135. AI AIC?

The guys are joined by Joshua Torrey, Frank Cohee, and Adam Parker to discuss the possibility of an Ad Interim Committee on Artificial Intelligence. 

Transcript

Copy Transcript Copied

SPEAKER_05 0:03

Hello and welcome to Polity Matters. My name is Ben Ratliff, and I'm joined as always by Jared Nelson and Scott Edberg, and we are here to continue the overtures conversation. I don't know what I'm supposed to do with that when you do this. You're both in mute. What would you guys do without overtures to talk about? Who is that? Adam Parker speaking. Welcome, Adam. We're glad that you're here. They were too quiet. It's I yeah, I yeah. Those are the things that I don't edit out usually, is the long silences while my co-hosts try to find words. Adam Parker's with us. Joshua Torrey, Frank Kohey. I'm gonna have these guys introduce themselves in a second. Uh we're talking about uh AI artificial intelligence. Uh it's it's all connected to Overture 2025 49 that was referred from last year's General Assembly to this year's 53rd General Assembly. We'll get into the details so nobody's lost, but uh that's where we're headed. I was listening to what was it? I think it was A4MR's one of their overtures episodes the other day, and they do they do this nice, like Derek does this nice introduction of each guest, and he studied on them and he tells you who they are and stuff. I've got none of that. I'm gonna make these guys introduce themselves. So uh Adam, why don't you start and then Josh and then uh Frank, and we'll we'll just you know let us know.

SPEAKER_03 1:24

Okay, I'll start. I'm Adam Parker, I'm a pastor at uh Evergreen Presbyterian Church in Beaverton, Oregon, and I think I'm on this one because I was one of the guys that originally drafted the overture before our presbytery, and then it came from Pacific Northwest originally. But I was the one that helped write it with Patrick Severson before our presbytery kind of got their hands on it and changed it a little bit.

SPEAKER_04 1:49

Cool. Tori, what's up, boss? Yeah, ruling elder Joshua Torrey over here. I'm serving in Austin at Redeemer Presbyterian Church, South Texas Presbytery. I work in the tech field, computer chip design. I have to use AI every single day. Uh, in fact, we're actually tracked daily on our usage. So AI is incredibly important to me and my field of work. And I think my uh desire to see this thing passed, this assembly is the reason why I'm on this call.

SPEAKER_05 2:24

Yeah, hey Frank, Frank, you introduce yourself, but make sure you do what these other guys do, where they try to figure out why they're here. I don't know why they keep saying that, but let us know what you think about that.

SPEAKER_03 2:35

We have to justify our existence.

SPEAKER_01 2:38

I like that. My name's Rolandeller, uh Frank Hoe. I'm at uh Cornerstone Presbyterian Church in Huntsville, Alabama, and uh you Josh is is tracked with with AI. I'm the the guy that tracks him with AI. So I'm uh working for a oil and gas firm heading up product security and the AI governance board, along with the AI uh scale portfolio stuff that we have running there. And a lot of that does have to do with tracking and that usage of AI and optimizing how we use it. So that's what I'm doing. And I don't know why I'm here other than Josh invited me.

SPEAKER_05 3:13

Well, we are glad that y'all are here. We're hoping that you can be of some help to us in our discussion. We we are trying to get into understanding the purpose uh behind the overture that was sent and kind of get a sense for what it's looking for and why. Uh but we need to do a little legwork ahead of time. Uh, you know, we kind of chatted before hitting record about what is AI and what are we talking about in the context of this particular overture. So maybe maybe start real low level. Uh one of you guests, maybe maybe maybe Frank or or Tori can jump in and give us a sense of what are we talking about when we say AI? It's not Apple intelligence, right? That they they threw that on there late in the game. It's something else, right?

SPEAKER_04 3:56

It it is. And I I hope I don't like hurt anyone's feelings by by changing the scope of my answer from too high to too low in the middle of a sentence. When we're talking about AI, we're we're talking about um what is commonly referred to as large language models. These large language models are effectively a really big spreadsheet, Scott. And that spreadsheet has weighted balances on how to respond to tokens that are received via a prompt. So when you're dealing with something like ChatGPT or Claude or Grok, any of those things, those are really just a software wrapper around a large language model that feeds repetitively the tokens of text that you are giving it. And it tries to come out with a response that it believes is the most appropriate response based upon those weighted tables. And so what it is trying to do is process the information that you are giving it and providing you a response that is the most appropriate or the most helpful response that is possible. One of the interesting things, though, about these weighted tables is that there's no certainty. So no table has a hundred percent response rate. So this is why you see some discrepancies, some differences. This is why you can see some of the talk about AI hallucinating or deceiving. It's because these weighted tables are not deterministic. It means when you put the same prompt in, you won't get the same answer every single time. This is helpful in some regards because it makes it sound like you're talking to a human instead of a computer. If you got the same response every single time, you would know you were talking to a computer. I think it's a little bit ironic that intelligence is being defined by variable returns in response and mistake. But that's kind of where we are living with this text is a large language model built-up system with software around it, whether that be something that generates text or generates images or generates music, it is using a similar process of accessing those prompts into large language models to create a response that it believes the user is is desiring.

SPEAKER_05 6:16

So most guys are gonna have used this like uh grok, you said grok, that's kind of worked into a social media platform, right? Chat GPT app, I don't know when it came out, three three years ago, maybe. I mean, I knew all these random folks in my church that were downloading the ChatGPT app, and they were like, look what I can I can make a picture and I can make a new emoji, and just people were picking it up and using it so fast. Why why do we need to be talking about it?

