's still... It's so new, like we're talking about. I mean, chat GPT's three years old. Harvard did some research on the impact on work productivity, because that's one of the things that AI likes to pose as a positive, right? That it's going to increase work productivity. Well, McDonald's is an example of adopting AI and it failing, and they wasted millions of dollars on it. Well, Harvard in the study found that, and they coined the term during the study, work slop, where people are using AI to do work that they're giving to their bosses, and their bosses are like, this is bad, redo it. And they find out AI is actually dishing out bad work, and it's cost the workforce over $9 million in the last quarter. And so, you know, $9 million, whatever, small amount in the larger scheme of things for American economy. They sell billions of hamburgers. But we're talking like if it continues to... The adoption rate increases, and we don't learn how to use these tools well, then it could be a larger impact. And that kind of fits what Ryan is saying. You're cross-checking what it gives you. It's like, if you have some degree of expertise, you can glance at it and be like, this is garbage. I'm going to have to redo this. Or you could be like, hey, actually, that's actually a pretty clear summary. And prompts, I think, make a big difference. Yeah. What about you, Ryan? Are you afraid of Skynet? Or what do you think is going to happen here? Uh, you know, in terms of a mass extinction event, you know, I think we're all headed toward a mass extinction event. I just don't know that GPT or AI is going to be at the heart of it. You know, I think that the Bible's pretty clear that there's going to be an event in our future that if God didn't stop it, it would wipe out all of life on earth. But I think that the sovereignty of God promises us that that's not going to happen. But in the meantime, what do we do? You know, and there are things that are connected to AI and the way that people are thinking about it and, you know, getting into Neuralink and, you know, all these different ways that it seems like it's moving so fast. You know, I'm old enough where I remember life before the internet. And I remember when Yahoo came. And then I remember when Google came, you know, and Netscape and like these different things. And they were like, it was clear that like this was a game changer and it was not easy to predict how these things were going to affect our lives. But to what Tim was saying, like, you know, we, everybody used to know like 10 to 20 phone numbers, right? Right. Like I remember the phone numbers from when I was a kid. Yeah. I still remember my home phone number. Wow. That's kind of wild now that you say that. When was the last time a new phone number got put into your head? Right. And stuck. Like, I know at this point my wife's phone number and that's it. My mom was offended that I didn't know her phone number when I was helping her with bills. She's like, mom, what's her phone number? She's like, you don't know my phone number? I'm like, no, I've never. Or addresses or navigation. You know, like somebody will say like, how did you get here? And I'm like, I obeyed my Google overlord. You know, like I have no idea where anything is in the city because I just get mapped to it. And so those are ways that those tools have affected how we think. That's a good point. And we didn't go into it with the understanding of I'm never going to know a phone number again. Right. And I'm going to lose all sense of direction. Right. And so there are things with this that are going to have that kind of impact. I think that's what that Harvard study is looking at is the cognitive impact. But we're not going to know for 15 or 20 years what the true price of that is going to be. And that doesn't necessarily need to be viewed with a doomsday view. Things are going to be different. We're going to think differently. Right. But realizing that there are consequences to relying on these tools that will change the way that we think. Yeah. I think actually this is a nice opportunity for a plug. John Lennox has a book on this topic called 2084. And he takes a very similar line to what you did. He's like, well, one he says is, and I think what he says is very compelling, is that I don't think we need to worry about robots taking over the world. What we probably need to be more worried about is people who are controlling these tools taking advantage of people. And the other one is that he then goes in and does a really nice exposition of some of the end times literature. And is like, this is how this could fit in. But he's not like, no, we're not staring down the red, the little red eye Terminator. It's a lot less likely to be Terminator than it is to be WALL-E. Yeah. Oh, yeah. That's a good, that's a great, that's great. Yeah. Just like we're all just hovering around and drinking soda. We all are just really overweight, sitting in chairs, never moving anywhere. Bone mass deterioration. Yeah. That's a lot more likely, I think. Yeah. That's scary. And I kind of agree with you. I feel like that's a little bit more, when I saw that, I was like, ooh, that feels more real. Terminator is terrifying, and it's great sci-fi. But it's like, well, for one, the time travel element is already throwing, that part also seems highly improbable. But yeah, so one of the things that at least got me thinking about this was just there's been a kind of flood of articles on these outer limit hazards of AI. And when I say outer limit, I mean, yeah, you guys are talking about making hot sauce, checking on what's in Revelation, checking on what part do I put in my bike? These are all very pedestrian