We Didn't Rewrite the Insurance Code for Excel: Commissioner Jon Godfread on Regulating AI

Paul Tyler (00:01) Well hi, this is Paul Tyler and welcome to another great episode of the L&A Hub Podcast. And and with me today is an esteemed commissioner from North Dakota, Commissioner Jon Godfread. Jon, thank you so much for joining us.

Jon Godfread (00:18) Yeah, thanks for having me, Paul. It's always it's always fun to talk technology, so

Paul Tyler (00:22) Well it i it is and I is it f it's fair to call you the father of the NAIC AI governance program?

Jon Godfread (00:31) I don't know if that that I don't know if I want that title or not, but kind of the way it the way it kind of came about is about ten years ago when I started in this role, I raised my hand to be vice chair of then the Innovation Technology Task Force. 'cause I don't think it 'cause I knew any better. the current chair at that time ended up resigning like three or four months later, and all of a sudden I'm the brand new commissioner chairing what is what became a pretty consequential

task force and then developed into the H committee and and as continue down that road. But I'll call it kind of an accident of how we became that came that, but it's it's something that I'm I'm really passionate about and it it was a good fit for me. And I think coming from North Dakota, we we we're kind of innovators by nature, right? We're farmers, we're ranchers. we we generally have to do more with less and so it's it's kind of in our blood a little bit. so it felt like a natural fit. It just kinda got thrown into leadership pretty quickly, which was which ended up being a good thing.

Paul Tyler (01:26) Yeah, well, th I'll tell you this is a great teaser for th for the episode. But before we do that, I I think maybe we just walk back through your career and how you got here because we were just talking about this before. You you've got to be one of the the few professional basketball players to actually be a an insurance commissioner.

Jon Godfread (01:44) Yeah, so I I mean I I've been blessed with a lot of great opportunities, right? And I coming out of school, I played basketball in college, and and really had an interest in financial and marketing and kind of businessy side of things. I like complexity, I like math. was in the banking world and and kinda got plucked and said, Well, why don't you come play basketball in Germany for a year? And so I I did that professionally. fully intended on going back into banking. Ended up going back to law school. again, also fully intending going back into banking.

But got exposed to the policy development process in North Dakota. And in North Dakota, we elect everything, including our insurance commissioner. So got exposure to that kind of process. ended up at the state chamber of commerce for a couple years, and and the my predecessor announced he wasn't gonna be running for re-election. It's like, well

I like complexity, I like math, I I I like the financial aspect of it and talked with my wife and we're like, Well, it seems like it could be a really good fit and and here I am ten years later and it's it's been probably easily the best job I've ever had. it's not one of those jobs where

You know, when I was in kindergarten, I either wanted to be a firefighter or police officer or insurance commissioner. I I I think even if you look back twelve years from now, I've been like, if you'd have asked me, Do you want to be insurance commissioner, I probably would have What is that? and and no. but it it is it is a really one, it's a really fascinating job, and two, this is such a fascinating time to be a part of it, with with the evolution that's going on in the industry. So I couldn't be happier with where I'm at, and and just really feel lucky that the the people of North Dakota have kept

kept me here and and a and allowed me to to stay in this role because it's one that I think I'm I'm fairly good at. And I get to kind of scratch that complex itch.

Paul Tyler (03:27) Yeah, well listen, I th I I think you're a perfect per person in a a terrific spot because you know, my I remember when was that? Twenty twenty twenty three? I remember around April or March, I started use you know, ChatGPT just kinda broke into the I I would say the the the knowledge or or awareness of of of people. I know a lot had been going on before and I you know, I thought to myself, my th this is gonna really change our business. Now

shockingly fast okay, the NAIC came out with some with you know guidelines and I maybe we talk you know, I'd like to just you know spend some time talking about that and I but maybe you start off by saying I I think at least people think regulators first instinct is to create a law. yeah.

Jon Godfread (04:16) Yeah, no, I think that's probably pretty accurate, right? I mean, I think I I've been again fortunate to work in this space for the last ten years and and work with my colleagues. And it's it's funny, it's or it's fascinating to watch the spectrum as they as kind of the my colleagues get into it or they they they're either new or they're just, my gosh, I've heard about AI now, now I've got ChatGPT, I've getting these complaints from my consumers or whatnot. The the first gut reactions is say, We've got we've gotta regulate this, we've gotta stop this, or we've got we've gotta get this under control.