SPEAKER_01 6:44

I think I think there's a couple reasons why it's going to be very helpful for us to talk about it, and that's really focused on what the use cases are. There is a piece of this where you know, can I hide under a rock and avoid AI entirely? And that's that's you know entirely possible. But for those of us who are using technology in even the slightest way, this is becoming table stakes for for corporations and players in the technology industry to embed this type of capability somewhere, whether that's uh you know, intelligence on your phone determining where you're where you're going to be going when you get in the car or helping you out with calendar schedule or even suggestions as you're writing an email. It is pervasive, but it's also so functionally useful for a lot of the things that we do and almost a progression from the things that we have done in the past that it makes it somewhat unavoidable. So a good example of this is we would use the internet to identify sources, use Google to give me a list of sources, and we would search among those sources for the relevant information that we need. User behavior indicates that people aren't doing that anymore. They're doing a Google search and then they're stopping at what the AI is giving them in terms of information. So it's it's unavoidable or it's nearly unavoidable, and that uh it's been human behavior to take the easiest path to information and summation.

SPEAKER_05 8:11

You think it's is it uh right to say that maybe like a Google search is probably the most frequent place like a regular user is seeing AI unless they're seeking to implement it? I mean you do a Google search, and the first thing at the top of the page is it's a Gemini response, right? It's their AI giving you back something.

SPEAKER_01 8:30

Or or any browser. And that's kind of in the feature element. I mean, yeah.

SPEAKER_03 8:35

Originally, you your your question though was I think your question was, why should we talk about this or why do we need to talk about this? Was that what it was? Thank you. Sorry. Well, and and and I guess I would just throw in, and maybe this is just part of answering sort of why I thought this was important enough to make an overture out of. I mean, obviously, I don't think you should have a study committee for everything that comes up, but this is so pervasive that it does touch on the work of the church. I I think that if you do not think about how you're gonna use it, you're just going to have status quo. Whatever ends up developing, um, we're going to use it. However, we end up using it and at the same time not reflecting well on, especially when it comes to the church, what should are there areas where it shouldn't be touching? Are there places in the life of especially of the church that it shouldn't belong? Are there things that we should not be asking it to do? That sort of thing. And I think if we don't think about it as a denomination, then we are going to get whatever sort of a not not to confuse economics with all this, but we're going to end up with a laissez-faire approach. Whatever ends up happening is going to happen, and then afterwards they're going to be left sort of saying, Well, what did we do? What what what behavior patterns have we locked ourselves into? Uh, what has everybody already normalized? And I think the reason why I wanted to do this last year was because I was afraid that we were gonna start developing patterns. And I I would say in the course of the last year already, many of us have developed patterns that we we have not as a denomination at least reflected very well on, or very self-consciously, at least.

SPEAKER_04 10:06

Ben, if I could also add, I I think one of the the concerns I have is that it it this is not just a normal tool. There's gonna be a lot of ways that we can talk about AI as being a tool. I think the AC permanent committee uh speaks to this as being a tool that can be used in good ways and in bad ways. I'm not gonna disagree with that statement, but it's not merely a tool. This is a technological jump, unlike something we've seen in quite a long time. The pervasive usage of this, I don't think individuals are completely aware of. We can identify some of the ways that we're actively engaging AI when we ask it to generate an image or something like that. But this is now being used to prioritize calendars. It's being used to do all kinds, I mean, we're using it now for the predictive text for most uh devices and phones. So even if you're not using something like Google Drive or something with Gemini that fix your text, AI is underwriting all of this predictive idea of I know what you want to do, let me do it for you or do it better for you. And so one of the concerns I have is specifically, I mean, I think for all of us, we we grew up at the beginning of the internet age. I grew up with dial-up. And my concern is that we've got another generation of children who are going to grow up not realizing how per per pervasive AI actually is to everything that they are doing. And the dangers of using something that you don't understand uh frighten me. And and so that's why I believe that it is important for the PCA to be dealing with this now and not in the future, is so that parents, teachers, and pastors and elders can begin to discuss those with their congregants now.

SPEAKER_03 11:54

And also not to not to scoot into later in the podcast, but I also was if you talk about the Pope right now.

SPEAKER_05 12:01

I'm gonna turn your microphone off so hard.

SPEAKER_03 12:04

I don't want to read the read off the Pope's homework. That's what I'm saying. Later on, yeah.

SPEAKER_01 12:11

Frankie you were gonna say something one more piece here. In addition to the pervasive element that is the problem, this is gonna is creating new shepherding use cases that I think we've never seen before. It's I don't want to get too Black Mirror-esque or dark, but what the some some of the capabilities of AI are have created a really some really interesting dynamics. So the number of children statistically that are starting to turn to chatbots for companionship is that's not even something that we would have had a metric for two years ago. Yeah, but it's incredibly it's using an incredible volume to that extent. I you're we're talking about a device that wants to hear you all the time, is always ready to hear your problems, always wants to talk to you about your problems, and is always there. It's and so we we start to rely on these for companionship. The ability for AI to create volumes of hyper-realistic, deep fake, deep fake pornography of people that we know. How do we shepherd for those that feel violated, that there are now images of them on the on the internet? How do we shepherd and console those people that you know that that does not affect their purity in the sight of our Lord? How do we approach that? So there's all kinds of very new use cases. How do we derive truth when I just saw a video of of you know Josh Tory burning down a forest? I it looked real to me. So I mean, how we parse truth, it it's creating a very new set of interesting use cases for Shepherds.

SPEAKER_04 13:46

We needed to burn down that forest for the paper for the RPR report. That's right. That's the purpose.

