Hashing It Out
Hashing It Out

Episode 75 · 2 years ago

Hashing It Out #75-Truman Esmond- VP of AAIS

ABOUT THIS EPISODE

On today's episode, Corey and Collin talk with Truman Esmond, VP of Solutions and Partnerships at the American Association of Insurance Services (AAIS). Why are we talking to insurance folks, you ask? Because they've partnered with IBM to launch the openIDL platform to help aggregate company data while maintaining privacy and compliance. We ask how it works, why they used Hyperledger, and many more details, enjoy!

Links:

Sponsor

AD Music: Chad Crouch- Algorithms

Donate to Hashing It Out

Discuss

Now entering work. I'm got a sponsor for you this week. This week's episode is sponsored by status. Status that let's chat. Browse Chan's act when the etherium blockchain. Take control of your own private, secure messaging steps on mobile and secure your assets. Download the APP to day, where you get your mobile apps, or at status. I am get at status, I am slash get bitcoin. PODCAST will also be in the TPP channel of the status apt to give out a little sant, but you play around with the features. Start chatting privately today. Enjoy the show. Welcome to hashing it out, a podcast where we talked to the tech innovators behind blocked in infrastructure and decentralized networks. We dive into the weeds to get at why and how people build this technology the problems they face along the way. Come listen and learn from the best in the business so you can join their ranks. Welcome back to the show. Passing it out. As always, I am your host, Dr Corey Petty, with my trustee cohost, Callin Cuche, say what's up. Everybody Colin. What's up? Everybody Colin Nice daily than you. I know you don't like that. It's fine. Let it's Lide. I wanted some his as and there wasn't enough and the trying times of today's landscape, I'll let it's live today's episode. If we have Truman Esben, VP of solutions and partnerships of AAIs, and there's quite a bit more there. So thank you for coming on the show, Truman. Why don't you start us off with the ordinary. Tell us what you do, tell us how you got here and we'll Bo once their. Great thanks for having me. Appreciate thanks, CO income, and hello everybody on this very interesting game. So if I'm with the American Association of Insurance Assurance Services, shortly, short candid AIS. Very quickly, what a diis is is a ninety year old advisory organization to the Insurance Space and we are a not for profit member association made up of about seven hundred and fifty insurance carriers on the property casualty space, so that's home auto, business owners, bar owners and a lot of other coverages like cargo and things like that. That can be very interesting guys, as it in our conversation. But where is night year old organization? And we have three major jobs in the first one is being a stad agent second one is a rating bureau and the third one is that advisor organization. will explain those very quickly. Are as a state agent, we collect data from the insurance carrier members, summarize it, aggregate it and and report it as in a mediary to the state regulators. Insurance is the state based regulatory ecosystem. We take that day to summarize an aggregated report up to the regulators so that they cannect get understanding of what's happening with insurance coverages and their jurisdictions, typically down with the visive code level, so they get an idea of what's what trends are in coverage. Typically run things like fire and water damage and things like that. That's the physical agent part of it and that is a license function that we provide your audible subject you audit by the regulatory bodies. Second part is we are a rating bureau. So we take that data that we use to report the regulators and get a little bit more and the reflect that data back to the industry in the form of standard lost costs. So that helps regulators as well as the industry understand trends at again at as a cod level for different types of coverages, fire flood things like that for losses, as well as coverage activity. So as new coverages come available, that reporting helps them understand how well the market is meeting those needs. The third thing we do, that's the rating bureau. The third thing we do is in advise you organization, we actually create standard insurance contracts, policies,...

...form language, the underwriting manuals that determine exclusions and inclusions for types of coverages and the market space, and we file those with the different states as vanilla sorts of insurance programs that are members can then use when they don't sally fail have experience in that product or in that in that jurisdiction. So all those things together makes us the data driven mechanism for the way the property casualty insurance industry moves and changes over time in a nice data driven way, and how the regulators are involved, to put making sure that they understand what's happening in the industry. Personally, a background is I've been in technology and marketing and the variety of different things over the last twenty five years, building network solutions in a variety of different industries. Joined Ais about seventy a half years ago as part of a really massive transformation of the Organization of this Agnes. You know, the time was about eighty years old, very traditional devisory organization or Insurance Organization, everything you would expect. We are now a much different organization, taking that purpose and applying modern technology and really disintermediating ourselves from that role in order for the industry to benefit and for and for our for a IUS to be able to do our jobs around the value to the industry and a much better way. So I think the obvious question here from from from that introduction, which is just thank you for that for extences, is why the hell are you going to show I are we talking about you regards to a show that talks about technical aspects of decentralization technology. What's going on there? But so we were in performing this role when I joined the as we really were looking to how we can do our job better and looking at this organization and our function how we would if we could, knowing what we know, how would we do it again? How would we do it differently? You know, the process that we use to do that reporting function is very old school, if you will. In fact, he's going to have six states that receive that data. We provide an aggregate and provide them in paper. So that's a very inefficient process and territ not terribly useful, and the regulators are doing a different process to change that. So we needed a way to allow the industry to work differently around data and it had to be a different sort of data strategy. And when blockchain is a technology matured, particularly in the enterprise space, we realize that this is the hammer we've been looking for. So we weren't, you know, kind of looking at it from a technology forward position. You're looking at it from a business solution. That we that required the promise of Blockchain as a matured data privacy, trusted can act, connections, structure transactions and mutable ledgers, all those sorts of things. So when we when we were exploring this for a few years and hyper ledger matured enough, we needed that sort of a platform. This is an enterprise space as opposed to a truly public space. It's open and that everybody can participate, but it is a enterprise space and that those closed orders, even though they, you know, the regulated body of the country or an industry or a you know, at least a mon anonymous community and then we can even more an importantly, our community of community. Each poems are going to make different rules and this is not one that drives, drives outward, but it kind of supports everything. One of the things I say in a calm you didn't say this, is that insurance doesn't drive anything, but it supports everything. And when insurance doesn't work, it's not available, we get delays and delivery of new technology like UBER takes years to get to the ground and get a general adoption, when it's unclear who takes care of the passenger in the UN an accident or those sorts of things. So when Hyper Ledge in maturiting up as exactly the model we wanted to follow, we're very, very big believer in the open source and community to driven standard for strategic platforms,...

