Insurance on the blockchain

Transcript

[Music] hi everyone um I'm here today to talk to you about decentralized insurance and where we are with that and what how how the next steps could look like quick one about who I am and who I'm working with so I'm alas I'm co-founder of B9 lab we do development training and research around decentralized applications and blockchain mostly ethereum at the moment

because it works the best uh but yeah other things as well I want to talk uh to you about insurance insurance and insurance is often has this connotation of being relatively dull and being old and being boring but actually it is a slow moving sector but it's actually could be quite exciting in the future and it is quite exciting in the sense that it underlies almost everything uh almost everything we interact with on a daily basis at least around these parts is insured or is is interacting with some kind of insurance schemes it ties into different marketplaces is so it really is quite uh quite

fundamental and the the options or you know the opportunities with decentralized decentralized systems could be quite you know could be quite disruptive for the sector now the insurance industry has not been disrupted in a long time hasn't really changed in a long time a few things have changed over time but it still looks very similar to to the way it looked 200 years ago new technologies have obviously changed things um but the basic model is still the same now the blockchain and and decentralized Technologies could present an

opportunity here to make it more accessible also to those who are underinsured or don't have access to the insurance schemes they need uh it could make Insurance more local and revive old models like mutuals and I'll get back to that and it could make Insurance more V varied and more custom the interesting thing here is we've got a we've got an intermediary or you know large intermediaries and obviously the technology is a an essentially disintermediating Force um and there are three core elements here that make it interesting and that creates a situation

where we could potentially build very interesting uh interesting things with this so we've got low transaction cost when we operating on a public or uh Consortium blockchain we've got low transaction risk so the risk that my transaction is being reversed or being interfered with by any other party is very low and we've got trustless computations so we can Implement relatively complex Financial models um at you know in formal smart contracts on a blockchain system and this is interesting in a few spaces and I want to define the problem space a little

bit and where this could be applied so we've got claim processing which is you know we I buy a policy and the policy sets out certain rules under which I get my coverage um and claim processing is one of the hardest things to crack in this uh in you know under this model so in in many cases I have to send paper documents or I have to you know go somewhere I have to send data that is not easily uh not necessarily there in digital formats or it is down to um interpretation so so a claim processor has to sit down and has to apply certain

rules and has to decide and so this is for for most insurances and we've also got index based insurances things like crop insurance for example based on weather in indices and they rely on data so we have the problem of being able to or needing to be able to bring data onto the blockchain to make decisions there so it splits into two main kind of problem spaces one of them is rule-based claim processing so this is bringing data onto the blockchain and then making decisions based on this data which also brings in you know how do you bring data on the uh how do you

bring data in and how do you make sure that it's the right data and not being fiddled with the second General problem space is other types of insurances that need that need some kind of voting or that usually have an have a person and deciding on whether reclaim is being paid out or not so this goes into building incentive models that that enable groups to decide which claims should be paid out and which claims shouldn't be paid out so this is for traditional Insurance models where where uh one participant or multiple

participants take out an insurance and there is a detached pool of investment taking over or taking care of the coverage the other model of insurance that is really really interesting in for this space is mutuals mutuals have been around for a long time they started out as fire insurance it's a model where group of people gets together and insures each other this could be by regularly paying into a pot and paying out when one of the participants has a has a claim or has a claim case this has already been done with uh with car insurance kind of a social car insurance with a group of

people and again the interesting part here is bringing in low transaction cost and low transaction risk so you can connect people very efficiently the second is that it is inherently peer-to-peer Mutual insurances are inherently peer-to-peer so you know the the blockchain technology is perfect you know is a perfect candidate to solve that problem and the the infrastructure in itself obviously when we build something like this is is trustless and on top of that we can then build kind of custom trust models so for example we don't have to trust the infrastructure but we can build a system where um we Define which level of trust we want to have to other

people and lastly the M Mutual model is a very it's very much a proven model it's been around for a long time and to this day is executed very successfully now where this is generally the direction this is generally moving in and where we think this will you know where this will culminate is kind of a Marketplace creation so through the lowering of the risk and the cost you're enabling the breakup of existing monolithic structures so at the moment all of these things underwriting liquidity modeling management processing is all in one place which is within one large insurance

