S2 Eps 1: AI, Automation, and You

Hi everyone, welcome to season 2 of the official ASNA podcast. My name is Enable Toepas. I serve as the weekly communications of ASNA 2025/26, and I'm the host of this podcast, where we try to bring actual knowledge right to your doorsteps. Hearing me today is my co-host, IBNOM, ASNA's future actuary, and of course, our very special guest, Tim Bisham. Tim is a seasoned veteran of this field, with over 25 years of experience in the industry. In addition to this, he's also the co-chair of the IAA, our International Actorial Association AI Task Force, and the Chair of the CIA Predictive Modeling Committee, as well as the co-chair of the SOA Exam ATPA Committee. They have made continuous efforts to promote the use of AI by actuaries, which makes him the perfect guest to share his expertise on our topic at hand today. The future of Actorial work, AI automation, and you. So let's get started. So before we dive right into the topic, Tim, we would like to ask you more about yourself, so how did you first get into the Actorial Science field, and what has kept you in the profession for as long as 25 years? Well, thanks, Annabelle, the pleasure to be here with yourself and Ivy. So the honest truth is I didn't actually start out trying to be an actuary. If you go back to school days, I actually took computer science and statistics, and it was a professor that was saying that the best thing you can do in computer science is not just learn computers, but actually learn an industry where you can apply computer knowledge and things like that. A great marriage of computing power and statistics and some business sense was Actorial. So I started writing exams and passed a few and had been on that track ever since. What's kept me in the field is I find that if you like learning and like doing new things, you will learn and do new things forever. You could do the same thing over and over again if you want, but you can never stop learning in this field. I've spent most of my career in insurance and then re-insurance and pretty complex business models. You can always learn more, especially if you've started a new national business. It's fascinating, a lot of fun. What drew you to the field of actuarial science into other fields of applied stats and computer science? Well, yeah, so I won't say when I graduated, but back then, there wasn't this flashy data science or so-called flashy type thing, and even some of the quantitative finance type fields, they were there, but they really weren't as prominent with Actuarial and this still exists today. With Actuarial, you could really see a progression through your career in early years, you pass exams, you get certain experience, and there's a really set path. You can follow, which is pretty beneficial in many ways too, and I like that. When I got into it, I liked the field, so I stayed. I see, so just to kick us off, how would you describe the biggest ways that the profession has evolved over the years? Well, complexity, everything's gotten harder. I think we'll talk about that later on, but yeah, with new tools and new power and new toys to play with, nothing gets simpler. The products get more complex, compliance with more complex, accounting gets more complex, regulatory gets more complex, it just keeps on going back to my first comment about learning. You're always having to learn to stay current. How would you say AI and other predictive modeling tools can help with this increasing complexity in the Actuarial roles? Well, I think it will help in some ways. Everyone knows how some tools, particularly the language models and such, can be really helpful in doing some of the so-called grunt work, give you a fast start on something. But to be honest, I don't think AI is going to make our jobs much easier, I actually think in the long-term it'll actually make things somewhat harder, because again, I think if you go back years and years ago, they talked about with the increase of PCs and cheap computers and there's several things they thought, "Actuarial field is just going to disappear because they don't need calculators anymore." But what you got was more sophisticated products and as a result of that, more sophisticated taxation rules, more sophisticated accounting rules, more sophisticated regulatory regimes, and then more sophisticated products to deal with those and take advantage of things and then more international trade. So I think AI is going to accelerate some of that complexity as well. Well, it'll make certain things easier on the front-hand, like maybe some data cleaning and things like that. The really complex work I think will get more complex because there'll be more tools to model these things. I see, that sounds really interesting, so where would you say has the much feel of actuarial science would have the biggest opportunities for actuaries to leverage these AI tools? Well, the leaders, if you look at, you know, always my opinion only, but the leading areas of actuarial have been the property casualty people. They've been early adopters, you know, more sophisticated, you know, generalize linear models and more sophisticated regression techniques. And to be clear, like, you know, when we say AI, I include predictive modeling and quite frankly, a lot of regression tools as part of AI. I know you can take a look at it and say, well, you know, regression is just a statistical technique, which is true, but it's also the backbone of a lot of these models