...for trusted platforms across different communities. Very much in the Linux Foundation and then engineering task force, where my boss came out of. So we're very much in the the open source world. So we were looking for a platform like that in order to allow all of our members to participate, given that we've gotten not only from the largest insurance companies but some of the very smallest ones as well that have, you know, very small off in six figure or less operating budgets. So it needs to be very, very inclusive. So we were looking for a solution like that and we so we created a technology of block chain Driven Platform Leveraging Hyper Ledger that would allow our members to basically keep control of their data and participate in regular Tory reporting in such a way that didn't require the implaicit trust in ais as the steward of that data. So they would keep their data locally in their fear node. We would have a trusted interaction with that data there basically allow us to make a query of that data and we could take the information resulting from that quarry, summarize it, aggregated and report it up to regulators, because that data is in that pier and a trusted fashion. We know how it got there, we know what the standard is and you know how far we can trust it because of the assertions of the the insurance carry you put it there, which is the level of trust we need for this regulatory reporting. These case, we were able to streamline that and basically automate ourselves out of that process, but also give us a way to expand that process, make it faster, make it more valuable, both to the participants as well as to the the regulators and really just society, so that let's to create the open idel or open insurance data link. I also say it's the open individual data link, of the open the story data link and and as a basically a data privacy platform. I say, and calling you and say it again as well. Only first meant that I think we solve data privacy only in this place, first for insurance regulatory reporting in PNC. But the model and the architecture that leverages a enterprise botchain technology at its core allows for a community, in this case PNC carriers for a regulatory purpose, to consolidate, agree on rules and have that transaction and dator reporting be far more streamlined and far more useful when we now have a trusted data stream of policy and claim experience, which is all the stuff that we have. You know, our house, our phone, our car, our travel, our media publishers all have insurance policies in those are all part of the data stream that we have. So the where I think it becomes important is when, now that we've got these physical assets tied to financial assets, insurance risk. We can then leverage that community of known pool of assets and then start to bring different sorts of interaction patterns, if you will, or or different types of community relationships across that pattern according and get it some agreed upon rules. Our challenge has been looking at from their trying to boily ocean. I think looking at it from what's the biggest possible impact we can make, as opposed to looking at where we can get some real traction in against specifically regulatory world. Long answer against are that was great. That was great. Okay, so let's let's break this down so to an audience that can understand the effects of what you're saying, if not necessarily understand the business side of what you're saying. And when you talk to me, you you have some real, really cool, direct applications for what you've just built the whole data privacy side of things, as well as the fact that you can kind of share in these federated, like you know, communication channels that enable you to sort of like get these sort of like anonymized high...

...level pictures of what's going on at what risk means from an insurance standpoint. Some of the things you talked about was automatically doing spot insurance, for like while you're driving. So I was wondered if maybe you could give us give some real concrete examples to ground what you just said in the vision of the reality of trying to build. Yeah, thanks, that's avide it. There's there's really a million different applications you can provide and there's a lot and a lot of them comes down to the the empathy with the with the stakeholder that we're talking about. So one of the one we can that I had that I mentioned a lot. Living in Colorado, I drive a Jupe. I like to go for wheeling. I am insured by Statecom when I'm on the highway and you know, and if I do something to my car while I'm for really you know, I may be able to get paid for but I've better be able to get it back to the way things like that, and they certainly are going to pay to recovery me if I roll it down a hill, for example on a trail. What we can start to do in terms of insurance products once we have that sort of trusted information in a in a private way. So all that information that's being gathered about me while I'm driving, my location, even things that are very detailed, but where my eyes are looking and so that it can rumble my scene to hey, look at the road. That information is being captured on my phone and the car and all that, and and we want it to be private and only used for very specific purposes. And I looking example the very good one where your eyes looking. It's fine, okay, to use it for brumble my seat, but I definitely don't want that data to be accessed after an accident, especially even if it wasn't my fault. But if that data shows my eyes weren't looking, there regards to whose fault it was, state farm is going to be paying for that, for that claim if they can, if that data is, say, suppeinable and can be brought to court and demonstrate that I wasn't looking when that accident happened. But in terms of the benefit we got to protect, it would be required data privacy. In the simple example of going for a reeling, my location is private, but my policy information is known, my driving history is known, the kind of power I have is known and stay farm may choose to offer a an add on coverage for recreational vehicles offer, driving, camping, things like that that would activate or give me an offer when my location interacts with a you know, is proximity to a trailhead, for example, which obviously is available from Google maps or something like that. I could say, look, you left the highway, you are within a hundred feet of the trailhead of this three star trail. Trigger the offer. Say Farm doesn't have any idea where I am, but they know that based on my car, my year, I experience, my anything else and my proximity of this to star trailhead, using a source that they trust. You do tours, USA, whatever. They give me a rate for here's twenty four hours of scratch in dent coverage and, you know, emergency recovery and if you do something too bad, maybe you get a rental daily driver. So for a hundred fifty bucks, you have a hundred. You have an offer of twenty four hours of off road coverage. I say yes, I'd like that. I drive a hundred feet in order to activate that coverage. So because I can't buy it after I've rolled it down the hill, for example, and as great your coverage is activated, maybe take a picture of my car at that point and off I go and then from there it can maybe offered different things if I travel out of a radius or if I've left the trail. If they get you, has available or could you say, Yep, see you twenty four hours or put the button when the coverage is done, when you're back on the highway, you're done with your trip? Those are kind of things that could be available. But there's a there's literally, you know, thousands of different potential interactions. As share of the Applications Governance Committee for Open IDL, my job in that in that group is to explore the potential applications of this privacy technology and architecture and building upon those policies and claims data streams that we have for different both,...

