How Much Better Will Life Be When Our Contracts Are Smarter Than Us?
Written by David Vandervort
Society is changing. We all know that, so let’s be more specific. Society is changing because technology is getting smarter. Search engines are getting smarter. Phones are getting smarter. Cars are getting smarter. Now, even contracts are getting smarter. This, by itself is not news. What is worth considering is that even as smart contracts change the way society does business, society will make demands that change smart contract technology, too.
Eventually, smart contracts will become so smart, they will be able to run businesses themselves, with little or no human input. As long as humans still get to collect the profits, I’m sure that’s not a problem. In the near term, smart contracts will soon be like cars, so useful to people that they will be indispensable. Like cars, the demand from huge numbers of customers for convenience and safety, will foster amazing developments. Corporate legal departments will also like the predictability of smart contracts (though as the intelligence of the system increases, predictability — from the human point of view — may decrease. That could be a problem). I can see politicians, too, requiring smart contracts in some areas as a form of consumer protection.
In fact, starting with Nick Szabo, who proposed smart contracts in 1994, cars are the standard example of what good a smart contract can do. From the seller’s point of view the smart contract is excellent because it will automatically send a signal to the car’s computer locking it up if a payment is late. But buyers will also have advantages. The smart contract will never lose a payment or apply it to the wrong account. More importantly, when the payments are done the smart contract can automatically assign the title (conveniently held on a blockchain) to the new owner. No muss, no fuss, no chance for bureaucratic snafus and it will be much quicker than waiting for the bank and the DMV and the insurance company to all do their paperwork. Instead of waiting 6 to 12 weeks after the last payment is sent (Electronically. Who sends checks through the mail anymore?)
You can have an oral contract, a written contract or even an implied contract (Although I should mention the all important caveat that I Am Not a Lawyer. I took a couple classes, which is not the same thing at all!). Now you can have an automated contract, a smart contract with terms encapsulated in computer code, making it (in theory) unshakeable. The recent DAO debacle showed that a code-based contract has the weaknesses of anything that uses software. In the future, smart contract security will be a tremendously important field. In any case, any agreement where there is money involved is liable to end up in court. In court, the security of a smart contract will be an important point to consider. Having them operate in conjunction with a blockchain goes a long way to providing a traceable trail of evidence. That may cut both ways, depending on what the evidence shows.
So we can see already that making sure smart contracts execute as they are intended is one of the forces that will shape how they are accepted by both society and courts of law. But this is just the beginning. As security improves, businesses and their customers will begin to demand smart contract features that matter to them. One of the most important features will be the need for smart contracts to respond to more kinds of inputs. Let’s change the car payment example to a mortgage and see what other terms and input sources might come into play. For starters, my mortgage requires that I keep the house insured at a certain level. In a future world of smart contracts, that means the smart contract I have with the insurance company has to be able to talk to the smart contract I have with the bank. (An engineering task that, all by itself, seems to ensure long term employability at very high rates for people like me. Did I mention that I have some knowledge in software security too?).
Once a communication protocol is developed, getting two different smart contracts to talk to each other is comparatively easy. But what if I need to file a claim against that insurance policy because of, for example, flooding? The insurance contract will require verification which might include pictures, a written description of the damage and a contractor’s estimate. In the current world, all these things require humans to process them. A claims adjuster, a contractor, a person holding a phone, all send their data into an insurance office where people input them into computers. Hopefully, they do all this in a timely manner with no mistakes.
Yeah, right. In a world where you click a link on your digital wallet and it sends a digital coin of some kind to a smart contract that automatically does the bookkeeping, it doesn’t make sense that people will accept having something as important as claims processing at the mercy of fallible humans. They are going to want the pictures taken by a drone, that directly uploads them to a system that interprets the pictures in a way that is meaningful to a smart contract (this is one variation on the concept of an “oracle” that provides inputs to the smart contract). There will also have to be computer interpretation of your spoken or written testimony of events and the contractor’s electronic estimate. It may even gather weather records and reports of damage to other, nearby homes, to verify your claim.
So now, we have described a smart contract system (yes, I’ve conflated the contract itself with the oracles and other systems that feed it data. I think this is justified because those systems will also be questioned in court and will have to stick to the same kinds of rigorous controls) that interprets speech, written text and pictures and probably decides for itself if it needs to do statistical analysis on weather reports, or if you fit the profile for fraudulent claims. This is an artificial intelligence. It may not be full blown sentient software (artificial general intelligence for the more pedantically inclined) but it’s getting there.
This is the world we are making. One in which the normal desire for routine transactions to be routinized, pushes us inexorably toward building and using artificial intelligence. While this may seem like coming at true human-level artificial intelligence through the back door, people looking back a hundred years from now will think it was all inevitable.