And I I mean I'll admit I probably had that same reaction too a decade ago, but the minute you start peeling back the layers of what this actually means, one, it gets really, really tough to to narrow it down. And two, we have an amazing system in the United States already developed for this, right? So we we've got the Unfair Trade Practice Act. Every state has the I think the tools available to regulate this space and and really this is just a new tool. I don't think we we didn't rewrite the insurance code when Microsoft Excel came on this on the on

the on the scene. I don't think we rewrote the insurance industry when electricity came on, you know, and so these are all evolutions. Granted, this one's kind of on a hyperscale, right? And I think there's still a lot of unknown that we don't know how it's all gonna impact. But at the end of the day, I have the tools in my regulatory toolbox to go hold bad actors accountable. It's it's and I would hate to overregulate to the point where we lose the benefits that are coming to this industry because frankly

The insurance industry needs to benefit from the technology piece of it, right? I mean, consumers are expecting that out of our out of our out of our carriers. We're not being judged anymore by you know, how do I compare one insurance company to another insurance company? It's really getting compared to how does my digital experience with my insurance company compare to my Amazon cart or my Uber, Uber piece of it, right? So those are the consumers are expecting if I have a loss, I want to be able to go on my app.

fill some photos and and that that process has got to speed up and and to the credit of the insurance industry they are taking advantage of that but again regulation could stop all that and that that's a harm to the consumer and it and it makes that loss even worse and and more painful and so it's it's we've really got to be careful and I think we've taken a really principled approach to okay what does it mean what is what do these tools mean and what do we expect out of our companies and it that has been a fascinating spectrum to watch our company

Jon Godfread (06:41) Evolve with AI and get involved in it. I think we've talked before. You know, 10 years ago you'd talk to an insurance company and they'd What do you know about AI? Or at that point, autonomous driving, or or whatever the the word of the day was. They'd say, We're not using it, we're scared of it, we don't know. You'd go two or three layers down into their IT department, like, yeah, we're using it here, we're using it here, this is speeding this process.

And so there's a huge disconnect between the board of directors and the operations. And that was a that was a massive exposure. That was a massive risk for the industry. I I I think to their credit, they've really s they've really closed that loop down quite a bit. Those these AI discussions are happening in every boardroom. and again, because of the exposure risk for these companies, right? If if they get it wrong, if they violate that consumer trust, it's you know, the regulators gonna be mad, absolutely, yes, but again, their their real harm comes from

Violating that consumer trust.

Paul Tyler (07:37) Yeah, no, I I b listen, I personally love your approach because A, the technology is of changing so quickly. How do you how do you create laws? Do you how do you how do you create rules that don't impede we do have laws in place that protect, you know, privacy laws, we have you know all sorts of advertising rules to f for actors et cetera. but your point so it's probably some different structure that we need. Now

Ma maybe kind of c come back though in terms of in terms of practice and I think they're i i if I were looking at at kind of bubbling up the the guidance from the NAIC, it seems to come down with to th sort of three concerns, safety, accountability, and governance. Maybe those are your three or what does that look like in exam today?

Jon Godfread (08:29) Yeah, and and I think we're you know, we're we're doing some pilot project, pilot testing on on how to equip our examiners to go in and ask the right questions, right? And and have those conversations. And so that's going on and I think twelve states right now are kind of in the process of of finalizing some of those AI examinations, we're calling them. and again, we worked with the industry to to to better understand, okay, what kind of questions should we be asking? and and and how do we get at these pieces? Because

believe it or not, I think a vast majority of the industry they're good actors and they wanna do good, right? And so engaging them in that conversation of, here's some of the here's some of the loopholes we found or here's some of the areas that we found that we've corrected, we're able to then apply that to the rest of the industry too. And that's really, really important and can't be undersold. But again, it gets back to those governance questions. I mean you touched on it a little bit with the speed of technology development

Regulation is not known for its speed. We're not known for being on the bleeding edge of this stuff. And so again, if we get overly prescriptive and how we wanna regulate this

It'll be outdated before it is even passed in NAIC committee. And and so we have to adjust and adapt to that. And and again, my my firm belief is we do that through governance style checks, examinations, and making sure that again, when things aren't the out the outcomes aren't expected, or when things have gone wrong or or whatnot, how is the company respond and how they corrected that? Because

Things are gonna go wrong. There are gonna be mistakes made. And and again, that's one of the other risks too, is when you scale this stuff, you scale the good, but you also scale the bad. And so and and I again I haven't had to

Jon Godfread (10:10) Thankfully, in the last well, month or so with all the you know apocalyptic news coming out about AI, I I don't think I have to sell the urgency to the companies anymore on on on on what that means. And again, inherently the insurance industry is an extremely risk-averse industry. And so it and so that plays to our favor too. but again it it's it's how do we look at governance, how do we approach the management and the control of of these tools? and and what does that mean? And again, this is gonna continue.