and the techniques that are on behind. So you can't really divorce one from the other and they often leverage, you know, ensemble models. They're not, you can't just look at them in isolation. So I include AI in an actuarial sense starting literally with regression and data, some of the data analysis. But the PNC field has been, you know, adopters, early adopters because, you know, they have a lot more data, you know, higher volumes of smaller coverage, the type of business model that really works well with, you know, data analytics and AI type tools. If you look at the more advanced things, this isn't actual specific, but, you know, there's property casually companies that are using image recognition software to do early assessments on claim damages for car accidents. You know, there's still going to be human interaction somewhere, but initial assessments, some of it being done by a photo you send in. So, you know, when you start getting to other fields, like, you know, life and health insurance, you know, they're quite frankly, they're catching up, they will see their day. There hasn't been the big disruption or a big application is just yet, there's been some incredible successes in just digging through, you know, your existing book of business and looking at the data and performance, just understand what's going on with your business because you have more ability to see the relationships that maybe you thought were there or it couldn't test or just didn't have the ability, but with some of these new tools, you can really get into it fast. So yeah, we'll see. I still think the biggest change to the actual profession will be that, you know, a lot of the logistical things will change. With better compute, with better data quality, there'll be a stronger push for more data-driven decisions as opposed to, you know, more business decisions. I think there'll be more analytics behind the business decisions driven by AI and data. I see. You mentioned the PNC industry is at the forefront of this AI innovation and they've kind of been like the first adopt this change. So what would you say is the main factor preventing life and health companies from adopting a similar approach? I think in any AI adoption, it's a question of trust of these models. So, you know, the big difference is that with, with property casually, you know, typically won your coverage, so if you're wrong, you can fix your mistake pretty quick. With the life insurance policy, if you're selling it for 40, 50 years, you're stuck. So people aren't, I say, people companies, you know, directors, executives, they're not as, you know, they're more risk averse, right? So until they can trust these models and trust these techniques more or find out that there's a business advantage or they need to catch up to the competition because they have a business advantage, then you'll see some much quicker adoption. Right now, I think there's still a bit of fear of, you know, betting the house on something that hasn't been in some people's views tested yet. Would you say that fear, like kind of a fear is that kind of the approach that company leaders and business leaders adopt towards AI? Well, no, it's healthy skepticism, right? So if you go right up to the board of directors, they have an obligation to support and protect those shareholders, right? So they're, they're going to allow very selective plays and uses these technologies. They're, they're not going to, you know, bet the farm on the big book of business or some huge product that that could potentially really disrupt the company. You know, there will be business decisions in this direction, but I just, I think life and the little more, just because of the business model, it's a much longer term. I think there's more desire to stick with the tried and true and sort of proven and accepted. But when, when people show that new techniques are as good or better, and that day will come, then, then you'll see some pretty significant changes. Yeah, of course, that's very interesting. So just circling back to what you said about PNC companies starting to adopt this approach. So what would you say has been the overall results so far from these companies implementing sort of new technologies into their systems? Oh, gosh. Well, there's a lot of applications because, you know, again, you know, auto insurance renews typically annually, right? So you're always fighting for market share and trying to get in front of your customer. There's a higher push for advertising and market segmentation. So, you know, the marketing teams are using AI and analytics for market segmentation for marketing and advertising analytics. You know, they have websites they can look at click through rates. Those are the things that AI is really good at. You know, high volume low transaction, you know, low risk type scenarios that click through is on websites and, you know, to a certain extent, you know, car insurance, highly regulated rates. But typically, one year, so you can play a bit more. I'm not a PNC person, I'm sure someone watching on the PNC side will take exception to what I'm saying, but that's okay. But you know, it's harder to apply that in other areas. Health insurance, health coverage and health policy, it's getting there. It's starting to get there. I think a lot of the analytics is more on the health informatics. It hasn't quite hit the actual site yet, but I'm it will, you know, the health, health actual is a different situation in Canada because