...you know, novel insurance applications as well as applications across industry and new, new, new ways of that could be used and applied, not only in the in even Hyper Ledger Babrid Technology, which is what we're using today, but taking that same data architecture and applying it in things like zaw to to allow more for that case I describe where you've actually got a note participating on a phone or on a watch or inside of a vehicle in a in a more disconnected way, allowing for that to those trusted interactions and keeping the data privacy closer to the edge of the the application I have listening to that just has my wheel spinning in a myriad of ways, mostly towards, mostly towards the what strong guarantees you have around the privacy model that, like the underlying, like fundamental technology is giving you, and what you can extract from there based on like reasonable risk assessments, because in order to have, if any type of like claim being offered up, you need pretty strong guarantees around the associated risk so that you can make way decisions, especially automatically, because this is potential to be massively scaled because she's to be automated. Like how are you going to like, what guarantees do you have based on how the privacy actually works under the hood? You talk a little bit about that. Absolutely, so that really the do things to understand is trust, and trust is different than faith. Right, it's not binary. It definitely requires context. So what we're when you talk about the guarantee, if, like you guarantee this, you know this process this far, and one of the quotes I use that my wife definitely doesn't like it, from Charlie CEA and maybe fields, and trust me with your life, mature money or your wife. You know, there's a there's a definite distance of how far trust can be cleaned. It can be a very implicit and deep level of trust, but only in a context, right, not in the casino, not somewhere else, but on the battlefield and a foxhold right, for example, same sort of thing that we're talking about here in this context. That data that's being used, in this case our first use case. The data is in the system, within that testation, if you will, by the data provider. In this case it's a Yep, I say, this is what it is for the purpose of regulatory reporting. Right. That that is a fairly low bar of trust. Rights something they're compelled to do so and it and the penalty for not doing it right is fines. I mean if you're really flagrantly not reporting and maybe you get an odd or something like that, but the the it's not, to use your case, you know, taking a trusted binary piece of data and paying a five hundredzeroll house claim on it or a million dollar liability claim on it. Those are much different levels of trust and but when we've got the data stream and the blariat reporting, example, the amount regulators trust that data they're getting today, when it's only being used and only being served to them in that one context, they've got to take with a grain of salt. Right. They don't have any set is an extract of a system that is, you know, closed to them, unless they did call for an audit, which they don't want to do, the very expensive exp intreases, and certainly we're trying. That's one of the things we're hoping to streamline, you know, so that instead of managing and enforcing by audit, they're managing and enforcing by DASHBOARDS and clearer guidelines. So we're hoping for a much better regulatory environment the future. But the the privacy and the guarantee of that is in the context of it. First of all, the guarantee is really provided by the data owner. Right. The idea is putting the control back in the hands of the data owner, in this case is the insurance carrier. They receive, you know, their own policies and getting claim information and their node is completely in their control and using their cloud environment is in their basement. It is in right, you know, on a server...

...that they control on their own hardware, whatever they want to use it, or platform leverages in coming that even in open ships, so that it is very componentized and can and create a trusted system on that pier of the different packages or containers that are that are linked together in a trusted way, as well as a code that runs on it, so they can deploy that wherever they like. They obviously are still protecting their data in their own world. So if they get packed, if you will, you know then that that node maybe at risk. But there's a lot there's that. That's pretty deep piece in the data that should be resident there, at least for the purpose of records are reporting. Isn't that private? Doesn't have names, doesn't have addresses, at least today, and down to the ZIP code level. It's not endorsement level. So it doesn't know, for example, when the know the exact moment that a coverage existed or something changed in the policy, you know whatever it happened to be. So it's not at that level of detail. But once the platform is trusted and people and companies can start to put more detailed data into their trusted peer, allowing it to be interacted with, you know, at a very permission basis, right, so they don't have to give their data anybody. They put their data into their peer in a truck, in a standardized Schema we called a harmonized data store, as the first of the patents we we have with IDM. That leverage is a very specific standard that that AIS as a regulatory reporting in your mediary defines and it and it's going to be open source. That Schema and the rules for data coming into that harmonized datast or are defined in known companies can put data in there according to that Schema and then as interactions with that data are requested, and this is the second of our patents of IDM. Is that the regulators? Great? Okay, I want to data call. I want to know how many, as a commissioner from the rat North Dakota the other day said, I want to know how many home daycare policies are out there. What you know in some information about them. That's a changing risk in our state. We need to have more up to date information about what's happening with home daycare be read the insurance industry. So there the craft a question of and increasingly work into it to empathize with the industry about what the right question is. That will actually add value. Why they're asking it so that the carriers are more likely to participate. The draft that posted to the audience, the audience and carries is okay. I would be going to do that once they get, you know, enough agreement. I don't want to say consensus because I'm sending something very different in our world. Once we get enough likes, in this case, the the regulust is okay, I'll issue this data call. Then it's issued formally and at that point what we call the interaction pattern or the extraction pattern is that pattered element. Is Basically the query of how they specifically how your data will be touched in the context of this question. The regulator asks. Right. So it's code level. It's the middle part of the smart contract, if you will, that allows the carrier to know exactly how their data is going to be interacted with and that will run that on that query, that interaction pattern runs according to the rules. Is it happening at midnight on January, you know, one January first? Is it happening every day? Is it happened as soon as I click. Okay, how that works, and that's what's going to happen. That that queries you will run against my data. The answer, the response to that query will be delivered upstream to the aggregation point, in this case Ais as the intermediary operating the basically in a mediay node, to pass that data up to the regulators. So the data itself isn't identifiable to Aius. We know that this particular carrier has participated. They both we know they've consented to the data call. Again, that consensus, but they said yes, I will agree to participate. We know when the data call fires, that you know the day it act keeps successfully and we know when data moves up. But we can't we never see the row data and Aius we never, we can't identify what...