I mean if you'd have told me

six months ago that there'd be, you know, news articles about the end of human civilization because of AI, I probably would have said, that seems a little quick. That seems I I I those articles are coming at some point, but it seems a little quick. But again, here we are. And so I I will put our our regulatory structure and the response that we've had in the insurance industry up against any other, you know, major financial industry, and and how they've adapted how they've addressed artificial intelligence. I think we've been

as as ahead of the curve as we possibly can be on regulation. And I think that's again has allowed our allowed our industry to evolve and continue to keep up with with all the all the changes.

Paul Tyler (11:20) Yeah, my my wife is is the concerned one. And I say to her, Well, the Roomba hasn't killed us yet. It's tried to kill itself a few times driving off stairs, but it you know, I but but I absolutely hear hear you and I hear hear hear the concerns as as well. Well, you know, it's you you mentioned a little bit about the tools for the regulators. do you do you see a day where hey look I put a chatbot on my website where th the

Jon Godfread (11:22) Yeah.

Paul Tyler (11:48) exam will actually have a set of questions they run against these chatbots to see what they say.

Jon Godfread (11:53) Yeah, I mean I think that we're already trying to explore with some of that. we we've explored a little bit even as a state government in North Dakota as a whole. we've got Teddy as kind of our AI bot at our for our state government websites. number one, it's available twenty-four hours a day. number two, we can also we can close the universe pretty aggressively, right? I mean, I think at least we we feel we've got good safeguards out where it goes with the information now. and so that allows consumers who again increasingly don't want to necessarily talk to somebody.

Or they'd want to do it on their own time, you know, is there, you know, maybe doom scrolling before they go to bed. They want to ask a question to, well, how do I fill out this application or what licensure do I need to do this? There's a resource available to do that now. and so that honestly that that helps our staff too, right? Because we're limited in the amount of hours and manpower we have to to take those kind of questions. And I think it's it's also increasing the knowledge that consumers are have having with their interaction with insurance. And again, all that is really, really good stuff.

the challenge is again keeping it in that closed y closed loop universe. But then we're also seeing, kind of the negative side of it is we're seeing consumers going to their favorite chat tool or GPT tool or or whatever it is and filing a filing complaints and, you know, feeling feeling that they've been aggrieved. These

We're we're we're putting together some stories for our for our legislative session of, you know, again, it's it's a little bit like Congress where everybody hates Congress but they love their congressperson, right? And so everybody is I think afraid of AI or doesn't like AI, but I really trust my my advisor, my chat bot, my my whatever tool I use. And so we'll get letters upon letters upon letters from consumers that are wrong, that are that are just inherently wrong or or s quoting wrong statutes or quoting wrong, you know, saying we we regulate Medicare and we don't regulate Medicare.

Yeah, whatever it is, and no matter what we go back and tell the consumer, they're convinced because they heard it from their AI that it's true. Yeah.

Paul Tyler (13:51) okay, so I just gotta double double click on this. People are literally going to the chat bot saying, This is what the insurance company did to me it's telling them, you were wrong. They're draft

Jon Godfread (14:01) Or or but I my guess is it's probably being more prompted of this is this is the settlement or this is what the insurance company did to me. how can I file a complaint or or w what give me some give me some options to to fight this or whatnot and who do I reach out to? Inevitably it comes back to the insurance commissioner or the insurance office. Then we get a filed complaint and it's like, well

this isn't right. Like w I don't even know what to do with this. And we s we kind of respond back with like, well here's you know, your letter doesn't make a whole lot of sense. Here's what they think we're trying to get to. And again it becomes a cycle of of no you're wrong. My my chat bot's right. And it's just like, I we're reading the code. It's not So it's it's an interesting consumer dynamic that we're having to navigate now. And depending on the aggressiveness of the consumer, it can lead to you know, again

Paul Tyler (14:22) Ha.

Jon Godfread (14:48) S some some really difficult conversations. Yeah.