it's public system instead of private. But, you know, it's, it's an area that's just ripe for applications of these types of models, right? Right. Could you also elaborate more on like the applications on the health side? Also, you mentioned like all the companies are trying to use this AI model, so get an advantage to grab more market share. So would there be a similar application over it and the health side is just more short-term as compared to like the life side? Oh, interesting question. I'm not sure how good an answer I have on that one, that's a big question. No, the, well, I'll stick to the Canadian side of things because it's a little safer for, I think, you know, actuaries could have, you know, in terms of emerging or non-traditional fields, I think actuaries could and should be involved with, with health care and looking at data and analytics to understand costing of programs and, you know, it costs a lot of money to do health care, right, and, and they watch things very differently in the states, you know, in candidates, a lot of it's driving around data, there's so many different hospitals that different types of health record information, different types of systems, huge amounts of, you know, regulation around getting access to that data and confidentiality, which there should be. But as technologies get better and the data gets more, you know, centralized, why shouldn't actuaries be at the forefront of doing analysis and helping people make businesses decisions around public health care? Yeah, that sounds, that sounds great, yeah, that's a great answer, yeah, so how would you define the success of the companies that have adopted this risk-taking approach and adopting like AI, compared to companies that are sort of more behind on this approach? Are we seeing more successes, more wins for companies that are adopting this approach? That's an interesting question, I think, you know, I can't point to anything that I can recall seeing that's been sort of a big disruption in terms of applications of AI, like radically new products or, you know, there's been digital-only companies and, you know, test companies that just try to be true, digital-only, like, you know, they've had some success, but I think, and I think this will continue for the next, you know, so many years, I don't know, five or so, a lot of the wins will be just on incremental gains on some of the logistical things, so, you know, will the products radically change? There will be some changes, but you won't be able to point and say that was AI that did that, you know, maybe it was AI that helped it, but you can't say that it was just because of AI that, you know, because a lot of these things depend on, you know, back office infrastructure and, you know, even skills of your staff to manage this book of business, it's a very complex business, so it tends to move a little slow that way, so I think a lot of the wins have been on incremental gains, and, you know, just getting more granular, you know, there's so much more ability to study your book of business, even if this is on starting to make inroads on the pension side as well, like, you know, looking at more granular and bespoke assumptions and mortality bases based on what you can analyze on your population and relate that back to, you know, relationships and other, you know, geographically, where these places are and just, I know that's more core, you know, sort of data and analytics as opposed to AI, it's a spectrum. You can't do AI, in our world, you can't do AI without, you know, the data and analytics component, and I think AI enables better data and analytics, and that's where I think a lot of the decisions are made. Making a lot of these models operational is, you know, significant investment, right? And ultimately, depends on what you're doing, like, certainly on the valuation side of things, like, I'm jumping around different industries and things like that, but most, most evaluation, you know, no one's going to sign their FSA on, you know, a model that's purely black box, right? So, you know, you can use the black box to do some discovery and some analytics and stress testing, but until people really trust those models and can make them, you know, highly auditable and a lot of high degree of comfort, so, you know, certain feels just won't jump over to using those exclusively. Yeah, I like how you touched upon the black box aspect of AI models. We'll definitely touch more upon that in the next section of challenges. Obviously, quickly, I wanted to ask, you've mentioned a couple times that there hasn't really been any big booms recently. So given that, are you, like, would you say that we're at the edge of a very big boom, or would you say that the field will more slowly evolve, so, more incremental changes for, like, looking forward the next 10, 15 years. So I'm in the realm of speculation, right? So this is my guess. I think there'll be a couple companies and maybe not even, you know, North American companies that will kind of get it right. Maybe not get into some of the details, but for certain reason, I think some international companies might have an advantage on being able to access and use data more prevalently. And I think it's a possibility for them to do something that we might not do in a more regulated environment, say in Canada or Europe. So, you know, there could be a big disruption there, you know, like, you can, if you start throwing names around, like, if Amazon wanted to get