...data comes from which carrier. So the privacy is maintained from a fundamental structural standpoint and their security is maintained because they can always not participate in the data call and if they and if nobody participates, the regulator can basically force it. But we're trying to get away from more of a Pote Force and actually understand that these two communities, the insurance carriers and the regulators have a common interest of a, you know, a flexible, resilient, more increasingly fluid insurance market for everybody involved. Like then it's the course, you mentioned a patent. You guys had the harmonized data store, and that's bounds. It's great marketing right there. That's good. That's that's a good part. Right. Yeah, very much. Describe that pattern in a little more detail exactly how that works and what what what you guys are why this is a novel invention and know how utilizing it. Yeah, so it's, you know, it's not rocket science, but it is an application of an idea of data abstraction, something that's very old or, you know, very common in the IT and system development. But what it does is a couple of different things. One, as I mentioned, the the Common Schema, which is not just a database Schema but also the rules of ingestion around that data. So how the data lands in the harmonized data store. That is a standard and obviously going to be multiple standards that are defined by the different communities. That will that that will that will govern that. So we're not in the business of telling auto manufacturers how they need to capture telematics data on their on their hardware. Right. They're going to figure that out. So it standard, like mobi would, would be adopted into the general framework as a way the data to be stored into a database, say we use for our purposes and going because it's component has the harmonized data store. The execution of it, how that is applied, could be in any number of different data stores. These mango DB today for broad application. It's on structured data, so it can be highly flexible and it's not an analytical use case at the point of capture, if you will, while it's resident in the harmonized data store. So we take that, that harmonized data that exists in mango and that's in in our particular case, and that's the trusted data store. We know how it got there. We know it all the the inputs of them credential that's met our quality rules. It's been gone through the ETL process in order to be resident in the harmonized data store or in an a trusted scheme. And I trusted you in a level of trust and some of the things that will start to talk about our things like trust scores, you know, like that data observed by human being and type the manually? Was it observed by trusted hardware and captured in a more objective way for exam yeah, and then's call it. You said Etl, but like this is basically etl at scale and with arbitrary data. Sounds like and just actually, yeah, it's essentially etl and it's very compount except you've built a scheme in front of it so people can do create their own standards and perform the etl operation in a way where they can join and unjoin a network and still be able to, you know, handle the etl from any data source that supported by that particular network. Now, are these harmonized? Are The schemes of a way harmonized? Is kind of a strange where it I was maybe, maybe you can like explain to me why you called it that, because the each data store that is resident in each insurance company, in our in our first two cases, node is singing from the same music for right. You know, they're there there. Well, they may have one policy in one claim or zero and zero, you know, or millions. The structure of that data, the rules...

...by which that data changes in is up there. The smart contracts that updates, that data is consistent across the rules. So, you know, because of that standard, the harmony across those data stores allows the trusted interactions without having to share of the trust of the private data, whereas if you don't have that standard and agreed upon rules by you know that have the data's got to be so good and on the ride ride this ride, without that agreement, you have to share the raw data and even then it doesn't give you any sense of the providence or the or where that data originated. Without again sharing more and more data, more and more data, more and more exposure, which is just really, really weighty, and a lot of the other platforms kind of reinforce that complexity. I mean you need to do that today. You know, for example, if you don't have a standard, you can put your data in an an Azure ledger database. Essentially you're an AWS ledger database. But without the standard you have no assertion of quality. You can just say it hasn't changed. Well, but if it's, you know, if it was bad when it went in. I mean again, how far can you trust it? And that's really the question that we're trying to define so that the harmonization is a getting question with how far you trust we can off sing in this key really well, but you try and change it. It's not that may not be the right application and in the one thing it's in where the the blockchain piece comes in, and I don't think I mentioned it yet, is that as the data is loaded into that harmonized data store, let's say it's a day's worth of policies and claims transactions, the evidence of that data, that one day's Cube, whether zero transactions or a thousand transactions, is loaded into the harmonized data store. Evidence of that cube is shared on in the hyperledged fabric world on a channel, a ledger, in this case with as. So we just have evidence that the that a day went by. So we have evidence of, you know, no date, no policy to happen that day or no claims happen that day. Or we have evidence when we go back to ask the query of that data later, that the data has its integrity. That Hash is extremely lightweight. It's not a it's an identity ash, it's not a it's not something that could be repress engineered back to the data. So it gives us that that that evidence chain without the need to continually copy data as it moves across the streams and because we have those evidence dreams, as when we do the second party extraction patterns, that extraction pattern creates a second cube of Metadata, you know, interpreted data on top of that broad data set, averages, totals, counts, things like that. That Secondary Cube is also has to share as evidence, again for its integrity. And then, using a again an open source but called the private data collections is inside the hyper ledger fabric, that information is moved, that subset, that Metadata, is aggregated to ais in this case from all the different peers into different private data collections on our side and aggregated basically unliked out sort of fashion. And then, once the data is moved the intermediate point, which is a new featured hyper ledger, those those private data collection can be destroyed. So we have an evidence of successful transaction. But we can get rid of all the intermediate weight at each point because we only care about, for example, the first quarter of two thousand and twenty data until we get second quarter numbers. Just to like reiterate on that, if I if I'm trying to try to understand this correctly, you've created a way in which people can attest to standardized types of transactions that are then subsequently auditible. So the blockchain piece of this is attestations of behavior which is then auditible. Does that? Testations don't show anything about the information. But but you would like to, you know, query it. You can...