Paul Tyler (14:50) Well well well this is interesting because I I've seen it in our company because you know we own Policygenius and you know for a p a point in time it w it must have been a couple of months, consumers were coming in to our agents saying, Here's what I want and it was recommending layered insurance, like buy a ten-year, fifteen years, twenty

Jon Godfread (15:09) Mm-hmm.

Paul Tyler (15:11) Same thing. It was giving them the same recommendations. But these people are convinced that this is what is. By the way, the ones who came through the chatbot have have had much higher conversion rates because it said, go to this company. Wha w wha why why does somebody tr trust a chatbot more than they're than real people?

Jon Godfread (15:30) I don't know. Again again it's it's that it's that a little bit of dichotomy of human nature, right? Where it's like, Well, this must be coming from the computer or or superintelligence somewhere and it's like, no w I mean

Again, AI gets better every day, but it's still not perfect and and and again it's still not it's still not you know infallible. And so it it's it's just it's really interesting. And I think it's I honestly think it's a challenge for our agent community too, a little bit where they're they're having to go out and kind of compete with some of these pieces and have consumers come in. I I I appreciate them being armed with questions or armed with information.

But when that information is wrong, it can it can it can lead to bad outcomes for the consumer. And so we're I haven't quite figured out how to break that code yet or how to figure that out with our consumers yet, but it's certainly something that's that's on our radar because the volume and complexity of the complaints we're getting are increasing because it's a lot easier for a consumer to go out get that information. Again, inherently a good thing, but if that information is wrong, trying to unwind that that twine ball is it gets a little gets be a little bit hard.

Paul Tyler (16:31) Well, l let me throw a hypothesis and then and and you can tell me what you think about an answer to this. One one is I think it's context of these of these AI systems that are reading these things. You y you know, I I

Jon Godfread (16:34) Sure.

Paul Tyler (16:46) Early on, this is before the you know, OpenAI and Anthropic had these huge context windows. One project I work with with a prior company was we can build a chatbot that answers questions off of contract. And pretty quickly I'd realized, okay, if you ask this question, you have to read the first sentence in the middle of this document, you have to go down ten pages, then you have to go down twelve. So like the RAG would not work. question answer pairs would. So

kind of pulling back this question all mainly about the I guess con the information people are putting on chatbots about the insurance and also probably from the regulatory standpoint, it's almost like these FAQs are becoming more and more important. You know, can we have a ground truth of questions answers that are

blessed by the regulators. You know, kind of like taking all those consumer brochures or c I I'm trying to remember what the name is for the, you know, information that has to go along with the contract. And do you think the NAIC will ever do that instead of the brochure? Yar, good.

Jon Godfread (17:48) We're already getting there. We're already I we're already we're already building websites like and adjusting our websites to account for AI agents, right? I mean, and I think I don't think that's groundbreaking or conf or anything like confidential. I think that's that's the way of the industry now where it's like, Okay, we're having an issue, we're having a discussion about f financial regulatory system broadly. We have to design our answers and and place them on our website in a in a way that is digestible for

these AI bots to go out when they go out and pull that information that it's spitting it back correctly and that we are the s we are a source of truth. And so that that's creating some adjustments on our websites already and it and and what we're what we're sharing. Cause I always talk to our NAIC comms team with we have a repository of fantastic work. I mean it is

massive volumes upon volumes upon it's 155 years of experience and it's all it's all available online. It's it's just that nobody can use it because you don't know where the where the hell it is. And so we it's it's getting us to that data organization piece and optimization and how do we how do we set ourselves up to be that source of truth for this next generation of of bots that are coming in and and and reading it, right? And 'cause that that'll if you can fight the fight before the fight even happens

You know, that that saves everybody a whole host of time. And so that's that takes a significant amount of resources to to look at that on how do you optimize your website, how do you do those those pieces, especially when you've got the level of technical volume information that we have at the NAIC and and state regulators. But my my my pitch is that it's probably never been more important to have a a well organized and a readable website that that answers those questions in a way that

Paul Tyler (19:04) Right.

Jon Godfread (19:33) you know, is acceptable to the chat bots and to these pieces because they're gonna find the information somewhere, right? And we might as well make sure it's right. Might as well make sure it's from a good source. And so there are people that are a lot smarter than me that are trying to do that and optimize that and and and have those conversations, but it we're already seeing that shift within the industry of of how do we how do we design for the future and how do we design our stuff for it for not the humans that are reading it. It's for the the bots that are reading it.