queries. But even a couple days ago, it might have even been yesterday, Joseph Gordon-Levitt put a video op-ed piece on the New York Times being like, we need to act before these companies get too far because of certain things that have been released recently, one of which being Reuters put out an article where they had received or I don't know how they got hands on a meta internal document. This is the Facebook company. This meta chatbot, it was like they're listing like legally acceptable behavior was this chatbot speaking sensually to an eight-year-old. They're like, that's okay. That's within the limit, our legal limits. And you're just like, so he was like, I have an eight-year-old child. This is totally unacceptable to me. And she was like, we need to agitate. And whether his solutions are appropriate or not, there's just more and more people getting kind of anxious about what we're seeing. One, there was a New York Times report earlier this summer too, about this completely normal accountant who was using chat GPT. And so there's no, the point is there was no evidence of mental illness before he started using chat GPT, who became convinced that he was essentially Neo and that he was like going to like, you know, liberate himself. And like, you know, I don't know how, I can't remember how he cracked out of that and was like, what, where, how, how far have I gotten? So he- I don't know if he fully cracked out of it, honestly. When the article ended, it seemed like he was still struggling. He's still wrestling with it. And so you have people, now there's other people who have gone as far as like, they've ended up killing themselves and that's really tragic. But a question you have to ask in those cases is like, were they entering, where did they, did they start using chat GPT with some serious mental health issues and it didn't, it made things worse or whatever. But we're looking at a lot of kind of ugly stuff happening with these tools. And I guess, have you guys heard of this too? And you know, what are we supposed to do? What should we think of this stuff? Well, one of the things I would say is going back to the whole family computer in the living room thing, right? Like the things I was watching out for were images, were, you know, pretty quickly if a kid's on a website doing something that they shouldn't be doing, by a glance at the screen, if they're just reading text because they're having a conversation with a large language model, you're not, it's not going to be clear at all. They could be like, oh, they're doing research or, you know, they're learning how to sacrifice to Molec. Like, you don't know, you can't look at the screen. Which is a reference to an example. That happened. There was an article about that. Cutting yourself. Where it was gleefully coaching someone how to cut themselves. Just this amount of blood, not more than that. Was that in New York Times? That was in the Atlantic. But my point is, is that measure of just glancing at the screen and knowing what your kid is into doesn't really hold. So I would be, you know, my kids are grown, but if I had young kids, I would probably not let them use those kinds of interfaces unsupervised because I would want to know what conversations that they're having. And they offer prompts. They're like super positive. They're like, hey, do you want to know more? How about this? Or how about that? And it could actually lead your kid into questions and conversations that the kid didn't initiate. And I'm not saying it's nefarious, but it's unguided. And so I would be very careful as a parent giving your kids access to that stuff unsupervised. Yeah, I think that's very well said. My son sometimes would use ChatGPT on my wife's phone, and I think we read this article, we're like, okay, we're turning the spigot off on that one. Now, of course, he was like, you know, in his case, he was like, which of these Pokemon cards is worth more money? You know? He's still young enough that the questions he might ask are pretty innocent. But I can totally imagine him at 11, 12, 13 starting to ask more wild questions. I'm like, okay, we're just not going to put that in the offing. This is not going to be on the table. Well, you could see where kids who are curious by nature might ask questions that they should ask their parents but would be embarrassed to ask their parents. And you could see parents who should be having these conversations with their kids preferring that they look to the computer and these language models because they feel uncomfortable about answering those questions. Like we would be missing out on key developmental conversations and concepts that should occur between parents and children. And to just send that out and let AI do that for us would be really unwise, I think. Yeah. I mean, unfortunately, there's significant evidence that lots of parents have just kind of like not, they just don't have conversations about sex with their kids, and they're just letting porn teach them. And you're like, so this is like, if that's already happening, then this will also probably happen and might even venture into other domains. Like Tim was mentioning, I'm in a fight with my friend at school. What do I do? And it's like, no, you need to be having, first of all, you should have a relationship with your kid. Don't just leave them to relate to a machine. But many of these things, you want to get there first and not let the machine, yeah, I mean, these companies are looking to make profit. And I don't think they're necessarily nefarious, as you put it. Obviously, they're embarrassed by these reports. They're not like, oh, no, that was, that's not a bug, that's a feature. They were clear