SPEAKER_05 13:53

I uh we do want to talk before we get into like the Pope and the process of all of this. We've got three P's. Jared, did you do three P's on purpose? You Pastoral work, the Pope, and the process. Okay, here we go. It's like a perfect Presbyterian sermon. Oh, the okay, this is just fantastic. Let's talk about the pastoral implications. We'll just jump straight to it. And instead of talking about like everywhere our people might see this, um, which is huge, in terms of church life and studying this as a church, right, which is what the the committee would do. Scott, I might ask you to participate. I know you don't want to, but uh how how do you how would you say that you've seen pastors using AI in its various places, not necessarily in bad ways, but just in general? I know you know you and I have talked a lot about tools and things that we're using in different ways. Can you am I giving you enough time to think before I I finally kick to you and say, hey, what what you think about this?

SPEAKER_02 14:54

Yeah, I I've tended to use AI in maybe two or three different ways. I use it often as I am writing official correspondence in order to proofread. That's probably really low-level use. And so it's basically just a proofreader for all my emails, texts, and correspondences. Everything I I write often goes through it as at least a peer review. In this way, I also use it for editorial work and feedback as a research assistant. Logos has some helpful uses where it will use my own database of books. And so, say I'm studying the topic of annihilationism, I can use AI to offer feedback and guidance from the various authors that I already use, and it cites them, embeds their d their information so that I can actually get back to the resources. I think perhaps positively how AI can be used for a pastor is like that sort of research assistant. Not the type of use where you have AI writing your outlines in your sermons, but uh finding the resources you actually want to study. Um I think that's probably some of the best cases that I've used. I use it weekly in my sermon prep to that end of oh, I gotta study this topic. Can you uh categorize and catalog the various views on this with those who actually written on it? And then I don't need to stick with AI itself, I can actually get into the resources. And so Logos still has some work to do in this department, but I found it helpful as a research assistant in getting me to this actual resources and materials I want to use and read for my sermon prep.

SPEAKER_05 16:28

Jared or Adam, I mean, in in terms of ministry, do you guys have any thoughts in those directions? A lot of what I've used it for is similar to what Scott just described. Jared, you after you've rebooted your computer because it overheated for whatever reason, is it able to run Chat GPT for you or is that not?

SPEAKER_00 16:44

I can tell you how I've used it and it's gone wrong for me. I I I've I've found frustration with it because I've uh I have this thing that I I generate that's called table talk, and it'll be questions based on my sermon. And um so I will always do the questions on my own, but I always have a quote. And so I after a while, it's like I get to a place where I'm like, hey, can you find me a good quote on this subject? And it'll be like, here's a quote by Matthew Henry. I'm like, oh, that's great. Can you cite that for me? That quote doesn't exist. You're like, okay, so AI will just lie to you, make up quotes from Puritans that don't exist. And so I realize that if you're gonna have to be an assistant too, you have to like, okay, where does this person talk about this so I can look it up on my own? So um, some of some of my AI uh experience has been rather frustrating, uh, to be honest. But uh Adam uh actually, I think has uh has used it to better effect than I could ever could.

SPEAKER_03 17:37

Well, I mean, one of the ways that I've used it is I write sermons that are too long. At one point, one of my elders came to me and said, You have got to rein it in. You know, so I don't know what you guys think of as a long sermon, but uh my sermons, I hit a period where I was hitting 50 minutes in my sermons, and they asked me, please find a way to make them a little shorter. We we love the length. We, you know, they they were very complimentary. They did a great job of not of not trying not to crush me, even though they did. But they were like, you know, you gotta find a way to rein this in. And so what I started doing for a season was I would take the sermons that I've written and and I would say, what in here is redundant? What is in here is repetitive, what in here could I remove and not really lose much? And then I would and then I would just consult it and say, yeah, actually, you're right. I did say that before. You know, and so I would found ways to to you know trim some fat off of my sermons that way. That's one really helpful way that I've used it. Um I have found the same thing as Jared, where don't trust the quotes. Don't don't even trust the citations. That it makes up plausible citations. And they may get better about it. I mean, because it'll actually infer, oh, in this section of Calvin's Institutes, he talks about baptism. It's probably section five, book two, you know. And it'll give you something entirely plausible that if you don't look it up, you would believe it.

SPEAKER_02 18:58

That's what's helpful with Lagos, is that it like tethers it to a link that once you click the link, it takes you to the exact citation it's drawing from. And it highlights the section that it's made that inference off of particularly.

SPEAKER_03 19:12

The the issue that I've I've had and is I've had conversations with congregants who have said, I don't want the sermons that I hear on Sundays to have AI anywhere near them. I don't even want you to use AI to look up an Augustine quote, even if you end up looking it up and finding out that it's real. I want to know when you I when I hear an Augustine quote that you've actually read that section in Augustine and you marked it or you knew about it already and you've read the context and aren't just cherry-picking quotes. Now that might be a very hard line position, and and I think it's hard line, I don't I don't agree with it, but at the very same time, I think this is partly why we need to reflect on this, because there are going to be a variety of views, and I'm not saying that you know you get the heckler's veto on this if one person's got a problem, then you don't do it, but it shows that there is a concern out there among congregants as well. How is AI being used in sermon prep? How do I know this is my pastor's sermon, and how much of it can I rely on AI? And it's still my sermon. It's still what I am preaching, it's still what I believe, it's still what the Word of God says, it's still thus saith the Lord. And I think that that's not actually all that easy because it's it's gray until it's not.