Paul Tyler (19:59) Yeah. Well, you know, again, s you know, s stepping back a little bit to you know some of the earlier comments about the regulatory structure. you know, one point here in New York there was a law in the legislature that was gonna restrict like information that and recommendations that were provided by chatbots. Connecticut had another law that was contradictory. I think Colorado came out with one. I think you you said how how are we gonna do this? Wha

Where where's the future? You know, now we've we've had, you know, calls from the AI companies for almost like a FINRA, which I I would hope we don't do personally, but well y how how how do you make sense of this? You know, wha what are the regulators how how's the NAIC approaching that?

Jon Godfread (20:32) Yeah. Yeah.

Jon Godfread (20:38) Mm-hmm.

I mean, I think the dust is still absolutely settling, right? I mean, I think we're still in the in a very much in a gray area of what is this gonna look like in I used to say, you know, six, ten, twelve years. That probably needs to be shortened down to six, ten, twelve months because of the way the speed this is going. but again, I I I do appreciate the the approach the NAIC has taken thus far. It's been very deliberate, very judicious, and and very frankly, again, operating at a high-level principle-based manner, right? We've done, we did the principles.

That was followed up by the Bulletin and now we're doing the AI testing. and so it there's there's logical next steps here and and how do we proof how do we trust but verify and and the governance models, all those pieces that we're doing. Now each state is certainly entitled to their own ability to do what they want. I think Colorado's been a a pretty good highlight of of some of the the more restrictive AI laws passed in the country.

I I I do think they're struggling on implementation, because for the very reason we talked about earlier where it is it is really hard to pin down these very complex models. And again, the models are built to change. And so if I regulate one piece of it, well, by the time I get to the action on it, it's already evolved and already they corrected itself or or what not. Again, the utopian view is that

these models improve and get better and better and better and you remove bias and you remove some of these, you know, the some of the inherent human issues out of it and it becomes kind of a perfect model, right? A perfect risk risk rating. I'm also concerned that we're not ready for that that level of honesty probably. but how do you handle the bad the bad the bad side of that as well? And so it's

Jon Godfread (22:21) We've been watching Colorado very closely. we've been watching obviously New York, what New York does, what c what Connecticut does, what other states are doing this space. I don't think anybody's quite got it right yet.

and I think you're seeing the the struggles in implementation with Colorado. They they got the governance piece down because that again, that's the one that I think we can all agree upon. But now when you get into the metrics pieces and and the data pieces, that's that's a challenge. I think this this discussion moves from how do you regulate AI to how do you regulate the data that goes into the into AI, data privacy. And we're in the middle of those conversations too at the NAIC. We're updating our data privacy model act and and and doing that, because again it all come for me it all comes back to consumer trust.

I'm of the generation and I think many like my generation are I assume most of my data is already out there. I assume they have access to it. I just want to know. I I and I want to know and if and I want to be able to have an opportunity to it's this is a really simplified way down of like almost a credit score, right? Where I can go back and I can I can fight if if there is if something's wrong in my data profile, I can go make the correction.

first have to know what you're using, first have to n and have to have access to that. I think that's eventually where we're gonna get to to kind of a credit check or a data check on on how we do that. Who does that? I my guess is it's probably gonna end up being Congress at some point. but again I think we've

We've taken a lot of steps as the insurance industry to not only position ourselves to have those have those conversations with Congress when they start talking about a Federal Data Privacy Act, but again, hopefully able to help steer that discussion s to something that's beneficial to not only our insurance industry, but I think the broader data universe as well, because of the work we've done on that front.

Paul Tyler (24:05) Yeah, a great no you know, even last month I would say opposition to the data centers have have just the noise has gotten louder and louder. Do you think that's gonna impact, you know, the framework you've laid out? Do you think the regulators are gonna be do think the approach is gonna be mu changed in in any any way as a result?

Jon Godfread (24:25) No, I mean I I y I I don't know to be honest. I I I think, you know, we're in the middle of the data center fight here in North Dakota, right? We're we're a we're in the geographical center of North America, right? We don't get earthquakes, we don't get hurricanes, we're we we're cold, we've got a lot of wide open spaces. We are we are what would be considered I think a a prime and we've got cheap electricity, abundance of water, all those so we are we are a prime source for these data center discussions.