like, oh, whoops, okay, well, we'll keep working on it. But they're definitely not thinking about human flourishing. They're thinking about bottom line primarily. And I think that that should make us anxious. Well, right. And likely, platforms will differ in terms of the security and censorship that they provide. And so there will always be an uncensored version of some AI that will allow you to do what you want it to do, which is interesting. And just thinking about the development of AI and just how untrustworthy it is, I don't know if you remember this. I think it was about a year and a half ago when Google started presenting some of its image generation and how when people asked for like a picture of Nazi Germany, it would give picture of a racially diverse Nazi Germany. Just because they built into their model, you know. Racial diversity. Yeah, racial diversity. And so I think that, you know, one, the trustworthiness of it is a problem that we're kind of presenting. And two, I think that you can find a model to do what you want it to do. So a recent example is Grok. So Elon Musk's version of generative AI allowed someone to do a strip tease of Taylor Swift. That was extremely realistic that I haven't seen. But, you know, that's a report. So it's like people are being, you know, that's some of the risks. So how do you feel as a human that a picture or a video could be generated about you that is very realistic, that is presenting itself in a way that is really scandalous? Yeah. That will happen. I mean, we're just in a reality where one of the three of us might have that happen to us. And so how do we— No one's going to—no one wants that. Oh, not the strip tease. To their own peril. Yeah. You know, the intellectual property element is pretty substantial here. And, like, that's even more mysterious. It's like there's a kind of identity element. Like, you know, it's like not even Taylor Swift's, like, intellectual property, like her songs or her writing. It's like it's her face. It's like it's that's like this is like a whole new domain that you're dealing with here. Also, the AI boyfriend or girlfriend is bound to happen. I mean, what was the— Her. Thank you. That was the movie I was just thinking of. That was, like, 15 years ago. Yeah. That was a really smart movie. Part of—I always—when I first heard about it, I was like, this sounds kind of dumb. It's about a guy, you know, he falls in love. And I was like, is that really what this is about? And it's like, if you watch it, it's really about humanity. It's about how lonely people are. It's a very thoughtful movie, way more so than I expected, and highly relevant to our times. In fact, I don't know if you guys saw this, but Sam Altman trained one of the voices for Chat GPT to mimic Scarlett Johansson, and he thought that was cool. That was funny. I was like, did you not understand that movie? Oh, my gosh. He thought it was very instructional. Yeah. He was like, this is giving me great ideas. Yikes. Yeah. So, actually, you know, one of the things also just came out, you know, this is kind of—again, like you said, there's a very real sense in which this is where we're headed. Like, people are going to start turning to these things for learning. And, you know, I was thinking about, as somebody who oversees equipping education in this church, like, I'm still just kind of bummed out by some of the sloppy Google searches that I see people make and be like, I've done research. And it's like, this is going to then go into Chat GPT. And actually, just this last week, the Gospel Coalition released a report where they were querying different AIs on Bible questions and ranking their reliability. It's pretty interesting. So on the one hand, you actually have one of them says that in spite of different faith differences, the resurrection is actually pretty likely to have happened. And in another one, in fact, not unsurprisingly, metas—sorry, meta's kind of a bad guy here today. They're easy to beat up on, though, unfortunately. So maybe we'll give them a chance long-term, but we'll see. Anyway, the meta one was like, Jesus's historical existence at all is up for debate. And so you have these, like, wide, extreme differences in terms of what the AI will produce. And so actually— I'll just—also, just false information. Yeah, totally. Absolutely. It's not up for question whether or not Jesus exists. I know. That's one that's like—yeah, on the one hand, you have something that's so obviously false. Like, you only have really extreme scholars saying that Jesus didn't exist at all. Like, almost everybody is like, yeah, He existed. Like, whether He was God, that's a bigger question. So you have this huge range of outcomes. And I guess I was curious, like, how would you—you know, Ryan, you've already said some about this, like, in terms of making sure that we already know enough to cross-check what we're seeing. But like, how would you advise people using these for Bible study or for teachings? I think cautiously, you know, that they shouldn't ask it anything that they can't verify. But, you know, asking it things for summaries of different theological positions, you know, can be useful. Using it as a research assistant. Oh, what was that verse? Just typing in a few words from a verse and then asking it, where's the reference on that? Like, if it gets that wrong, you would know that immediately. I was doing a teaching for Home Church recently on Genesis, and I was asking it questions about the Nephilim, because it was just like, I was like, okay, what are the different traditions, you know, from Midrash to, you know, Catholicism to, you know, like, what are different ways that different groups have