into insurance, what do you think they would do? So, right? So exactly. You know, regulators, you know, love actuaries, right? In Canada, there's a legal definition of what an actuary is. An actuary is a fellow of the Canadian Institute of Actuaries, full stop. And there's certain things by regulation that only a fellow of the Canadian Institute of Actuaries can do. But however, that doesn't mean that anyone else can't do that and then hire an actuary that gets comfortable enough to sign and there's your disruption. Or you come up with a business model that somehow is, you know, sort of skirts the regulatory environment, which is going to be hard to do, but if someone finds a way to do that, it gets popular. There's your disruption as well. I see. You mentioned like international companies and a difference in regulation. So, would you say there's a country or a region that's leading the way and integrating AI into actuarial practice? Oh, um, leading. No, I think it follows the, yeah, I think it follows the provider. So, you know, it's, it's, it's the US, it's China. It's, you know, it's the big, where the big AI companies and leaders are, you know, just how, I think it's one of your questions that might be jumping ahead, but just, you know, different regions. The regulation is very different. The different actuarial bodies locally have very different approaches, right? So, you know, Canadian Institute of Actuaries, even though I'm the chair of the predictive modeling practice committee, you know, there's, there really isn't such a thing as a predictive modeling practice in Canada. Predictive modeling is a tool that is used by insurance, pensions, risk management to, to, those are the main practice areas, you know, health, the, you know, actuarial evidence. I'm not, I hope I'm not excluded anywhere else. But, sure, I missed that one. But, you know, we're the tool used by other people. We don't have data science actuaries in Canada per se. You know, Germany, Australia, you can become like a certified data science actuary. You know, they have these type of credentials in the site of actuaries. And I think the cash, you can get big certificates in data science and things like that. So, you know, different approaches for, you know, applying and asserting the profession in a certain way. And then ultimately you have to follow within regulation and regulation drastically changes from region to region. So that's a whole other topic. So would you say regulation is kind of defining how actuaries and companies and different companies approach AI? Would you say that's the main thing? It will have a very significant impact, right? Right? If you can do something quicker, faster and more complex because there's less regulation around data and privacy and things like that, those regions will have an advantage from a pure technological sense. It's highly debatable whether it's an advantage from a public protection and a public interest perspective. But that's a very political and a very different discussion, which I probably shouldn't engage in. Okay. I see we've kind of touched upon regulations and the limitations of AI, for example, data privacy. So I think that would be the perfect segue to our second segment on the challenges of AI. So I'll be handing over IV to ask the questions on this segment right now. Yeah, so thank you so much. So far, you've prepared a great insight. I'm sure a lot of students will find it useful. So yeah, we've talked so much about the potential AI we have touched on challenges as well. Let's, what do we want to go more into depth about specific challenges that appsharies may face? So for example, what are some of the biggest challenges or limitations that you've seen when it comes to applying AI in appshary work? So maybe it could be day-to-day purposes or maybe on like a more global level as well with the IAA. Yeah, well, I think a lot of it goes just back to the data. The accessibility of the data, the quality of the data, you know, so many times you hear just in good old-fashioned predictive modeling or things like that, it's, you spend 80% of your time cleaning and understanding the data and then, you know, then you're fitting the models as easier. And then the storytelling is always hard and interpretation and, you know, giving your results. That's a dark and a craft in itself. But no, the biggest challenge is definitely on the data. Like, what data do you believe? And again, this is, it's an interesting situation that AI is far more commercialized as a technology than most things we've ever seen before. Like, you know, it really wasn't that long ago that, you know, people didn't really know what appsharies did and they didn't really care. And there was kind of behind, you know, people in the corners of insurance companies, pension funds and things like that that did some sort of magic stuff, right? But now with everyone using advanced models and AI and analytics and highly advanced statistical techniques of different forms, now everyone's kind of like an actuary, but kind of not like an actuary. So, you know, even, even, you know, media outlets are hiring computational journalists to, you know, hire like a PhD and, you know, and pay and stuff to find interesting stories to talk about on these topics. And they do find things, you know, there was a little off topic, but there was a case in California where someone working for it might have been Reuters, but a media outlet, a quantitative person just got a