Kartik Sakthivel, Chief Information Officer, LIMRA and LOMA and John Keddy, LazarusAI, discuss relevant AI topics including historical roots, current applications, rapid adoption, and the critical importance of AI and its growing influence. Read the transcript. 

Welcome to Insider Insights, where we dive into hot topics facing the financial services industry. Today, Kartik Sakthivel, chief information officer at LIMRA and LOMA and John Keddy of Lazarus AI explore the historical routes, current applications, rapid adoption, and the critical importance of AI and its growing influence.

Hey, John, good to see you. How you doing? Afternoon. How's everything in New Hampshire? Everything is great in New Hampshire. How's everything in Kansas? No major complaints. At least nothing you want to listen to. Well, you know, we could talk about complaints all day, but we are here to talk about something super duper exciting, John. You and I going to spend a little bit of time talking about, the thing that we've been talking about all through 2024, and that is A.I. and all things A.I.

So, you know, this is our podcast series, John. We've been we've been talking to our audience about timely, topical, relevant things, right? No more timely, topical, relevant topic than the topic of AI and all things Generative AI. So you and I have had many, many conversations in the past several months about where we've been in terms of AI, look, AI is not new in our industry.

Right. And why is it that everybody's talking about AI? Maybe you can start there and then and then just let's demystify the concept of Artificial Intelligence and what everybody's been talking about specifically Generative AI. So maybe just talk to us about why is it that now that we're talking about AI all day, every day, and in every single forum?

Yeah, you're so right. It's really in some sense, a new conversation. And so I do a lot of presentations with executives, including boards. And so far, the first AI conference I can find is in 1956, in New Hampshire, actually. Yes, I remember that. I wasn't there for it. Neither was I, just for the record. But certainly AI has been around a long time, and that's often why I start that discussion because executives feel I'm hearing that this has been around forever, but yet it feels very different today.

We're having conversations I didn't have 24 months ago; make sense of it. So what I found is starting with a basic chronology helps. And I said the first I can find is in New Hampshire in 56. But then data science, statistics, etc., cloud computing, which you cannot talk about AI without cloud, advanced over decades. And then especially in the last seven years, we've seen significant increases in understanding both in the technology and capabilities.

And then, of course, Open AI's ChatGPT is a blinding light. We'll have to look back in time to discern was it more blinding or enlightening? That's a debate unto itself. But it gave us as a society, not just the insurance industry, as a society, permission to discuss these things that have been going on for decades, building in importance, and now it comes bursting through into our economy and into our society.

and certainly the insurance industry is a huge part of our economy in our society, isn't it? No 100%. So, John, you talk about Open AI and ChatGPT. For those of us who don't breathe, live and breathe, Generative AI, AI, look, AI is a sprawling field of study, right? It's a vast, vast industry. People don't realize it, but we've been surrounded by Artificial Intelligence for many, many years, not just going back to New Hampshire in 1956, but every time you unlock your iPhone with facial recognition.

Right. That is a rudimentary form of AI. Every time you use S-I-R-I, because I don't want to trigger your device. Right. Or A-L-E-X-A, at home. That's a rudimentary form of AI, right? It's speech to speech to text. Every time you use social media, every time you use Netflix, every time you use Google Maps. We've been surrounded by AI all day, every day.

What do you think it was about Open AI and ChatGPT that basically thrust us into the common lexicon? Yeah, absolutely. And just to pick up on your theme, you are so right. So often when I start my talks, as I progress, I set insurance aside and say, and one of my former roles, I was a CSO.

So I say AI has been used in cybersecurity technologies for years, thank God, saved our bacon more than once. AI is used in our home security systems. My oldest son has a part time job at McDonald's. AI is used there. So you're correct all around our society. AI has been at the edges and increasingly coming into the core. What

OpenAI's ChatGPT did, though, is the conversation that you and I have been having, CEOs and board members and investors were like 'Yeah, I sort of don't get it. Are we talking about Bitcoin again? I mean, I try to study it. I try, I don't know if I get this.' And suddenly, with Chat GPT, they'd say I got this.

I understand that.' Now, by the way, we might say, 'Well, hold on. You don't understand all of it.' But there was this very visceral reaction to, 'Oh, I get this. I see what this might do. And where I don't see this, I understand the risk it might bring.' And that gut level, visceral reaction of CEOs and board members, in my mind, is the fundamental difference.