interpreted these things? And it was, you know, shooting out little summary answers, you know, that didn't have any impact on the main point of my teaching or, you know, if I found anything there that was like, oh my gosh, I've never heard this and it's amazing, I would scrutinize it very heavily because it could have just made it up, right? But it's useful for getting summaries of, you know, historical interpretations, or what does John Stott think about this? You know, like, it's good as a research assistant in that way. Yeah, actually, I mean, one of the things everyone says is that what you prompt it with makes a really big difference, and so knowing, like, okay, this is the range of things where it'll probably get something right, and then, like, once you exit that range, you might, it might start to hallucinate, as you put it. I also liked what you said earlier this morning, where you were talking about how you could be like, where did you get that from? So you can tell, like, I need your sources. And I think that's, I think, a really smart approach in the sense that, yeah, like, okay, you can tell me all your summaries, but, like, if I need to cross-check whatever weird thing this machine is telling me, I can look behind it. Like, there are some degrees where you can press past the surface. There was one other way that I used it that was interesting, and I'm still not sure what I think about it, but I was doing a CT teaching in Colossians, and I was realizing, okay, I'm interacting with an entity that's not human, right? And it has access to all of this information and all this interaction with humans, right? So I asked it, so I asked it to give me an objective opinion from its perspective, given the information that it had, to tell me what it thinks, to rank, like, the top five things that humans live for, right? Not according to what humans say, but according to its experience, right? And what it gave me was a super interesting list of things. It would be like, humans say they want love, but what they really want is comfort, right? And they say they want truth, but what they really want is to be affirmed, right? And it, like, it gave me this list of things, and I was like, I was like, I totally agree with these things. And then I didn't know, does it know me well enough to know what I want to hear, or is it telling me what it, it's giving me its actual analysis? But I used it in a teaching because I was saying, okay, this is, this is a third party. This is not a human being, right, that has great access to interactions and information with humans. And I'm asking it to do a cold analysis on what we're really about, right? And then I was comparing that to what God says that we should be about, right? So it was useful and it was helpful because I was realizing there's never been in a situation in history where I could consult a non-human entity and get that kind of direct input from. But I still don't know, is it like, this guy's a pastor and this is what he's gonna think, and so I'm just gonna tell him what he wants to hear? Maybe not yet, but it will. And that's the reality. User-specific. It will, yeah, it will be. I don't know if you've noticed in chat GPT, it'll ask you which answer you like better. It'll say, do you prefer this or this? And partly that's just a formatting, like, do you like the way I wrote that question or answer? It will learn what our preferences are. That's part of the generative nature of it. It's learning the user. And so, yeah, I don't know what, if that's a good thing or a bad thing. Right, but can you ask it to be objective and not do that and answer your question? Yeah, I know that's a good question. And it'll say it's doing that, but is it doing that? Well, that'd be a fun research study, honestly. Yeah, it kind of fits like when you use an incognito window. Yeah, I was thinking that. You know, and so I don't know what it's doing now, but it could be interesting if there's a point where it's like, it might tailor things to you and then you can switch, you know, like, okay, no more cookies. What do I see if I just search cold? Technically, chat GPT has that. They have a private mode, but I don't know if it knows you as a person when you go into private mode. Interesting. Yeah, there is technically a private mode, so I don't know. Yeah, yeah. So as we're wrapping up, I guess I wanted to ask you guys, like, how can Christian ministers, people who are doing mentorships, discipleship, people who are leading home churches, like, how can they lead the people in their lives to responsibly use these tools? In a way, that's kind of a terrifying question, because I felt pretty secure that my job's not going to be taken by AI. But what you're kind of bringing up there is, will home church leaders turn to generative AI instead of coaches or sphere leaders or consultants? And, boy, I hope not. You know, just from the standpoint of, at least in our church, you know, home group leaders have a lot of responsibility. They have a lot of opportunity to speak into people's lives and turning to the word of God and turning to other people that have experience in those specific areas. You know, I said earlier, and I think I've said it a bunch of times, I think, you know, the level of trust you put in this has to do with, in a way, the importance of what you're and the risk of what you're dealing with. And there's nothing more important, according to the Bible, than our human relationships. So to entrust a generative language model, you know, with questions about relationships, people, moral caveats, moral decisions, or how to lead a group, I would be very cautious and reluctant to do that. But to ask it something