...provide the information and a proof that it happened at any given at testations that correct correct correctly, you can see, you know what question was asked against what data. That was true at the time and we know that it worked at time. Now we can we can verify that it worked correctly at that time. But of course, when we talk about things like GDP are and information, the underlying data can obviously change. So if I was to run that data call, you know, for the month of January today, you know, with some mytorical information, right, what happened in January and whatever, and then I run it again tomorrow on that same time period, I may get a different answer, even though both, you know, both were answered truthfully, if you will right. So we got multiple versions of the truth because the underlying context change and we need to be able to understand that. So before all kinds of fun applications when we start talking about things like ai stuff like that. So we both things. Need those types of tools and data sources. Both need to be network flexible and accountable and a jet and objective to this is where we start to do some really fun sorts of applications that are more infrastructure, which is definitely where we're going, and I imagine that's why you chose hyper leisure in this context. Is so you have that I don't want to call it a mutability, but ability to pivot based on contacts, as opposed to always having specific date data locked in some data store that's then a creed upart, like open systems like a theomor pick order exactly. It's very flexible, you know. It's community driven. You know, we're starting from one point in a general framework and seeing out how far we can ride the few the cycle before we need to put another tire on it, you know, and that's really the exciting point. But the key, you know, and hyperligees a lot of reasons we went with that with so many of the community projects that are out there. You know that by having this on a peer, this policy claim information on a hyperledge of peer that gives the ability to interact with you for her potential other networks, all kinds of cargo networks, all kinds of other emerging networks in the hyper lugged budger community. And if you look at them that are out there, you know all of them are looking for a lot of them seem to be looking for traction and we're talking with a variety of them to help them get there because we start to bring them into the real you know, into a regulated framework. You know, you know when we talked about before, the much like the financial community, the insurance community is a regulated framework and it's not going to allow itself to be disrupted. You know. Now, is it moving too slow? Yes, but but it's it's still got to be the stable and untrusted ecosystem that that we need that work with industrial control systems for a while. That's, if you want to know, slow. That's slow. Yeah, yeah, we laugh. We're we talked about in being in a race of turtles between insurance carriers and regulators and we're just streaming at them on the sidelines to try and go faster. The regulator seem to be running. For the record, the Nice thing is that you can take two weeks off and go on vacation and come back and they're still in the same place, more or less. We've been at this a couple of years now. It'll be two years in the spring, and we've where we've got traction and broad interest and participation from some of the largest carriers in the country, from state regulators, wearing an entire right now. With North Dakota, that will be our first real, you know, regulator, issue data call on the platform for North Dakota homeowners and it's not, you know, terribly exciting market. Obvious North Dakodea not, you know, not that. Not milliing do millions of homes up there. But the Commissioner is the chair of the Innovation and Technology Task Force at the National Association of Insurance Commissioners. So basically it's the federals assembly of all the different insurance commissioners of our state based insurance market and he helps lead the innovation and technology that those commissioners and regulatory bodies can adopt. So we're getting some traction. These insurance commissioners are also insurance commissioners...

...for life and health and are typically in the division of Finance and definitely regulatory affairs, but financial regular regulation in those states. So we start to get to a pretty quick overlap of where all these things start to touch. So I got a question. So you're actually a lot of traction for a, you know, a decentralized sort of projects right now, and that's that's good to see, even though it's not kind of in the space we typically focus on, which are more open protocols. It's really nice to see that same language, the same behavior, that same concept of being able to join and unjoin to bring more custom Taylor, you know self, sovereign data control to individuals. This is good. So I like that. How are you actually doing this proof of concept, though, because my experience in the past has been if you are going to show somebody at that high level something, they need to get it within the first ten to fifteen minutes or they are gone, and you must have some pretty good proof of concepts and demo workers and to actually get them hooks onto the idea and engaged in the process. What have you guys been building to actually show this off? So it's it works now. The technology was, ironically, kind of easy part of it. The harder part is getting people to to listen, spend time, understand and and they'll be empathy across the stakeholders. So we know the technology that I said it was. We were able to build that and able to solve through two major releases. The first one we built obviously experienced a lot of challenges. You learn quite a bit in round one, and then in January last year we released what is our long term architecture on the hyper ledger fabric framework. So that's that's we're moving forwards. At that point the technology was largely solved and then it was getting the humans to figure it out. Everything we have done is been leveraging design thinking and bringing the stakeholders into a room to help them understand their problems and get to the point of where we can agree again. How far are we able to trust each other? where, where to our common interests extend and when do they diverge? And that's really been the critical piece to getting the the understanding and in the different constituencies together to do that part. And that's the it's been working ever since. We've had working proof of concepts, working technology that can pop open the hood on show you exactly how it works, how the data is kept private, I'll you know, a little plug. If you go to AIS onlinecom open idel industry tests, drive is putting your email. There's a series of videos and very specifically the I think the middle three go through the architecture and great detail, goes through the actually how to stand up and node, and then the third one actually goes through the data, called the data experience, how the data is moved, kept private and ultimately reported back out to the to the regulator. So the technology has been working and been mature, if you will, and stable for for over a year the but in order to make it work, put it in production, it needs real data to be brought into the system from the carriers, which is obviously a sensitive process in and of itself, because we're talking about blockchaing, because we're talking about very large carriers, because we're talking about all these different things, including in the relationship to ais. Has We wind up needing to get approval or at least understanding and an okay, let's take, you know, a little tiny step forward across the entire sea suite of a company, you know, data, Innovation, technology, underwriting, compliance, finance, all those different areas of the organization need to be in involved because we're talking blockchain it, you know there's going to...

