Idea: Using Decentralized Courts to Develop Insurance Products Requiring Human Judgement

Hello Etherisc Community!

I’m a long time Chainlink supporter, recently discovered Etherisc through their channels, and have since fallen in love with the idea of a Decentralized Insurance Protocol. Having spent the past couple weeks contemplating potential use cases, I’ve run up against a problem (or perhaps a better word to use would be “limit” instead of "problem), and devised a potential solution.


Most insurance products require real world data outside the blockchain in order to settle claims. Some of this real data can be fed into the blockchain thanks to oracle networks like Chainlink. Weather data oracles, flight information oracles, and so on can be built and deployed to make decentralized insurance products such as lack-of-rainfall insurance, flight delay insurance, and so on possible. However, many insurance products require data reliant on some form of human judgement to settle claims. Such instances cannot simply be settled using an oracle solution, because Oracle solutions rely purely on API-pulled data and can’t apply the complexities of human judgement.

A great example of this can be found with car insurance. When making a claim to your car insurance company (whether it be Geico or Progressive or whomever), the procedure often involves determining things like fault (who caused the accident), magnitude of damage (where is the damage, how bad is it, and what are the cost of repairs), and dispute (handling disagreements).

One of the limits of an insurance protocol such as Etherisc is that determining fault, damage, and handling disputes can’t be resolved without human actors applying their individual judgements. Granted, perhaps in the future, artificially intelligent programs will be able to make such calls. However, that’s at least some time away. For now, some sort of subjective or intersubjective human adjudication process is necessary to resolve issues of fault, magnitude, and dispute for many types of insurance products. And the present inability for the Etherisc protocol to make decisions based on human judgement limits the types of insurance products that can be offered.


Kleros is a dApp which bills itself as “an open source online dispute resolution protocol which uses blockchain and crowdsourcing to fairly adjudicate disputes.” The way it works is that, when an individual or business files a dispute, Kleros selects a panel of jurors and sends back a decision to settle the dispute. Anyone can participate and make money as a juror, in effect creating a decentralized digital court system.

Imagine if Etherisc could leverage Kleros’s courts in order to offer insurance products for situations that require human judgement. As an example, let’s consider car insurance. Using the Etherisc revenue model for protocol users, a developer could hypothetically create a car insurance product. The problem is, the car insurance product would have no way of knowing if you got in an accident, whose fault it was, and how much damage was done. But what if, using Kleros courts, an insurance policy holder can file a claim? The claim would automatically be sent to a Kleros court, which would determine the fault and extent of damage and feed the result back to Etherisc. Etherisc would then pay out settled amount using funds from the risk pool.

And voila! Using Kleros courts, Etherisc would be able to factor in human judgement to provide insurance products for which such judgements are necessary to settle claims.


Maybe we could set up our own system, and this would be another use of the Dip token? Or upon partnership, Dip would be used as a way to pay for the Kleros. I dont know.

Auto insurance is a huge market but it has a lot of steps that aren’t data driven.
I’d imagine that even if you make it easy for people to buy an insurance plan, pick their coverage rates, and even upload accident documentation like police reports or pictures, you’d need people to go over the documents to determine fault and damage payouts.

I have a friend who is a lawyer for Geiko. She is one of a fleet of lawyers they employ and all they’re there for is to ensure that they pay out as little as possible.

I also would imagine that auto insurance providers would need specific licenses to operate in each state / country. Like Geiko might need an office or business license in each state. This might provide scaling issues.

If the tech works, it might be easier to sell or subcontract it to existing auto insurance providers than to become the provider itself.

What drew me to Etherisc and Dip is their products seem to be data driven with automatic payouts, which is why it’s built off of chainlink I’d imagine. The data is Oracle generated and verified, and then if requirements are met, payout.

Maybe instead of auto insurance it could be “smart car insurance” and only be available for cars that can automatically provide car velocity and photographic data. It’d cut out some moving parts? Idk even then there’s still the police reports, and personally if my car automatically contacted my car insurace I’d be very unhappy because I’d be afraid the slightest ding would result in a headache and an increase of my insurance rates.

I have a feeling Esthetic should prioritize insurances that are more data driven and have less moving parts. As Dippies increase in value, the Dip held by the Etherisc foundation will increase too, and they’ll be in better position to explore and attack those auto insurance scaling issues. Plus, smart car tech might be a little more advanced, and maybe a really AI that can parse through police reports to find fault, or that can replicate the stingy lawyers, will fall into their laps.

(Disclaimer: I dont own much Dip yet soooo here’s hoping the rocket to the moon is still a bit aways because I haven’t packed my bags!)

Agreed. There is absolutely no reason to dip our toes into the murky waters of traditional insurance offerings where ambiguity exists and subjectivity reigns free. There’s a ton of stuff we can cover that only needs deterministic inputs to settle, like billions of dollars worth of market share there. If there’s need for human intermediary, then the use-case is simply not viable for DLT / smart contracts at this time, however of course that may change as we mature and evolve.

1 Like