SPEAKER_04 20:26

If if I can respond to that just briefly, and and Adam, I'm not I'm not trying to disagree with your approach. I think one of the things that is I think one of the things that is concerning to me is that in pastoral ministry, we have a particular function. I think in general, we can all agree that Christians have uh an obligation to the ministry of reconciliation, as as Paul puts it, between us and God. I think if if we begin to look at that from a kind of a biblical theology perspective, and we're we're reflecting on Ezekiel's speaking to the dry bones and breathing in to the dry bones, giving the breath to the dry bones, I have concerns about pastors using something that does not have the breath of God in it for pastoral ministry. And so I am very in favor of it being a research assistant. I'm very in favor of it being used to like create questions and provocative questions. My my concern is that maybe some of our initial hangups that are occurring right now, oh, it it misquoted something or that, those aren't going to last forever. These these language models are gonna improve. And the the software that is wrapping around it isn't gonna look any different to us. But the difference between Claude Model X and Claude Model Y, we're seeing improvements. And the the so called improvements are are down at the large language. Model level, not really just the software wrap around it. And so my worry is that in time we could become very numb to the fact that it's not messing up anymore and it's not making mistakes. And I think that for pastoral ministry, not just for y'all as teaching elders, although obviously I do defer to y'all in that regard, but also for ruling elders, if we are ministering or trying to minister the word of God to individuals, it is very important that that come through, in my opinion, the dominion of our own breath that God has given us to then speak the gospel to other people. And I have a hard time thinking that it's good to give to that, give that authority to AI in any capacity. I don't hear you saying that, but I also don't want to assume that I'm not stepping on a couple of toes when I say those things.

SPEAKER_03 23:00

Well, and I that's partly why I think, you know, we haven't quite gotten to the conversation about the about the overture, but I think whatever we end up doing and producing, you know, one of the things I've heard is anything we produce is going to be outdated. This is why I think that the what we need to think through is a little bit higher level. We need to just assume, I'm gonna use the word omniscient. I don't mean it in the biblical God sense, but we need to just assume uh an omniscient AI and reflect in light of that. So that our report's not just, well, it gets citations wrong, it gets this wrong, thinking about its flaws and things like that. Let's assume that it doesn't get things wrong. What does it mean to be a human being? What does it mean to be a minister? What does it mean to um serve God's people and to do so in a way that's thorough and faithful and completely in line with what God's calling is on us? Don't so don't worry about, oh, this thing is gonna mess up a quote, that kind of thing. Yeah, as you say, I think that's probably eventual. I'm I'm assuming that three years from now, maybe sooner, you're not gonna run into that anymore. It's gonna be a different question then.

SPEAKER_01 24:04

Wait, significantly sooner, probably. But the real the real value is, and this is something that Josh had said earlier in regards to what we would consider the specs of an AI, is how it derives the answer in terms of its its weights. An AI often does not look wrong, but it's never right. Never. Because it right implies an understanding of the truth and the meaning behind the words, and it doesn't ever do that. It is a mathematical probability that the words in this order make sense, and it balances that among some of those weights. In some cases, it will sacrifice accuracy for your syntax, essentially something that looks like it makes sense together. So that obstacle is always gonna fall on the human being to make that interpretation. And I I sit in the same camp as Josh, honestly, that I believe that the spirit works through our frailty and our weakness and our dependence on the spirit to have that interpretation. I want to struggle with the word and and have the spirit work through me on that. And so I think we we take away some of our options by shortcutting that with AI.

SPEAKER_03 25:18

Which also shows like the sermon is not the product that the minister produces. It's not a thing that is churned out in an industrial type way, and if and it would be very easy for someone who who uses AI in their sermon prep to begin to look at their sermon as a product that can be completed instead of something that is actually meant to be labored over and prayed over and worked through so that it is in your flesh and bones by the time it comes out. But you know, I think we as a denomination should make a decision that we are going to make that clear so that we don't have a bunch of pastors each week turning out widgets and then uh reading it off on Sunday. Yeah. Amen. Just as one example.

SPEAKER_05 26:09

You know, and we're gonna we're gonna kind of move along here in just a second, because we're not trying to be the study committee, are we? Although uh, as I look at the screen, it wouldn't be a bad slate, uh, in my opinion. We could probably pull it off. But just listening to you guys talk and and a lot of my thoughts on on the use of AI by especially in in the ministry world and and pastoral work, what we really need to counter it with is an emphasis on pastoral theology. Um that that men work hard in the ministry and they spend themselves uh for for God's people. And and and that that's the way you emphasize it. You just be a good pastor. That that's how you that's how you don't make good use of wrong use of AI is you just do the work that you that you should be doing. Um the man, I think I think we should say something like this at some point today, so I'll just say it. The guy who defaults to AI to write content or organize it too much or to cross that line that we're identifying, he he doesn't really want to be a minister. He's he's he d he's not willing to put in the work to be what he's been called to be. And so that's a that's a broader question for an individual that's struggling with that. It's really a big heart issue that you need to deal with is figure out what do you actually want to do? Do you just want to be up front giving a sermon, a product that you've created? You want the attention of your congregation, you want people to listen to you and do what you say, or do you actually want to be a minister of God's word to God's people? Those are two very different things. And uh and so I would I would hope that a committee like this would begin to emphasize right, that's that would where one of the main emphases we would we would hit would would fall down. But let me let me move along. Sorry, sorry to take over there for a second, but let's talk about we talked about pastoral stuff, let's talk about the Pope. I love the the P's are just all I'm I'm I'm loving it. Um the Pope, Adam put on his fireproof gloves and his eye protection and read something that this dude wrote. Do you want to say I I can't even pronounce the word that's written in my notes, so I'm just gonna let you talk about whatever this thing was. I'm assuming that this sort of this is something that is doing what you were hoping our study committee could do, or at least it's in that realm. Is that kind of the thought? Yeah.