It's been fascinating to watch this on a political scale to see, you know, again I I like we're we're a very red state and I but I think everything not to get too political, but

We've gone so far right that we're coming back left and now it and and old is new, up is down, we're having these conversations about you know land use and and what does this mean. And it to me it all goes back to well, this is artificial intelligence and we're afraid of it, we don't like it, we're nervous about it, we're nervous about what it does, and so we're gonna we're gonna shroud it in the fight of the data center, the land use of the data center.

I I don't know how that shakes out. you know, I think we've done a good job in this in our state of trying to, you know, avoid any kind of moratoriums or any kind of piece. But the the reality is to continue to evolve and to continue to use these tools, we need those data centers. They they need they need that those engines to to to run this.

I'm hoping calmer heads up heads prevail. But I I think again the real the real fight I think will exist in Congress at some point because no matter where you sit on the political scale, you hate big tech for some reason. Right. So it it it depend like you're gonna have a bunch of strange bedfellows getting together for different reasons because their hatred of big tech, whether you're on the right or the left or or wherever you're at, and c and frankly Congress needs a win at some point.

Paul Tyler (25:53) Yeah, it's it's crazy. It's crazy, yeah.

Jon Godfread (26:09) and so that that to me is a is a is kind of an area that I'm watching pretty closely because if there's something if there's someplace where the political parties can become aligned upon is their regulation of data. And I think consumers are asking for it, the the citizens are asking for it out of concern and and and all those pieces. So what that Federal Data Privacy Act looks like, I I there's you know there's probably ten different versions floating out right now.

But you know, we're trying to watch and trying to get an idea on which one actually is gonna gain some steam and gain some power and and and then and then inject ourselves into those conversations because again, we've done a lot of the work on the insurance side. I don't think I think it's we're a little naive to think we'd be able to get an exemption for the insurance industry from those those type of laws. But if we can help shape it, I think that's that's that's really probably a more realistic view.

Paul Tyler (26:57) Right. Well, I you know I'll tell you right now it's complicated. You know, we have for instance our company, we have offices in US, we have offices in Montreal, whole set of different lists. So e even our website has to be comply California, Montreal regulations, EU because we have some operations in EU. It's it's I don't know, th th whatever comes out it's it's gonna create a lot of work for a lot of people.

Jon Godfread (27:23) But I also think that I mean you've we've seen that on the websites they've all everybody's complying with the most restrictive area, right? It doesn't make sense to to well we'll do North Dakota's website this way and we'll do it it that just it's not reality. And so I again I think that's that's the argument for, you know, my my current governor was a congressman for the state. We used to have a good friend and we used to have a lot of conversations and he's like

Paul Tyler (27:30) Yeah.

Jon Godfread (27:46) It's really, really hard to have a federal data privacy law without having a federal data privacy law, Jon. So please don't come and talk to me about exemptions and and and and trying to s you know, sliver yourself off on these things and

It makes a very valid point. It's like it's it's hard to have consistency across the country or across the globe if you don't have consistency across the country and across the globe. And so I again that's where I think having the discussions on the front end before the law is passed is where we're gonna get most of our our fruits for our labor. and again, I'd I'd position us as a leader in that space in the financial industry, financial space, because of the work that we've done at the NAIC level to be able to have those informed conversations with Congress as they start to find come coalesce around what

Yeah.

Paul Tyler (28:29) Yeah. Well gr great. Thanks. The I guess last topic would be f AI and and and financial regulation. Now you've got y again, unique perspective. You've come from the the financial world. You also know what know AI. I w I would describe

And I've I've had a little bit of financial work in my career. The whole t you know, tiering of capital has been a pretty much a blunt instrument up to this point. How could that reshape how you look at the financials of a of a company when you do your quarterly or annual exam?

Jon Godfread (29:01) Well, I I mean the the biggest power of AI is to take huge sets of data and analyze them. Right? I mean and and and that's what our feder our financial regulators do that right now pretty manually. I won't say manually, but you know, that's a bad analogy, but they do. and so be able to speed that process up and get to our conclusions, it'll it'll allow us to spend more time on some of the nuances within the financial regulatory space.

Our companies are already doing this. Our companies are already figuring out ways to what's how do I invest the best to get my best return on capital? How do I do this reserve? How do I keep my reserves as tight as I possibly can so I can have that capital to deploy and to develop? And again, the insurance industry is in a very unique place, is our our liabilities are longer term, right? And so we have we have the ability to absorb some of those inherent economical shocks. I mean, you look at

COVID, you look at, you know, you want to go back to the great financial crisis, you want to go back to World War Two, you can go you can go back on and on and on. The insurance industry has positioned itself to withstand those type of economic shocks. This is this is no different. I think you're gonna get a little bit more granular with with how we set that up and and what the companies choose to do. And it it's a challenge on the regulatory side when you talk specifically on financial regulation right now, because

Stocks are good, bonds are good, public bonds are even better. We you know, those that's the traditional world we've been living in. Well, now you've got some unique features in private equity, you've got some unique features in in in you know collateralized loan obligations, you've got y some of these unique instruments that aren't new to the investment space are new in insurance.