company and this is really because it's it's very very difficult to get access to the the liquidity that is that is necessary it's a large numbers game so you need to spread your risk and it is uh very difficult to connect different players different parties in the market in a way that is efficient enough to execute this so a decentralized model could be really interesting uh in in the way that that is outlined outlined on the right side there we could have through you know through trustless computation through smart contracts we can bind different participants in the market to certain rules and we can you know we could see

the creation of an insurance Market where you you can create a policy yourself you get access to risk analysts risk modeling you get access to liquidity pools TOs and all of that is tied together automatically now there are a few things that we've been looking into so theoretical research and a bit of experimentation because a lot of this is just really down to trying out how these model work these models work and uh the game theoretical elements there are a few areas where it's really worth looking into and the first three are mainly around weather uh so we've got crop insurance bad weather insurance Insurance energy markets with

weather and things like a double spend micro insurance now I want to quickly go into the proof of concept that we've built so far which is which we'll publish it's still in testing we'll publish next week which is a a mutual it's a simulation of mutual contract it's a betting game you have a limited registration period and you pay an initial premium you join the you join this this mutual and then you have to pay in every interval so for example you have to pay five ether every week and once you stop paying you get your payout so you know it's the equivalent of dying

in a life insurance model the incentive is that if you drop out first you get back less than you've invested the longer you stay in and the more other people drop out the more return you will get and the last person will receive about three times their original original investment now this is to simulate interest uh interest accumulation um the first the first learnings we've had from this is machines have a huge Advantage with this so if you have a script that executes the transactions regularly that is much more efficient than than a human which might seem obvious but you know we are recording all of these and also if you run out of money you are basically

dead and so liquidity always wins at this game the next step is we're going to publish this and I want to give you a demo but I don't think I'll have enough time I'll see we'll publish this and you'll be able to to to join and try it out it is initially a betting game but we're trying to solve Insurance research questions with this uh we we will monitor the behavior so this is interesting to see when do people engage and when do they drop out we are starting to engage with insurance companies and we've been speaking to a few insurance companies to improve models and to see where this could lead and ultimately we will iterate on this so this is the first research

question we will you know we will finish the paper around it and we will move on to different research questions now if you are interested in joining when we publish when we publish the contract and we publish the proof of concept go there leave your email address we'll add you to slack Channel and if you're interested in the topic in general go there join us and we will keep you informed and as involved as you would like to be now I had to update my I had to restart my computer so the test environment is not running if you bear with me for a second right

I I don't I will not have enough time to to run the demo unfortunately but if you're interested in seeing the demo uh come speak to me afterwards follow us on follow us on Twitter and email us if you're interested in any of this and yeah find me find me afterwards and I'll show you the the demo and I'll show you how you can how you can interact with it thank you very much and if there are any questions AB uh yep we have time for a couple questions um there's one thing in in insur that I'm not sure that um at least

native cryptocurrencies are are good for it's um so so comp insurance companies they basically they don't have the money to uh to cover all claims so they rely on facts a bit like Banks they relying on the fact that no old claims will be um yeah I mean will be processed at the same time um so isn't that weakness if you would have insurances uh in in crypto native crypto this is this is actually uh this is a very complicated problem so in if you've got a Marketplace and you have to you have to provide full coverage for all policies taken out it becomes very

very inefficient now theoretically right now you're taking a bet that your insurance will be able to pay you and if theoretically if a decentralized autonomous organization becomes big enough and has enough liquidity and has spread the risk far enough it's a it it could approximate to a similar bet so you could go you know I'm I'm betting that this is that it's going to be all right and they're going to be able to pay me it is a very complex problem and and I think you know it's it's going to take some time to come up with the with the solution and also this technology doesn't solve all problems uh at all it's it's very well fit for certain subset and it might that

might expand but for the moment this is one of the problems that that doesn't really work also funds are tied up so you can't go and invest them and get a return So at the moment it's really interesting for for mutuals for Mutual schemes any other questions great thank you well thank you very much