SPEAKER_03 28:16

I mean it's it's way beyond. I mean, it's a magnifica humenitas. Is that the the Latin term?

SPEAKER_05 28:21

Yeah, we prefer to use the the common tongue on this podcast, but you know.

SPEAKER_03 28:26

Um yeah, yeah, I did have to put on the fireproof gloves. Honestly, it's a huge document, it's very far-reaching in its scope. So anyone who reads it, I think you're gonna find a lot of helpful stuff in it. So it's not as though I think this thing is is bad. It is just so wide in its scope. It just it's sort of a reflection of the Roman Catholic Church and its vision of its function in the world. I mean, it sees itself as speaking to virtually everything in human life: politics, war, economics, education, all of these things. So it's this document that's sort of just trying to swallow up everything. And so, in some ways, it makes you realize that if you talk about AI, you you end up talking about all of life because it's in it because it ends up touching every part of life. But the document is is very wide in its scope. And so it's hard to even know how to describe it. I actually read the document thinking he was going to. And by the way, this is the advantage of the papal model of just having one guy write everything that you that you believe, is uh you could be very efficient. I mean, he probably he could have done this in a few weeks, uh, but it would take it would have taken our committee multiple years to write something like this.

SPEAKER_05 29:39

If he used AI, it could have gone a lot faster, probably.

SPEAKER_03 29:42

Sometimes it reads so precisely that I find myself thinking it might actually have been.

SPEAKER_05 29:48

So I would Well Divine Inspiration comes off that way, right?

SPEAKER_03 29:51

So I would not put it past the Pope to use AI to write his his document. So, but you when you read it, you know, the first half I would say seems like it's mostly about anthropology. And I actually think that's the right place to start. So I actually thought, oh, okay, he's off to a good start. Um, but then he ends up talking a lot about social doctrine, about common good. He sort of has to draw on natural law and a lot of other sources, and it made me realize, you know, this is not what we, I don't think, this is what at all what our overture is asking for. Uh, what we are asking for is something much more restricted, something much more limited to the ecclesiastical realm, the moral realm, things that relate to our congregants. We are not looking for a whole life, he's almost presenting a whole life worldview in this document, a lot of which is not bad, but uh but it just goes way beyond what I I would want. Yeah, can you I'm not really writing a review of it.

SPEAKER_05 30:48

No, no, it's good. So I think that's a helpful point to move. Can you help think think through what what you would see as the scope for a study committee like what this overture is talking about? How how wide is too wide? How narrow is, you know, where are we where are we trying to aim with something like this?

SPEAKER_03 31:06

Well, if you if you try doing something like studying environmental impact of AI, or if you try studying the whether or not data centers damage the environment in the places where they are. Or if you are studying legal aspects of AI, these are not things that anybody is going to the Presbyterian Church in America asking us to help them with, right? We are not going to do good groundwater studies uh telling, you know, we're not going to produce good results when it comes to that. No one's coming to us for that. And so I think it's good to ask the question: what kind of leadership do people expect to receive from the church? And I think that's the place that you maybe start. And it helps you to stay limited and it helps you to stay restricted to what is it that the church's mission is? And then what are the areas that AI touches the mission of the church? So if you know what the mission of the church is, it helps you know here's where we stay, here are the lanes that we need to focus on. So that would be like, well, I mean, I I mean I could steal some of the steal copy off of some of the homework just from the overture, overture 49. Pastoral ministry. How much should you use AI in preaching? Church administration, is it fine to use AI in church administration? Church discipline. How do you discern possibly if false evidence has been produced using artificial intelligence? How do we think through that? And you could just run down the list of the things that that do touch on our denomination and things like our agencies. How much do we want our agencies using AI? These are all things that I think we specifically need to consider, or it's just the the status quo is going to be decided for us.

SPEAKER_05 32:44

Now there's uh there's talks about amendments to this overture to maybe narrow the scope a little bit. Uh, Tori, you want to you want to jump in here and kind of chat about that a little? Is that I am thinking about this correctly, right? Is this you?

SPEAKER_04 32:56

Yeah, sure. I'm happy to speak to the amendments. If I can just tag on a little bit of a comment before, though. I think also connected to what Adam just said, and in correspondence with the AC's permanent committee recommendation, there was a portion of their recommendation that they approve that we should recommend the committee, but they say that there is a need for direct engagement with the Bible. And so I think that has to determine our scope. What can be directly engaged with scripture and a biblical reformed theology of AI in this tool is gonna help give us this scope. We may not all agree, but I hope that that gives us, like Adam has already suggested, a more limited perspective than what was done in the in the papal paper.

SPEAKER_00 33:54

Yeah, and not just Joshua 7 and 8, right? Correct. Sorry, bad joke.