whole bunch of information on automotive rates in California and was pretty convinced and wrote an article that there was basically bias and discrimination in the car automotive rates being offered in the state of California across different providers and different networks. And it turned into something. The regulators actually believe that to step in and act on that. So, you know, here you have a journalist having an opinion on actual work on auto rates in California. Like, whoa, whoa, whoa, whoa, like, you know, where that comes from, right? So, you know, those are going to be some of the early disruptions in that we're going to have to have potentially more transparency and accountability because more people will have, you know, access or an ability to have an opinion of what we do. So, um, but, you know, the question was about challenges. You know, it really comes back to, um, yeah, it really comes back to the underlying data. If things are clean and you trust your data, you can do so much faster if you have to call these, you know, clean and scrub and, uh, sorry, I jumped over to the commercialized aspect of it. Like, you know, data, people, companies sell data and you don't necessarily know what's behind it and how good it is and how much do you trust, you know, your product to design based on something that someone's selling the same with AI models. Like, you know, there's a lot of hype out there about what these, you know, language models can do. Um, and they want to drive up their stock price. They want everyone to subscribe to their system, uh, and it's incredibly expensive to build those, like the, the, the, all of them is incredibly expensive to build those. So it's, it's, it's interesting that we're in a situation where before technology could just be to the academics and people applying these things. Now, there's, there's a whole aspect of selling the science to, to, to make money. Um, and um, yeah, that, that part's really different to deal with. So that certainly hasn't caught up with the actual side that it might. Yeah. Uh, times are so evolving so fast. Um, I know in the field of acting, you know, like you mentioned, um, like, yeah, yeah, usage is a heavily debated topic. Um, so you mentioned, uh, talking about the California race case, um, how it's important for actuaries to maintain things like, uh, transparency, accountability, and professional judgment. Um, what do you think, uh, are some steps that we could take in the actual world to, to do that. So to maintain all these things as AI is so, uh, heavily and quickly evolving. Uh, well, you know, start with the easy stuff for the most transparent of the things that are closest to you. Um, so, um, I think, I think, actually starting out should, you know, read section 1,000 of the actual standards of practice for the Canadian subactories. It's an easier read than some of the other sections. You don't have to have a lot of industry knowledge. Uh, but, um, you know, there's sections in there that talk about data, use of models, uh, you know, obligations to report things that you know that you don't know, uh, and it outlines a lot of things about good professional practice on the sort of the basic nuts and bolts of modeling and basic actual, uh, basic actual work that all actuaries do. That's why it's the section 1,000 is the general section. And the Canadian institute is actually updating that now my committee, the predictive modeling committee. We submitted quite a bit of, uh, proposed changes to make it sound like it knew the standards should sound like it knows that AI and predictive modeling exists and it should address certain things that are unique to those, uh, those tools and techniques as well in terms of types of risk and professionalism considerations. So, so yeah, my advice would be, you know, be aware of what good professional practices and have some understanding of what the regulatory requirements are in your industry as well. Um, they, they let you know, there's a common theme to these things and they lay out, uh, you know, steps to take, uh, and good professional practice skills to use as you do your work in terms of communication and, you know, not hiding assumptions and certain things, there's certain things like that. Um, so, but start with those. Um, the more you can show that to your, your new manager, the more they trust you because they know you're going to follow the right path and, you know, uh, that just adds to your credibility. It adds to your value. It's, and then it's the most basic thing. Everyone has to abide by those. Yeah, thank you for that. Um, so you mentioned that, uh, you know, it's actually your standards. Um, we can kind of read into them. They already existing regulations. Do you think of the current ask your standards though, uh, could be evolved a bit more to suit the, the changing trend, uh, with the AI? Or do you think, yeah, it's fine as it is. Um, well, no, it is being reviewed. So, um, the actual standards board created an investigative group to look at, uh, guidance and standards around use of models. Um, I sit on that investigative group as well. And like, again, I've over committed on my volunteer work, but, um, but there, there's two more broadly. There's two schools of thought in that. Like, you know, AI, any of these things, they're just another type of model. So one school of thought is that existing standards of practice talk about good practice when using models. And if you interpret those existing