You know what? I think you're spot on. I also I'll also add there was the inherent ease of entry, right? It's really easy to use ChatGPT, really, really easy to use ChatGPT. It was also the cost of entry, right? Zero. ChatGPT 3.5 was zero. Which is why the adoption skyrocketed. What, five days to a million users? Two months to a hundred million users. Those are the kind of things that allow people to truly democratize AI and prosper in the lexicon.

You and I have talked about this before, John. My personal theory, again, is that anything that starts off in our in our personal lives, right, and then goes over to a professional life, has a lot more staying power than the converse. So while we're on the train, when you talk about insurance, you talk about the value of AI and Generative AI across the insurance value chain.

Look, you know, AI within the insurance industry has been in place for a while, right? Whether you're talking about some of those rudimentary chat bots that we had, right. Underwriting is a wonderful domain when we think about artificial intelligence, whether it's automated, accelerated underwriting, the hallowed ground of straight through processing, fraud detection and fraud management, as a form.

So you can you can attest to that. Can you talk about what role in your opinion is AI? And then also, GenAI, Generative AI, is playing in the insurance industry today. Yeah. So today there's multiple, very important use cases. And generative, as you said earlier, it's very important for people understand that it's just one field of AI now.

And I think last year it felt to me, I almost felt like that shirt 'Got Milk' where people were going around saying, 'Got AI'. And I say, No, no, no. AI is multi-faceted. It's not a thing. You don't got it or don't got it, so to speak. And so understanding those different components is just absolutely critical. Now, Generative AI, create something that was not there before;

an answer, computer code, an image, video, my voice, which should be scary, not only because you may have to listen to me more, but we can see this as threat vectors, right? If somebody can easily mimic our voice, what happens to all those institutions where we've used a voice for authentication? Suddenly we can see the risk.

Yes. So those are all different aspects. And again, generative is one field and generative is the one that one of the fields, rather, that has made the most progress, especially over the last seven years. That's really where we see significant impact and we're going to see more. Now, having said that, there's fields that I would consider analytical AI that are very important and vendors have been using those tools.

There are other fields that I would put as a subset of GenAI, you quite correctly brought up in life insurance underwriting, being able to take a complex series of documents and summarize them and then interrogate them as if you were talking to a person. Those are powerful things that we've seen just the tip of in some of the products and tools,

and we're going to see far more in life insurance. Remember, everything that we're looking at today and are impressed with, and rightly so, is the worst it's ever going to be. I love that. From here it goes up very sharply. 100% today is as immature as AI and GenAI is going to be, right. Completely concur with that.

Now, listen, what's interesting is, I mean, you've seen it over the past year, year and a half, right. The accelerated adoption of Generative AI across our industry has been pretty remarkable, whether we're talking about, as you mentioned, operational efficiencies, cost reduction, to be able to automate these rigorous, repeatable, manual processes. Right. Huge benefit with it in terms of summarization, huge benefits in terms of having a more humanlike interface between us and a machine.

You can attest to those. Right. So every single company last year, starting last year, had like vibrant, vibrant use cases of generative AI ongoing within their organizations. Now, earlier on, earlier on some of these use cases, and I think some people mad when I said that, but some of these use cases were a lot more artificial than intelligent.

Right. But in less than a year, we have seen these mature to a point where they are realizing some real tangible value, whether it's cost reductions and cost savings or whether it is being able to give people time back into their day and it's pretty remarkable to bear witness to. You know, while generative AI, John, has consumed all the oxygen, how is AI today improving the insurance value chain. Talk about,

talk to us about a little bit about not just generative AI, but AI in general. Yeah, absolutely. And I will start with generative because there is definitely some real forces there. So number one, use case that we see in insurance the most is still coding to be able to code. And I think that companies are also using some of these tools to a lesser degree to document legacy systems.

I see that as a huge opportunity that's being under leveraged, frankly. But John, none of our listeners representing carriers, none of them have legacy systems, right? That's not a thing, right? Nobody. I'm sorry, I thought this was an insurance webinar. To be fair to insurance, banks have this issue some capital market firms, airlines. Yeah, right. If your industry's been around over 100 years, I promise you, you have legacy systems.

Absolutely. And like I said, I think it is an under leveraged opportunities to use language models, even to least to document legacy systems and then to go forward and start preparing some base code in terms of rewriting them, not saying go take that and start slamming things into production tomorrow morning. That's not what I'm saying. But when you use these language models, especially business trained ones, to go through, do documentation and then begin rewriting the systems.