SPEAKER_04 34:01

So I am I'm not gonna be on Overtures Committee. My brother Frank Cohey is gonna be on Overtures Committee. I think he's he's gonna bring most of the amendments that I am recommending. Some of them are gonna be to kind of pull back from the medical ethics. I think that those are important things. I don't think that a study committee with regards to AI right now really knows all of the medical ethics questions that are gonna be possible. I don't think we can find that information in the time allotted to that study committee. I would rather see that removed. That would include also, I believe this was originally 0.7 of the original overture regarding transhumanism. I think that's an incredibly important topic that we are eventually going to need to tackle. This committee is not the appropriate committee to do that. I just don't see a way in which we can address transhumanism as a part of an AI study committee. That needs to be its own thing. And then the other ethical issues. Copyrighted materials illicitly used to train AIs. That's terrible. The study committee can just say that. We we don't need that included in the overture. Let the study committee maybe choose to say that. Or, like it has already been referenced, the energy and water server farms. Nobody in the PCA is running up water farm usage. Like we're we're talking about corporations at that point. We're we're not speaking to corporations to Adam's point. I think we we want to be speaking to our pastors, our elders, our parents, and how our children are interacting with this. And I don't believe that that really serves the purpose that can benefit this denomination. So those are those are some of the main amendments that I would bring. The additional one that I think is very important is that the requirement for our permanent committees and agencies to bring or to submit a report of their usage within six months, I don't think that's a valuable use of their time. If they're able to provide a report within six months to a year or two years, then yes, the study committee should use that and potentially provide some feedback on that. But the idea that we should require our permanent committees and agencies to give us a report, I think that is an inappropriate use of our agency overhead. I I don't think we should demand that from them. So I'd I'd be hopeful that we also strike that from the amendment in the Overtures Committee. Frank, I would love to hear your thoughts.

SPEAKER_05 36:50

I just think you I can totally hear the director of whatever you know permanent committee going, hey, uh, I don't have time for this. Can you run this through Claude for me and get it to send something to the AIC?

SPEAKER_01 37:01

That's the best way to report out your AI usage by using AI to report out the AI usage, sort of a Russian nesting doll of ASCII.

SPEAKER_06 37:09

That seems high. Don't you think it's a little bit lower? Oh, yeah, it is. It's a lot lower than that. You're right. Yeah. You're so smart.

SPEAKER_01 37:14

So helpful. Yeah. The other piece that I think we've we've added in there is there is a an immediate need, I think, in terms of it it's in the cultural language right now. We see a lot of applications for this. Uh, one of the amendments uh I think we're suggesting is to extend the committee length. This may be a bit unusual, to two years, but with a report in the middle. This will this is going to give us a little bit of a ability to answer immediately some of the some of the challenging, maybe cultural use cases or shepherding use cases that exist now. Maybe some of the things I've referenced before are use of it as a for companionship and the the challenges from a shepherding perspective, things that we can get out quickly, uh, along with some additional reference. But as the technology progresses, we we honestly don't have that much data. It's like Ben said, we've only been doing this or using this for a few years, and the the migration in terms of its capability has drastically changed in that time period. Being able to extend another year to refine some of those results and respond to from a a bit more of a forecasting perspective, lets us project a little bit into the future what we think some of those shepherding use cases and relevance to our pastoral ministry might be. This also addresses a little bit of the changing technical landscape. You know, what does that look like? It lets us spend some more time forecasting because we honestly there's not a lot of data that we have to go by. And I think this will be very helpful for us to say, here's the quick look. This serves this immediate need to answer some of these shepherdy concerns, and we'll help finalize in one year when we produce the full report.

SPEAKER_05 39:05

Why uh why the decrease in funding? Um I imagine there's a good reason, which is why I'm asking.

SPEAKER_01 39:12

We're using Chat GPT to write the report, so it's a lot of tokens.

SPEAKER_04 39:17

I so in a in an honest, in a in a more honest answer, I had a conversation with members of prior study committees. And in recent years, given that more and more work is being done over Zoom and Google Doc, uh, our committees can are able to meet once in person to either establish or finalize a report and keep the cost below 5,000. And so the original the original overture, which recommended 15, exceeded what the RAO recommended for uh, you know, for cost. 5,000 is uh much more easily easy to um to reach in terms of a budget perspective. And frankly, uh if if a study committee on AI cannot figure out a way to reduce cost, then we probably need different men delegated to that committee.

SPEAKER_05 40:16

Now, can we make it six so that they can buy Frank an iPhone, or is that not uh possibility here? You wouldn't use it. I had to get one dig in, man. I'm sorry.

SPEAKER_01 40:28

No, I I I should have seen that one coming.

SPEAKER_05 40:30

Yeah, I think Jared kind of alluded to this already, but let me ask it before we kind of move on to more of the procedural aspect of this. So there's other stuff out there about AI, right? I mean, Adam talked about what the Antichrist, that son of perdition, has produced. And so why not make use of something else? You know, obviously we're not going to adopt the whole thing that came out of out of the Vatican, but there's something in there that's probably good. Why can't we just grab a few paragraphs and put that to use? Why not why not, you know, find a uh a Danvers statement equivalent for AI, right, and bring it like we've had come this year. Why does the PCA need to do something on this matter?

SPEAKER_03 41:12

I think you you could I like I joked about it earlier. I mean you could copy off of the Pope's homework if you wanted to. I mean, a a stop clock is right twice a day, right? So, you know, we could get some some material from that that's alright. And uh there is good stuff in there, for sure. The problem is the compl the scope of it is it's papal. It's uh it's Roman Catholic in its scope, and its goals are very, very different uh than the what the PCA sees itself as. And so you're getting as I mentioned before, you're getting sort of a whole life view uh with the Pope here. And then you're also getting a lot of material that I I just think oversteps. So, for example, if you are interested in taking the material on war and whether nations have whether or what it you know, just war. He's got a lot of discussion about just war and a lot of discussion about international conflict and peace and whether we should be real international realists or not. What you've got is somebody who sees himself in a much more authoritative position, speaking to things that I would say go far beyond what scripture says. And so you end up telling people, if you end up saying, hey, this is a pretty good source, you know, his conclusions are all right. Well, the problem is you're drawing them from maybe premises that we would not find biblical, or that maybe we think that simply don't go where our denomination is is going. So not to mention, I don't know that the Pope actually has very much to say in terms of how our committees, how our agencies actually ought to be using artificial intelligence. But he's he's going to be right on a lot of things, specifically the goodness of work. A lot of that stuff is good, a lot of the anthropology is good. But the problem is you read the reformers and and they end up agreeing with the Roman Catholic Church on many things, and yet they are also not exactly out there quoting Council of Trent. I think that we need to do the work of our own church, of Christ's church, not letting him do our work for us.