practices appropriately, those would naturally encompass new tools and techniques like AI, predict the modeling, whatever else. Um, I'm of the belief, which is more to the other side, which is, okay, that's true. But there's a whole lot of people who've never really worked in other fields of actual, there's, there's, there's younger actuaries who've just worked in data and analytics and, and, you know, AI, they, they don't necessarily know pensions and interpretations of the standards of practice in broader context. So we need more guidance in general to, to, because the standards do have examples for life, for health, for pension, for, you know, P and C. So why shouldn't they also have specific examples and, uh, additional guidance related to predictive modeling and AI and things like that? So, um, there, there will be more on that. Because there's the standards practice, also education notes within the convenience of actuaries. Uh, the, the education notes, they, you don't have to follow those, they're not requirements. But if you don't follow them, you almost have to justify why you're not following those, those guidelines and the education notes. And there's a 2017, uh, education note on use of models, which, uh, almost certainly is going to be updated in the near future, uh, to, um, to up, uh, up the coverage in terms of predictive modeling and AI. Thank you so much for that. Very comprehensive, um, definitely gives students, has up on what they should do to prepare for the changing industry. So that actually moves on, uh, right to our next segment, which is staying future proof. Um, so you did mention a lot about how students can, and working actually as well, how we can all prepare for the future, um, but aside from reading regulations and what's already coming to your, are there anything, any other tips that you might have for students? So, um, maybe like a mindset or, yeah, just any general tips that you might have. No, um, it's definitely a mindset thing. Um, so, um, I would encourage people to be avid learners at a high level. Um, the best thing you can do is learn what you don't know, you don't know. So, um, a friend of mine is an accountant and, um, I like this story. So if you give me a few minutes, um, anyway, um, he asked the partner at the accounting firm, he was at a question and the partner went over and pulled off the book and started looking something up. And the guy's like, well, well, we'll wait a minute. You know, your partner, shouldn't you know this stuff off the top of your head? And the partner said, well, no, no, no, I, I, I've qualified as an accountant. I've got the experience. I have the right to be able to look it up to make sure I have the right opinion. You know, you're studying for your exams. You have to know what off the top of your head. So, the point of getting it is a lot of actualities come into the field thinking they need to just remember everything. Um, don't, don't keep that mindset. Like go learn things about, you know, whether it's, you know, in AI, whether it's neural nets or image recognition or, you know, different things that may not seem like they have a direct application, like image recognition, like, okay, some of these language models, how they're built up, learn about the concepts at a high level so that if somehow, you think you need to know more, you can connect the dots and then go in and dive deeper. I think it's better to have, and this relates to business knowledge, communication, have a broader skill set, you know, that 80, 20 thing, try to get 20% of a lot of different fields and trying to, instead of trying to get 100% of one, because, you know, well, then you're just a one shot wonder for this is if you have the ability to kind of see the map at a broader level and then go in and attack and learn more as you need it and learn fast, we all learn fast. We had to write these exams, right? So, yeah, so just get that wider knowledge and don't try to be an expert at everything, try to be a, you know, you need to be an expert at your job and your current things that you're doing fine. But more generally, like, learn a lot about different things, like, what are the underwriters doing? What are the marketing people doing? You know, what are the regulators of, you know, AI and pharmacy doing? Like, you know, they're doing designer drugs, you know, driven by AI models. Okay, how does the regulation of that work? How could that apply to the regulation of AI in our industry? You know, you don't want to be an expert in, you know, designer drugs to do that. But, you know, you can understand the concepts. You might learn something and apply it in a new area, right? That's often the, often it's not the big inventions, not a completely new thing. It's just the application of an existing technology in a new field. Like, the black sholds, you know, model was heat transfer models and engineering and then they found a way to apply it in finance and boom, you got a Nobel Prize 20 years later. So, you know, yeah. I think keeping like an open and curious mindset is always necessary for these advice for students, but I know we sometimes tend to go off on a tangent event and try to just focus on one single field. Yeah, so Tim, thank you for the valuable insight. We just have a couple more questions for you, a couple of last questions. So your journey has been very unique. You're involved in leadership positions in a lot of different communities, particularly the IAA and the study predictive