And so how do we best understand that and how do we allow for the continued development of this industry? Because again, I'll not to get too broad with this, but we're in the middle of a retirement crunch, right? Europe is worse than we are, and and you continue down that phase, and so capital has to be available to go out and support our real economy in the United States. And the insurance industry provides a lot of that support. If we don't have the capital available to

Jon Godfread (31:08) sell annuities, sell life insurance, sell different products. That hurts our economy, it hurts our consumers. And so we can get overly restrictive very quickly as regulators and say we're gonna we're gonna clamp this down and become a much much more liquid economy like Europe is. I would argue that's a that's a mistake because again you look at the retirement issues in Europe, they're coming off of a social retirement system or have to evolve off of that at some point into a more privatized retirement system.

We've been in that step for about three generations in the United States and we still don't have it right. Right. And so I would encourage Europe to look at us as a as a model of of how do you kind of continue to evolve in that space and allow us to continue to lead and do what we do in that regulatory space.

And again, AI is a tool that's gonna help us analyze and help us ensure that the solvencies of the companies are there, that the are it it'll give us the regulatory tools to go in and again analyze these very high, highly complex decisions, capital framework, capital structures to get us to a level where we're comfortable and we understand it. And there's a lot there's a lot of spotlight, there's a lot of magnifying glasses on the insurance industry right now, especially around the financial regulatory space. and we need to do a better job as an industry of explaining our work.

And and how we're going about doing it because again, the private equity discussions, the the private credits, the CLOs, all those discussions, we've been having those for a decade. We just are really bad at telling our story. And so but again, all these tools, they're they arm the regulators with the ability to go in and do that analysis, which is which is highly complex but so critical to the financial backbone of our US economy.

Paul Tyler (32:49) Yeah, I b well, and I think I think AI combined with what I'm starting to see in quantum computing will be able to c create unbelievable ability to stress test portfolios, change projection because you who who knows? I d I I've I've talked to a number of people who say AI actually is gonna extend lifespan. Well, wait a second, if my liabilities were

Jon Godfread (32:54) Mm-hmm.

Paul Tyler (33:10) I mean I I was assuming a fifteen year payout period now it's go to twenty. That would it that would be or even fifteen years to sixteen years. That's that would have a material impact.

Jon Godfread (33:18) Well and and I mean we we saw that in the sixties, seventies and eighties, right? I mean the life the the more morbidity til tables for life insurance industries continue to adjust further out. We've they've stalled a little bit now, but yeah, I mean that was a big issue for the long time. It's like, well, I wasn't expecting to pay this annuity for, you know, thirty-five years and now all of a sudden I'm paying it for forty, or the long term care questions of it. I mean, we went through that long-term care crisis. And so yeah, I mean as you as you've got as the

Paul Tyler (33:40) yeah.

Jon Godfread (33:46) the human experience continues to evolve. again, the insurance industry needs to c continue to evolve. And again, I I'll put that up against any any other issue, any other financial regulatory body. It's we we have 150 years of experience in this sp in this space. I think we've done a very good job of of navigating whether it's a bank crisis or great financial crisis or you name the you name the issue. And not that we're gonna rest on our laurels and say and get overconfident with it, but it's

There is some built in shock absorbers in the industry and and that that provides some really good cushion to the US economy.

Paul Tyler (34:20) Yeah. Well I g I guess last question for you is you know advice to carriers actually trying to advance and and and do some interesting things in in AI. I've seen s a couple of companies on name and one now company has a huge huge public target for efficiency with AI. I've seen other companies and talk to them actually InsureTech Connect last week or this week and LIMRA where they're saying,

You know, w we're we're very conservative, Paul, you know. What what what's your advice today?

Jon Godfread (34:52) I would say get get moving, right? I mean I think I I think the the the companies that are gonna survive are going to be the ones that adapt the quickest, right? And I and I guess, you know, it's it's there is so much behind the scenes legacy systems and other just paperwork. And I pa I'll use paperwork as really broad term that exists within the industry.