...be some level of media attention to everybody's got to be able to answer a few basic questions, which takes some work. And then we get then we get them to agree to a pilot with all the understandings and data contracts that we need to have, even though we never see the data. This is a foreign concept right today's constant contracts and data agreements don't work like that. They don't assume I get to keep my data and you get to use it. They assume you're taking it. So it's a different sort of agreement and getting some folks to understand that it's very challenging the source of contracts and agreements across the parties. And then, of course, the regulatory side. We literally in some cases have to get laws changed in order for regulators to accept data in a different way or make different sorts of policies. And you know, for those of us that speak in an addial world, it's a lot easier to create effective policy if you know why you're doing it and you say that instead of saying turn left, turn right, you know, and being much more instructive or enforcement based or letting folks go in particular direction and then slamming the door on the once they're down there. So it's much better to say hey, we want you to, as regulators, stay inside this guard rails, if you will up doing insurance to my state, of your concentration of risk, of your pricing in Dene outs, your claim payment ratios. You've long you stay in here. You're cool. And if we'd that data that they can see is trusted, they'll be less subject to audit, they'll be able to get more, allowed to be more innovative in their space, you know, doing more of those types of innovative products that leverage my personal information without putting me at risk. So let's talk about one of my private data. Yeah, personal information. And clearly you've said you said a couple times, like we're you know, I am an operator and they could add things that, for instance, check my attentiveness while I'm driving. Did my eyes steer off of your throat or something like that. It's creepy, it's it, but it doesn't have to be creepy. It just when we all think about it. Because right. So how? I maybe I missed this in something you said, but how are we actually gatekeeping in siloing personal data from the rest of the world without like am I encrypting this data's like it's talk to me as as the engineer that I in, a background that I have. I'm in a car. They're tracking the state as a data stream. Okay, it's a data stream going out. I'm deciding to either report it in batch or I'm actually streaming the data live using fight G or whatever. So another road play intensiveness. How are you actually making sure that somebody can't view or unlock that data or monitor the stream they how do I know the NSA isn't monitoring my stream? Is I'm driving down the road like is? Are these things we can press against, or is that a rep in irrational fear? or how are you do not at all. Again, the question is how much you trust that particular thing, because what you're dealing with when you can, when you're driving a car, you're actually participating in any number of communities of data right. You know the car you're driving, drive a jeep. Jeep has got a series of data networks that are monitoring everything from my brake system to the fuel line system to the air conditioning system and all those things are basically little networks. When we when, depending on what you're trying to do, as you're driving your car, before we even talk about it going up stream and then it broader things, you're going to have trusted networks in your car that will allow me, for example, to go from the sixty five mile and hour fast lane, where state arms my insure, to the hundred and ten mile an hour super fast, a v Lane, which cars are six inches apart and they are aware of each other, they are aware of the road and may have met a very high threshold of of function. Right. You know that the brake system is intact, the FUELIAN system is intact. There's it's not overweight, it's not doesn't have a weird drag coefficient, of all the sensors are working properly. That's going to have to be trusted at a very high level, just in that community within...