So we're seeing that in a huge opportunity. Secondly, is some of the things you and I have talked about, so let's stay with life insurance as appropriate for this forum. A lot of the technologies and tools today have embedded some element of AI into them, like you said, around medical record summarization, APS summarization, things like that. But the next class of tools we're seeing are so much more powerful, and literally will take a complex series of documents and X-rays and EKGs and just let someone interrogate them as if they were speaking to a person.

And you can see, and I apologize, I'm going to switch to a future opportunity, so you can see if you can use that in a multi-mode environment, could you also take those same technologies, and apply them to enterprise search management? What is the fortune in the amount of time we've spent trying to get Google search appliances to work, to get SharePoint to work, get all of these things, to find things in our infrastructure,

and now tools are available that we can begin to do that. So that's where I see both some actual use cases today, but just huge opportunities for us as an industry right in front of us. I completely concur. The ability for us to be able to surface our institutional, enterprise knowledge and have it be accessible to employees is going to be pretty tremendous.

John, something we should acknowledge, right? So while generative AI, most organizations have use cases ongoing, at LIMRA and LOMA, we have the governance group with 72 business and technology executives representing over 40 firms across the across the industry diaspora. And literally every single firm has some amount of use cases currently ongoing that are actually seeing some real value. Now, we should also acknowledge for our listeners that generative AI. as a technology. is being baked into most of our vendor products, right?

If you think about Office 365 has I think it's called Co-Pilot365, right? You've got Adobe with a bunch of these tools with AI being incorporated. Windows has Copilot AI. Talked to us about what that future, in your opinion, what does that look like in the next year or so? Yeah. And I will address this.

I want to first address the group you just mentioned. So I want to really applaud, as I always want, to LIMRA, how active you are in doing really important things. And I will point out, however, not every insurance company is spending time and effort in this. I'm very fortunate, in myy role, I get a very broad view of the industry.

So we see people implementing AI today. I gave some examples earlier, we see companies doing proof of concepts, we see lots and lots of ideation and there is a fourth category of companies that are not doing anything. It's the ostrich strategy and it is extremely dangerous. I believe it's the highest risk strategy you can take in today's world.

So again, really applaud what you're doing to drive forward and to do that and ensure that people don't deploy the Oxford strategy. Now, when I'm talking to executive teams, one of the points I make is you have to get educated on this. You cannot, effectively, stick your head in the sand because your vendors are employing this today.

You must come to an understanding of these tools and technologies, because people are making choices on your behalf. And if you can't be an educated consumer, you are putting your organization at risk. So, not engaging on AI is high risk in the age of AI. So I strongly encourage every insurer to come to basic understandings of these tools and technologies.

Think about and understand the choices your vendors are making on your behalf. Be educated consumers so you can say, I'm not comfortable with that decision, or that tool, or that product. Or drive them to say, 'Why aren't you doing more here?' Because it seems there's opportunities to deliver value. So because of that linkage between vendors making decisions on the behalf of insurers, it is evermore imperative that insurers are educated on this topic,

they're educated consumers, and they can ensure their vendors are making the right decisions. Completely concur, John. So I think two things, for me: number one, AI literacy is going to be important. We have people who still don't know how to successfully Google for things, right? So, to be able to capitalize on generative AI, you need to know some basic understanding of how to interact with it.

Did you know, Jamie Diamond? Jamie Diamond's a smart guy, right? At JP Morgan Chase, every new employee at JP Morgan Chase, has to go through mandatory training, right? Some level of rudimentary AI, basic AI, AI literacy training. I mean, they're onto something here, you know what I mean? They really are onto something. So I think AI literacy is going to be important from a 'protect your posture' as well.

You don't want your employees posting intellectual property onto public activity, for example, right? And these things are moving so rapidly, and moving so fast, it's going to be important to have a base level understanding so you can continue to grow your knowledge base. So, John, I think this has been a great conversation for us around the where we've been with AI and where we are today.

But there are some there is a dark side of it as there is a yin to every yang, right, or a yang to every yin, you know. Are you going to come back and talk to us about the challenges with AI? And then we can we can also end of a high not, and we can talk about some of the future applications of AI within our industry, how does that sound?

Always a pleasure to engage with anybody in the LIMRA community. Wonderful, you're not just saying that because I'm here, John? Mostly, but I very sincerely believe that. Awesome. Awesome. All right. Well, we'll pick up this conversation next time, John. We'll talk about some of the challenges, what organizations need to know in navigating the age of AI.

We'll talk soon. All right. Talk soon.

Thanks for listening to LIMRA’s Insider Insights Podcast Series. To hear future podcasts, subscribe at LIMRA.com/podcast.

Subscribe to Insider Insights Podcast