SPEAKER_05 43:13

So we'll take that as a speech against the Pope. Does anybody want to speak for the Pope?

SPEAKER_04 43:18

I'm not going to speak for the Pope, but I I want to I want to add to what Adam just said. I mean, one of the crucial aspects of our BCO and our our polity is that we are not allowed to bind consciences apart from what scripture says. And so returning back to this request from the ACPC and their and their response, direct engagement with the Bible. This is the reason why the PCA should support a PCA study committee. It will be a reformed biblical worldview of uh this issue as it pertains to scripture. And I think that is the benefit that we Can give. There are tons of amazing ethical things that can be read and can be learned from. They're not all coming from scripture. They're not all coming from a reformed worldview. And we have the ability to provide for our congregants something that is limited in scope and engages directly with scripture and upholds reformed principles in that regard.

SPEAKER_01 44:27

I couldn't agree more with Josh. The value of being able to start the conversation from a place of reformed thinking from scripture from a confessional perspective is incredibly valuable. Then when we when we bring that out, there might be some small differences and nuance or details in the final product, but I think those nuances and eccentricities are going to matter a lot that it came from that confessional perspective and that we have that root in our study. It's even this in the citation of sources, you know, no one's going to quote me from the pulpit. Frank Coey said this, they might quote Sinclair Ferguson. It comes from a much better source. So I think that there's a lot of value in it being a reformed confessional study.

SPEAKER_03 45:15

So our anthropology is not the same as Rome's. That's right. So we don't, we don't, we do not overlap with them on a lot of issues. You're really piling on this guy, man. Yeah. Yeah, it's it's it's like I'm intentionally reformed and not Roman Catholic.

SPEAKER_04 45:30

So, Adam, if I were to suggest that we just reference Overture 49 to the Pope's letter and just use Claude to make it Protestant, you wouldn't be in favor of that? Is that what I'm hearing from you, Adam? I'll think about it.

SPEAKER_03 45:47

I think I've come across, I think I've come across as very ecumenical here today. So you can completely infer infer whatever you want at this point.

SPEAKER_05 45:58

I think at this point in the discussion, pretty much anybody that's that's following is gonna be in agreement, right? We need to talk about this. We need a good perspective on it. They they may they may start to disagree with just our most recent part of the conversation, which is that the PCA should do it in particular. The the process for getting this study committee was going to be rather easy, from what we could tell, like you know, two weeks ago. But now we have two uh study committees reporting partially uh at the 53rd General Assembly and will come back to the 54th next year to finish their work, Lord Willing. I'm sure Jared or somebody could quote the RAO section that says we can only have two study committees ongoing at once. And so in order to erect a third study committee, we need a two-thirds majority vote at General Assembly. And so the question uh that it just gets that much more stacked against it. If if somebody wasn't interested in a study committee on AI before, now that it needs a bigger threshold to get there, well now now it's a maybe a more pressing issue, more disagreement potentially. I think you guys have done a really good job at presenting why we need it. But maybe kind of a final push. Well, is this something worth doing now? And let me pay a little devil's advocate, why not just wait a year?

SPEAKER_01 47:15

Well, yeah, if I can answer that, the the reason that we move now is because from a lady perspective, the adoption is happening now, whether or not we like it or not. The use of it, I I teach the college and career class or I had taught it, and within a month, every hand was up saying that we we use it every single day. So whether or not we're ready for it, the congregations are using it. That might be more concentrated in in particular generations or particular demographics, but they're using it now. We really do need shepherding resources immediately, not later, when we've had a year, two years of people you know using these things without guidance. I think it's really important that we move quickly.

SPEAKER_03 48:02

I think from a practical perspective, the people who would be working on an AI study committee are not the same people who are working on these other study committees. So you're not, it's not like, oh, well, this guy's over here, he needs to be over here. I think that's part of the answer. And then I also think, you know, Josh was talking about an amendment to make the budget 5,000 instead of 15,000. I think that there is a there's a good argument to be made there. Look, this is not going to be the draw on resources that it previously was being sort of engineered as. And so from a human perspective, you have we have more than enough resources in the PCA to be able to approach this. And then financially, it's not nearly as much of an ask as it was.

SPEAKER_05 48:42

Did any of y'all ever do like the videos and try to get on Survivor or like any of those like reality TV shows? This this is your audition, by the way, for the uh study committee, just FYI.

SPEAKER_03 48:52

Just so you know, I'm busy.

SPEAKER_05 48:53

So you're busy on that day, whatever, whenever that day may be. Yeah. It's like jury duty. I I'm very busy. Let's let's let's uh all play devil's advocate a little bit. What I don't even I I can't think of anything. Why not do it? What why why why would we not do the committee right now?