Paul Tyler (34:54) Yeah.

Jon Godfread (35:14) That is low risk, high value targets for this kind of AI stuff. And that that's where I'd be deploying my resource right now if I'm a company. It's like, how do I clean up this mess of books behind the scenes that again doesn't impact underwriting decisions, doesn't impact consumer availability, doesn't impact pricing, doesn't impact claims, doesn't impact the the kind of that consumer-facing decision pieces.

Let me get my house in order first, and then we can start talking about the higher risk, higher the the the the those kind of consequential pieces. There is still a ton of work to be done on regulatory technology, administrative development, administrative processes, data exchange, filings, reviews, licensure, all that stuff can benefit from the use of AI. And and and I and again I think

I I think the companies that are being successful, that's where they're targeting right now. We're starting to see the tip to to come in and say, okay, we're gonna start looking at how do we use it to just underwriting decisions, how do we do consumer of eligibility and and those other pieces? But again, from what we've seen so far, the human has remained in the loop pretty aggressively on those pieces. And so the way I like to l think of it is it takes, you know, 90% of the decisions are probably pretty black and white, cut and dry. They're they're yes or no.

But now with AI, you're able to get through those ninety percent and really focus on the 10% and say, okay, these are the ones that need human intervention, these are the ones that need the decision the the the analytical piece of do we want to pay this claim? Are we paying this claim? Do we need to do or how how do we respond to whatever this is? So instead of spending all my time on the 100% of the problem, I'm spending on the the main ten percent where the the work really needs to be done. It's it's really

a good tool to filter through some of that stuff. And that again, that ends up benefiting the consumer with faster claims, with faster processing, faster all faster decisions, all those pieces. And so that that's where I'd say I'd spend out spend my time if I'm if I'm in the industry because I said this years ago, but this is a this is an industry that could definitely benefit from some of that evolution, revolution, disruption, whatever you want to call it, because it it's it's desperately needed, right? And

Jon Godfread (37:17) It's it's fascinating to watch the new entrants come into the space, the the you know, through InsureTech Connect or through Plug and Play or where wherever you want to be and and run up against some of these legacy carriers who are, you know, built legacy systems on DOS mainframes and are un and and are unable to adapt. They don't have the ability to do the APIs and and and jump into this stuff without massive investments. The smaller ones are able to move quicker.

And and we're seeing the legacy carriers move on this on this piece, but it it there's gonna be some fascinating shaking out of the next five, ten years of of who ends up surviving this this revolution. And if you're not if you're not investing into it, if you're not looking at it, if you're not if you're overly conservative in this space, I guess I'd be concerned.

Paul Tyler (38:05) Yeah. Well you you just did a commercial for my company. Don't worry. Yeah, said Commissioner Jon Godfread. Commissioner Godfread, thanks so much for your time. And I guess if if people have questions about either, you know, what's happening in your state or happening broadly in AI, what you know, who should they follow? Who should they track? Yeah.

Jon Godfread (38:08) unintentionally. Yeah.

Jon Godfread (38:25) I don't think I I I really don't think we're that scary. I think I think there's a perception that regulators can be a little bit intimidating. I'm I I can't speak for all my colleagues, but I can speak for most of I think.

I enjoy these conversations, right? And the the biggest thing I think we hate is surprises, right? So if you've got an idea, if you've got something that that is on the cusp of kind of release or or kind of a new innovative way to look at how we handle the process with the insurance industry, reach out to your local regulators, reach out to reach out to the NAIC, reach out to the H committee, reach out, you know, there's a number of ways to do that. You can th through the NAIC website, through my state insurance partners website, and have the conversation because I'd much rather be up on the front of that conversation and be able to say, boy, this isn't gonna work.

or or and or here's some things you need to think about because again with the new entrants this is this is a heavily regulated space and there is a lot of nuance to it

But I think for the most part our regulatory system is is is ready for that and ready to have those conversations and certainly willing to give guidance where we can. and it's always good for regulators to know what's coming and kind of how to react to it and how to how to adapt to it. And so we spend a lot of time trying to do that. but if you've got some interesting, innovative ideas, I love having those conversations. And I think most of my regulator friends do.

Paul Tyler (39:37) Great. Hey. Listen, thanks so much for your time and th thanks for the listeners. Be sure to tune in next week for a another great episode of the L&A Hub podcast. Thanks.

Jon Godfread (39:47) Thanks, Paul.