modeling one, like the 19th and Introduction. So, could you elaborate more on the work that these committees do? Well, yeah, it changes by committee. So, the easiest one to explain is the Society of Actories work. We write and grade the exams in advance topics and predictive analytics that qualifying actuaries under the SOA have to take, right? So, that one's easy to understand by everyone. It's pretty straightforward. You know, we do review the syllabus, you know, should, you know, aspects of AI be considered and things like that, but that's a whole other story. With the predictive modeling committee with the Canadian State of Actories, our main mandate is to promote awareness, education, you know, anything from webcasts to, you know, pining on the standards of practice. We wrote a paper in response to Office of the Superintendent of Financial Institution. They're changing their model risk guidelines to encapsulate AI, cover start covering AI, and that they're changing scope as well. Previously, it was just covering banks and sort of risk in capital models and banks. Now, it's going to start being applicable to federally regulated insurance and pension funds. So, bringing the actuaries under their wing as well. So, so the CIA, the predictive modeling committee is responsible for anything to do with predictive modeling and AI, you know, in response responses from the Canadian Institute or, you know, within the Canadian Institute. The International Actuarial Association, just for clarification, because a lot of people aren't as familiar with the IAA. The IAA is an association of associations. So, the members of the IAA are the Canadian Institute of Actuaries, the Society of Actuaries, the casualty Actuarial Society, the IF away in UK. You know, it's the IAA is the global voice of the Actuarial profession. So, its mandate is to represent the Actuarial profession to these super-nationals. So, like, you know, the United Nations to the OECD to, you know, governments and things like that. And it also helps establish actual organizations in developing countries as well. So, but they created their artificial intelligence task force about a year ago and we're studying, you know, the impacts, opportunities, threats, a lot of different perspectives on what AI could do for the profession. I would encourage people to go to AIForActuaries.org if I could make one one advertisement. That's been my project for the past year, year and a half. So, or search for #AIForActuaries on LinkedIn or any of the socials that you'll pull a lot of information there. So, what we're trying to do is we are creating an online community for people to share information, have discussion, ask questions amongst the community. We are looking to bring in, you know, industry experts from outside of the actual profession to contribute as well. So, it's, you know, it's a small profession. So, if we can work together online as a global community, I think we're a lot more powerful that way with this, you know, very new changing, fast-changing and potentially disruptive technology. So, yeah, the IAA provides that high-level guidance. And there's another number of other things the IAA does, which, you know, like they have, you know, example standards of practice and they're updating their global standards of the standards of practice to reflect AI and things like that. And that's rippling through the organizations as well. So, I'll leave it at that. Yeah, I think that was a very insightful comprehensive review of what all these communities do. As students, we did mention we can go on some, like the hashtags and also just read some things, but are there any other ways that students can get involved with these committees? Yes. Well, most of the time you need to be a member of the organization to, to go so it depends on the committee, I guess. So, with the Canadian City of Actuaries, as long as you're, technically, you don't have to be a member. I was actually out of committee before I was a member of the CIA a long, long time ago. They had to have an executive vote to allow me to do it, but, so my point is put your hand up, show interest. If you want to participate on these committees, like, you know, the website lists what the different committees are, you know, just show interest. There's a volunteer page where you can express interest and volunteer and show you what you list your skills. And that's that's looked at on a regular basis to identify needs to staff other committees, but there's also work in groups and different other areas that the people can help. So, most importantly, just show interest. It's very different with the IAA. The IAA, you have to be actually sponsored by your local organization. So, my work with the IAA, I actually had to be approved by the CIA to represent the CIA in the IAA work. You can't just go to the IAA and say, I want to do something. You have to go to the CIA first and ask to do it. But again, show interest. Thank you for that. Yeah, I think that's great advice. So Tim, that brings us actually to the end of this episode. Thank you so much again for your time and incredible insights. I think Edible and I both have learned so much about AI and also the actual profession. And we hope that everyone tuning in did as well. And so to all our listeners, please make sure to check out our website and socials that will be linked in the description for more actual content and updates from ASA. And thanks for tuning in, and we'll see you next time.