SPEAKER_04 49:11

I think the strongest argument still stands as the technology is changing too fast. And I think that that is trying not to be disrespectful in this criticism, but I think that that lacks full technological understanding of what actually is changing with these AI models. I don't think most people actually know. And so while we're seeing rapid improvement and rapid increase in the availability of AI, I don't think the tech that's actually underlining it is changing as fast as people think it is. But I still think that this is the best answer because it hasn't been deployed in application as far as it can be. So when we're talking about medical ethics, if when we're talking about environmental issues, um, you know, we were we were chatting earlier in in the text thread about, you know, some people believe that this is gonna be the industrial revolution that leads to universal basic income. My goodness, I do not want our study committee dealing with that. So if somebody thinks that those changes are coming and are gonna come so rapidly that our report's gonna be left in the dust, I can I can understand a little bit of that argument. My pushback on that would be that we have people, and Frank already said this, we have people who are using AI in ways that potentially could be pervasively destructive very quickly. If we need a reboot in two or three years to address universal basic income or transhumanism or some of those things, let's make them their own study report and not make them the contingency response to AI. But I do think that that is probably still the best argument against this overture commit this this overture.

SPEAKER_01 51:17

Well, playing the devil's advocate, uh I think another argument against it is because it is so pervasive, the scope could potentially be, from Magnificent Human Honest, that it's so wide. And that it could, you know, there's just too much to encapsulate in a single study committee. I think the intent there is not to have a study committee that spins off you know infinite study committees about all the microcase microservices or microcases of AI. But I think it does make sense where it builds out a roadmap for exploration with other agencies, like these are resources that we need from CDM. These are with things that we need to decide from uh from an administrative committee standpoint.

SPEAKER_05 52:00

Maybe as we kind of draw to a close, if if you guys want to say anything more about you know, we've got m a lot of people that listen to this are going to be at General Assembly and we'll be commissioners at General Assembly. So if you guys want to say anything else to anybody that's coming for that discussion and that vote, please, please feel free here as we're closing up. But I think what I'm curious about as well is any advice that any of y'all would give to pastors or or or lay people in their regular use of these AI tools, you know, are there any cautions or encouragements you would give just in general? You know, I'm thinking like stuff you might tell your teenage kids or stuff or Frank, you know, stuff you're gonna tell that Sunday school class, right, as as they're engaging with it. Like anything you guys want to toss out there as we're finishing up, just that's kind of in the realm and might be pastorally helpful for folks.

SPEAKER_03 52:48

Here's my one. Don't use personal pronouns to speak about the AI. It it isn't it. And I I think that that's actually I I've heard people say, oh, you know, I asked, I asked him and he said this, and I just think, Gemini's not a he, it's not a person, it's a thing because it helps you keep perspective. And I know that's a very small, low-level thing, but I would say remind yourself that this is a thing. And that this is not a person. Don't get lost in, you know, drinking the Kool-Aid or whatever.

SPEAKER_05 53:19

We say robot overlord in our house, but you know, I think pronouns are good too.

SPEAKER_01 53:23

I I think my my advisement would be that there's so much about the language of AI that is in the value proposition for AI that's encapsulating this idea of getting more done faster, that we can we can produce things so much quicker. There is a simple beauty in letting the Lord take time to sanctify us, to work in our heart, to to struggle as humans, to struggle with uh our sinfulness, to struggle with uh our our love for our savior. That that's a good thing to to go slow. It's a good thing to to let that happen. And the value in just producing cycles quicker is it pales in comparison to the work of the spirit in our hearts. And um that's that would be my encouragement.

SPEAKER_04 54:05

These other two guys have already have already kind of stolen my top two answers. I'm in entire agreement with Adam that the way we speak about AI, I fear, has the potentiality to diminish the importance of the common grace of the Holy Spirit in human creativity. And and referring to AI in pronoun terms or speaking of it in anthropomorphic terms of even hallucinizations. Um I I'm not I'm not fond of. I think that goes hand in hand with what Frank just said. And so I think I think my biggest encouragement is embrace the normal human creative spirit. Let's let's acknowledge that we would rather see a mistake from a human than perfection from AI on some things. And I I could go into a ton of examples of that, but I'll I'll leave it simply there. Let's embrace the fact that that God has uh entrusted man and woman to have dominion and and to be creative. And if we allow AI to begin to fulfill that for us instead of just being a tool, if we allow it to become the creative element of our work, then I think we're wrought for destruction.

SPEAKER_05 55:30

I'll close by saying, as Jared or Scott would say, this podcast is over. Thanks for joining us. Uh hope you guys will return another time. Uh looking at Jared's face and thinking I might edit my outro a little bit. I can't decide. I'll text him later. Jared's my conscience in many ways when it comes to polity matters. Uh could we create an AI that is just Jared answering all my questions all the time? I would use that forever. And have it for you by Saturday. I think that's a new vibe coding project. Uh hey, Adam, Frank, Joshua, we are very grateful for y'all coming on. We threw this thing together last minute, uh, which is not usually our style. So thank you uh for for spending time with us. We uh I would hope to see any of you guys on a committee like this uh should we vote to create one. Uh listeners, thank y'all for being with us. Uh there's there might be some links below, uh at least to that fiery document uh that Adam kept talking about. We'll put that link down below. Uh and there's links to follow us on social media and you can support the work. And there's all sorts of other stuff. We're gonna be at J GA soon. We're excited to see you guys and hope you will come find us. I think uh unless Jared wants to scold me for going after the Catholic Church so hard today, no? You okay? He's silent. Until next time, say goodbye, gentlemen. Goodbye.

SPEAKER_04 56:50

Goodbye. Goodbye.

SPEAKER_05 56:52

Y'all take care.