...the car among the caught, the different community of the cars among each other, the different community of the car talking to the road. You know, those are all three different things before you've even, you know, gone into a broader network of using, you know, g to take some information and have a broader understanding of them. I'm looking at the road, if I'm a commercial driver, to rate my performance as an employee, you know, more broadly, but so that so that can day to control winds up being happening at a very grandular level, all the way down to what's on by watch or again, what's happening inside that car, and to the degree that that data's being, you know, sniffed or swept up by something beyond that, then it's a question of how how much integrity that local system is that is providing the raw control of that data, how it lives, how it's brought in, where it's where it stored, how it stored and when it's destroyed, and then, more than to the degree that there's another smart contract that takes some of that data and sends it up stream, how granular that data can be. Very can be trusted to a very high degree, because we've got, you know, if you will, signatures of all the different components that make up the brake system, that makes up the engine system, and Geep has very tightly controlled those supply chains. I would look at a company boor set up io leveraging. We need that trusted hardware right. You heard about the chips coming back from from China with freeloaders on there that are doing not quite true what compromising our future. FIVEC networked, among other things, you can make chips that that have entirety, that can, that can even though they're in the field. We can trust they are what they are. We know if they've been compromised so that we can trust their data and we can start to get to six signal levels of data accuracy and reporting at that level. And we need that because if my car goes into the fast line and I'm using crappy break parts or I did my own break job at my in my own garage, I don't get to go a hundred ten miles an hour because my car doesn't stop according to the specifications, I could take out twenty cars, not just my own self. So the when you get into that point again, it's what happens is now what we did with the Internet was created very huge pipes of data with very little boundaries within them until you get into very proprietary sorts of networks. What we're seeking to do now is develop trusted conduits for a particular purpose with much more efficient because we know about it. You know uses of data, much more efficient interactions because those interactions are trusted and consistent and verifiable. So the will be doing a lot of the same thing, but the where it's not a question of faith, it's a question of trust, and that trust that context, very, very specific context. That's a user from the user experienced perspective. Am I going? I am I going to be able to verify my own trusted context, like, yes, I trust my car until the moment I find out that my car has been exploited or, you know, somebody came up and literally like, because my car is mobile and I don't have control of her physical access to any degree by any any any sort of typical security standard. I have no physical control, physical security on my car. That's kind of my definition of the feet of the way the vehicles. Somebody could easily tamper and and you know, just do this regularly with with vehicles. They could even remotely do it without even touching the vehicle. Yeah, you know things I want to do. Make my car look like an ambulance. So when new system, when we're all autonomous vehicles and cars are aware of each other, my my hacker bring to the art. How do I make all the cards around me. Think I'm an ambulance and get out of there. You go right. So thinking of that sort of thing helps learn how to how can that be prevented? You know, hopefully I'm ahead of myself. Phone. That is an incredible by the way, that it's I had no idea what...

...you when you first said angels like, okay, sere goes to this. That is actually an incredible example of what I'm talking about because, yes, if you have an automated car system and you have automated yeah, like whatever the being able to spoof the type of vehicle you are is is also is a potentially transparent and and personally beneficial way of manipulating your life. Is A prostical life. That's almost that's what Tou say it in terms of you need that contextual trust. You gotta have that rob all the way down. No it maybe I'm the fire chief, right, and I'm driving my station wagon, right, you know, I'm the fire chief and I get a call, I need to be able to flip something that assert I am who I am, even though I'm driving my my family truckster. I need to be urgently so I need all of a sudden everybody else to recognize that not only my emergency vehicle, but I'm an emergency vehicle on away now. Right. Yeah, maybe they didn't know I was an emergency veable before, but now I am. So, I mean, but the context, I guess you're saying is is literally a kind of fingerprint of the vehicle itself. Of any one of those pieces changes, the whole kind of thing breaks down. Is that what I'm misunderstanding here? It's a hierarchy, right, so there's there's that. The brake system is one thing, right, and they, you know, I jeep. It's okay, this is my trusted suppliers. It's going to be trusted for certain things. Their conditioning system may not care, right, you know, we may not need as much trust in the air conditioning system to give them the AV lane as we do in the brake system or the engine performance. Right. We don't want it to be, you know, you know, two thousand miles around an oil change, going one hundred ten miles an hour. That could be bad. What I'm seeing then, is basically each part, each kind of like major critical part, is its own serialized number, essentially. And so if we were to build sort of like you could create assentially a mercle route all the way up to the top that represent your vehicle, and if any one of those parts changes, there has to be a registered change in the in the global system. As well as that, your mercle root might you know you are changing this part and if that doesn't match, then you have basically an invalid vehicle. So let's say you wanted to say this supervation comes in. It able to sort of like have the the lights go and to people would be aware that, hey, I'm an ambulance mode right now. But if your car doesn't possess the ambulance mode, that breaks the whole security, the whole the whole validation check of your car. Your car itself loses its own validation and the rest of the cars around you go ay, that's an invalid vehicle, which of course, flags you would and potentially can bring it down without having to know anything that's inside of your vehicle. They don't have to know anything. NOPE, exactly. This is where I start talking. You know, when we've been bringing it back to kind of, you know, the bitcoin concept in the theory, in which I mean you guys talk a lot about. This, is where the what I kind of go microcurrencies, in these private currency start to exist, right they the the ticket to ride in the AB lane will be a result of my having a token that allows me to join, having all those other things that are that are true according to other trusted and increasingly private networks, jeeps break system as opposed to all break systems, you know, and how that works relative to the question is, you can go on this Ab Lane if your break system works. You're at you're get, you know, all these other things. You're true, you get a gold coin and now you can join the Ab Lane, you know, or now you can identify yourself as an emergency vehicle or do all these different things, and those will all be different types of of you know, of of you know, non human currencies, if you will, modern monetary cuarrencies that will be attributed to specific value. And a lot of these, a lot of the stuff is trying to build solutions based on the like the aggregate questions of what can I prove and then what can I build from all of those proofs? That right and and the communities to find their own standards right in the communities of communities that you know that define an an overall met you know, organization, and then how those communities are going to interact and when that's again where, I think, you know, insurance has got a great opportunity,...