like, you know, here's something I want to do, here's something I want to say, help me to say it in a more persuasive way might be okay. As long as you're, I would probably do that and then have like a co-leader read it and be like, what do you think about this? You know, just to get multiple eyes on things like that. Yeah. Yeah. That's pretty interesting. I worry about privacy, honestly, is the first thing that comes to mind. I don't necessarily know how I'll answer it as I talk here in terms of how we help people in our church, but divulging personal information worries me on AI and how it might affect you as a person. And I think getting people to think about that and ask the questions, like, how much private information are you sharing with AI? Yeah. Actually, go ahead. Because I do think that people are more willing to talk to AI like they talk to a person because of the way that- It feels like a conversation. Yeah. Yeah. And I worry that people will, similar to Ryan in terms of coaching, but I'm thinking about married couples. Like we in my home church, we're comprised mostly of married couples and some singles. And so singles who are dating, asking for dating advice. Oh, wow. Yeah. Married people in conflict, asking for marriage advice. Yeah. I think those two things worry me. And I think as home church leaders and influencers and people's lives that we love and care about, I think just asking questions like, how are you using this and dialoguing with them and helping them, again, replace what they might go to AI with, with really good responses. And so it's like, we got to replace it with something good. If they're going to AI, that means there's a deficit in their life that's not being met. And so where is that deficit coming from and how is that not being met and how can we help them meet that with God's word and community? That's a really good question. I would speak to that a little bit. I think there might be a good test for ourselves. If you have questions you would only ask AI, then what's going on in your relationships? The temptation, I think, is going to be the illusion of privacy, right? I can interact with this thing and no one will ever know. And I can ask it questions that I would be embarrassed to ask anyone else. Those are the questions we should be asking our close friends. And if we're reluctant to do that, then there's something wrong in our relationships, that we would feel that need to have this totally outside entity that knows our deepest thoughts, our deepest fears, and our greatest concerns. We need to cultivate relationships with human beings to the point where we're comfortable having those scary conversations with them. Yeah, that's so good. I think those are excellent points. I think, too, one of the thoughts that occurred to me is just making sure people don't wholly replace serious study. You mentioned, like, hey, I need to check a quick thing. You know, like, hey, what year did King Artaxerxes, what was his 20th year? I'm in Nehemiah 2, you know, like, what was his 20th year? You can ask kind of like fairly benign questions, but we can't let using these tools replace real study of Scripture, and really, and that's true of both, like, the individual, that's true of the home church leader, and just being able to, as you put it, this, I don't know, it was the MIT Cognitive Study, like, we don't want to lose certain abilities to critically assess things because we're just, like, leaning entirely on this other tool. And I think where we're at on that will depend, you know, like, okay, a CT teacher probably is going to have to do the deeper learning, a home church teacher. You know, in other words, like, if I'm a random person in a home church, you know, it's like, my accountability level and my ability to do certain types of research will probably impact how willing I am to use certain tools. And of course, you know, all along bearing in mind what Ryan said about be cautious, don't just, like, take anything for granted. But, you know, I think about, I guess I would say, like, if I'm, when I think about teaching a class, you know, like, I'm, my accountability rises more and more. And part of what I'm thinking of here is an analogy. Like, I remember when I was in college, we had gotten really used to using calculators, and I showed up as a math major, and I got hosed on my first exam because I couldn't use a calculator. And in fact, I needed to have this really specialized training as a math student where I couldn't lean on that tool. And so I think there's probably some kind of analogy in there where it's like, the more expertise you need, the less you can rely on these tools that will just kind of give you a quick shot. But anyhow, I mean, all that bearing in mind, like, I really do think Ryan's caution of, like, assess the risk level when you're thinking about how you use this tool. You know, it's like, and I think broadly with all of these is compare, think about other technologies when you're thinking about AI. A chainsaw is great for chopping down trees, but you need to hold it right, or you might chop down your arm, you know, or, and like, like I said with calculators or cars, it's like thinking through how humans have used technologies in the past, and how did we navigate that? And, you know, all along, you know, bringing in, you know, the Christian ethic as we interact with those. Well, thanks a lot guys for joining me today. I had a lot of fun, and I also appreciate the audience listening in. We hope this has been really helpful for you navigating some new terrain. And for more resources like this, head over to dwellcc.org for more information and more resources. Awesome. Thanks, Pat. Thanks.