...because not only are we, you know, the the financial backing for a lot of these things in place, but we also recut we need to create that economic communication, that transaction across regulators, roads, cars, other cars, you know, in trucks and commercial vehicles, and stop lights, and you know, to the real that we even have those anymore. And then make sure all that system has integrity right, because you don't want that to be hacked. If we can all of a sudden say, okay, every car on the highway, I want half of them to stop and half of them to Florida, right, that would be a problem. So this is this is really great for you know, I think the car analogy is fantastic. I think the real estate analogy you gave me when we firstly we're talking, was also fantastic. With flood insurance, what was our that's our first working by the way. Blood what other so we do have to kind of wrap up here soon, but I feel, like a lot of other industries, you're not mentioning here, and one of them we met at a industry specifically around the financial sector and dealing with like Golden Sax was presenting there and NASTAC was there. So I'm curious what you see this technology doing for, say, trading organizations or something like that. Like is there? Is Their potential use there? What is the what other use cases are there other than these these real estate and car use cases which people in the general audience would understand. But like what kind of like more enterprise e use cases are we seeing here? Well, enterprise, you can get it comes back to the community in the standard and enterprise of its own community, so as they and they can define their own standard and nobody else has to. You know ingested, you make you know to worry about it to the degree that it looks like something else. You know, for example, I man a opportunities are will become pug and play that. You know that sort of thing. If you, if you've got your you know the financial experience in the policy experience in a trusted data store according to a trusted format, and you get acquired by a larger organization, they already know everything. Maybe you know about you and you are now a reading the operations in a plug and plate way, and we've got carriers that are adopting the platform right now, the meanting that they didn't do get a year ago, having just bought a very large organization, and it's not part in plays years worth of unraveling data. So that's more application. We talked about it to the to the broader financial solution. That's where we starts to get instant in again, because if that financial day we're doing insurance in its tregulatory reporting data, but financial reporting data is the next thing. You know that we're looking at making sure that carriers have the the right liquidity, the right level of assets. You know, especially as those things change like the market, all the sudden, if the market goes down twenty percent, you know the bag of money you had to ensure the hundred houses is as big a bag as it was yesterday. Can you still ensure a hunter house houses? What do you what do you do? How much the way you need to have so and as that sort of thing becomes becomes more trusted both by regulators as a framework of trusted system to get that information a more real time basis. That allows the whole thing to move forward so as houses and cars and most physical assets are brought into a digital streamline world, the opportunity for not only digital currency, you know, be a Bitcoin or be at a digitized dollar, has much more opportunity to be directly trot tied in real time to physical assets and make real change in the world. So, and it works both ways. One of the things I talk about that's kind of most exciting one, you know, is because of the challenge we have in the Internet is the inn the lack of trust that we seem to have across our population among the world. Insurance underwrites everything right, every every product that's sold on the market, every plane that gets in the air, every every publisher of content. One of the things we can do very quickly, when people talk about proof of Insurance, is one of the first use cases that insurance can provide. You know, when you get pulled...

...over, you have to show your show the top, your your proof of insurance that you're driving. Increasingly they know that automatically by a different systems, but another smaller context, having a proof of insurance is extremely valuable. For example, as you're just going through facebook and you're seeing news stories, maybe one day there's a one pixel box around a news story and you hover over that and it says, Oh, the publisher of this content is verified. To have to be to have a million dollar or greater media publishing coverage. We're going a bit rated insure. Okay, extings will be that. As preying. I see more of those green boxes that store those things seem to be more and more trusted, are more and more legitimate. The things that are not. everying boxes are less and less. Look less and less legitimate, more and more like Click Bait. One day I can have a little switch that says I only wants you stuff from from verified publishers. That could be me and my basement as a you know, as a technology blogger. You know that. I told State Farm Ay, I'm you know, I'm going to give you a call at apple decides to see me. Or The New York Times or The Wall Street Journal or the National Enquirer, a legitimate publisher. You can choose how much you want to trust the information, but they are in fact legitimate business as opposed to, you know, the thing my dad sends me. Yeah, I think that's it's a great way to wrap this up. That that brings true to what you said earlier, like insurance kind of supports everything and and finding better ways in which it works more efficiently is certainly at a giant endeavor, but one worth doing. Where can where can people go to learn more about both kind of the business side of this as well as the technical underpendings account works? So the first place going to the AIS onlinecom as a as onlinecom, slash open idel downloads to basic information. there. I mentioned the open ideal industry test drive. There's a link from that page there. That's a great place to go to get a good overview in a video, since by doing that you'll sign up to email, will give you a give you more news letters and more information. We are not for profit by your organization, and we seek to follow the hyper ledger in the Linux model, and that means our marketing budget is not huge. So all our content would be is is being generated rapidly and if you have anything of the question that encourage you to reach out to me directly. My email is Truman e at Ais Onlinecom or again, if you sign up for that, there's a room for comments. If you have any questions, please let me know. The happy to reach out to you. All the presentations are available. Everything we're doing is is as open source as we can make it. We are seeking to be an open source project as part of the hype, as ideas the next communities, as are as our organization matures. So there's also a wiki open idel dot at last see and dot met as I got that right. Basically confluence Wiki with open idea at the beginning and then you can see the membership agreements and information on our floodworking group, things like that. Right now the membership is is organizations, but it's very much wide open and we're looking to once, once becomes open source, then we'll be able to have a much broader individual, live community and we're looking for ways to do bad and looking for lots of ways to just state. So looking forward to talking with a lot more people and a lot more industries and perspectives. Yeah, I'm looking forward to seeing a lot more come out of this. I do think that what you're building is really great. I'm glad you're open sourcing it. I hope the patents won't be kind of a barrier for a lot of people for an option. Not at all know they will be there. There it's they won't be again. That's why we are joint pads with with the ask from open and we had vision donating those pads to the open source organization as IMA chips. That's fantastic. So yeah,...

I'm looking forward to seeing some of this in real world, especially as this will power the auto you know them, the automated cars that we're dreaming about and all that kind of stuff. So I'm looking forward to seeing this all kind of like manifest in reality and I'm glad to hear that you're getting a so much traction. It's really be great talking again, Truman. Thanks for having me. Appreciate you to.

In-Stream Audio Search

NEW

Search across all episodes within this podcast

Episodes (118)