People have long used technology to help them make and carry out their contracts. Even an old-fashioned vending machine offers goods to the public, understands an order for particular goods, recognizes execution by the buyer through the deposit of the appropriate payment, and delivers the goods.
However, we do not usually think of the vending machine as smart or its implied contract of sale to be a smart contract. Apparently we need to see a computer to attach that label. A smart contract these days is generally considered a contract the performance of whose terms is judged by technological means, and often whose execution or enforcement is carried out through such means, without human intervention.
Some people see the blockchain as either required for a contract to be called smart, or as the ultimate expression of the smart contract. Here too, it is the automation of the whole system, its subjection to unalterable system rules, that attracts the interest. Let’s get the contracting parties, with their competing interests and subjective evaluations, out of the picture. Let’s push aside those pesky lawyers and courts with their unreliable judgment calls. Let pure technology do its job.
The chart at the head of this article (developed by PWC) shows an imagined progression of smart contracts, through smart organizations to smart societies – with the blockchain assuming the role of intermediaries and ultimately of government. We are not there yet, and some fundamental questions of relation of citizen to citizen need to be answered before we get there. Not everyone has an equal ability to understand or monitor the rights involved in these activities.
However, we can avoid the speculative range today and stick to contracts. Our legal system gives a lot of leeway to freedom of contract, so if parties want to set up their relationships in this way, they probably can. In many cases this will work well, and save time and effort (once the programming has been done.)
There are a few legal and practical issues that need to be taken into account before one rushes headlong into the smart contracting era, however. Lawrence Lessig and Joel Reidenberg pointed out nearly 20 years ago that “code is law” – the choices about how computers will operate have consequences for the exercise of legal rights, making some easier and some harder to invoke. Smart contracts are a conscious exercise of this principle, coding to enforce the law set by the contract. To what extent do they or should they be allowed to change the law as they do it?
Context – consent
Looking first at the context: Is there real consent from both parties to subject their relationship to the judgement (and execution) of the computer? Are these contracts being set up by browsewrap methods, i.e. “agreements” where a non-tech-savvy customer is bound by terms of a contract merely by making an order online, without actually seeing the terms or actively consenting to them? Courts have often been reluctant to enforce such contracts – U.S. law at present is particularly unfocussed – and if the customer is seen to be prejudiced by the ‘smart’ elements of the contract, courts may give remedies even in the face of automated execution.
Some parties have special legislative protection that smart contracts should not be allowed to circumvent. Ignition interruption software may be a non-violent way of controlling automobiles as collateral, but statutes govern how rights in collateral may be exercised. Enforcement by computer program may be ill equipped to estimate the practical consequences of shutting down a car, too. Why is the car to be used? A routine trip or a family emergency? Other forms of automatic enforcement may produce greater costs on the party on the receiving end than mere contract would normally be allowed to do.
Contract law rules
Aside from legislative protection, contract law contains a number of other protections for all parties or for vulnerable parties, including laws on mistake and misunderstanding, and the newish (for the common law) duty of good faith. May more powerful contracting parties simply remove such protections by being “smarter” than the law? This is a similar concern to the one often expressed about technical protection measures for copyright, namely that they can prevent the exercise of rights long given by the law (or the functioning of defences, for those who prefer that characterization of doctrines such as fair dealing or educational use.)
A digression: Canadian copyright law, modeled on US precedent, prohibits the circumvention of technical measures used to enforce copyright, regardless of the legal rights being asserted against it. Presumably there is no such barrier to circumventing the automated enforcement of smart contracts. Doing so may breach the contract itself, but the circumventing party would have the opportunity to argue that there was a legal right that justified the action.
Practice and principle
The ideal smart contract is intended to be a closed system, where the rules and their operation are all prescribed in terms of technology. But sometimes that does not work. The main example from this year is the experience of the “Decentralized Autonomous Organization” (DAO) in running stock trading through a blockchain-enabled value exchange known as Ethereum. According to the rules, investors could as a collectivity make investments through Ethereum, and all transactions would be transparent and irreversible. No brokers, agents or lawyers were needed. Everything was very smart.
Except that somebody figured out a way to extract about $50 million from the system, playing by the rules. While the rules had not been consciously designed to permit this, the person who figured out how to do it was entitled under the rules to do it, though no value had been provided in the normal course. The organizers of the DAO had almost an existential dilemma: change the rules to get the money back, or maintain the principles of the closed smart system.
The organizers did a quick survey of members, and when a majority of the small proportion who responded favoured a fix, the system was changed to restore the money to the system. Some people have claimed that this was a “victory” for technology – the legal system was not needed to restore order. However, it may have been a costly one for the smartness of the system – it needed outside help, a programmer ex machina.
Self-help by parties to contracts has always been available – but the need to prevent strong parties from helping themselves led over the centuries to a system where neutral agents – courts, ultimately – would decide who could do what, and the state would enforce the decision.
A financial journalist put it this way:
… the broader point is that every contract, every human agreement, is embedded in a much wider set of rules and social practices. In the regular world, this is easy to understand, because contract law is just a subset of the much broader rule of law. If we sign a contract saying I can murder you, I can’t murder you, because no one in the real world is silly enough to think that a contract can fully determine the rights of everyone involved in the contract. In the blockchain world, they’re still figuring that out.
Another article drew similar lessons:
The law came to the conclusion long ago that, sometimes, it needs flexibility to do what’s right; contracts don’t always automatically work like they should, whether on paper or on the blockchain.
Don’t kid yourself, techno-utopians: as smart as your contracts may be, you still need legal principles to save them from your dumb mistakes.
One is reminded again of the differing mindsets of engineers and lawyers – except that a number of lawyers seem to be on the engineers’ side of this debate. Perhaps, as a colleague suggested, the young dogs need to learn some old tricks.
Technology can certainly help administer contracts and make enforcement easier. Nothing in the nature of technology makes these applications fair between the parties. Especially in consumer contracts, one party will normally be more sophisticated about the technology than the other – and will probably have developed the technology to suit its purposes.
The law has wrestled with the fairness of contracts of adhesion and has applied notions of unconscionability to their enforcement. Smart contracts of the same nature may need recourse to the same notions – but if they are self-enforcing, how will this happen?
In short, even when smart contracts work as intended – which is not guaranteed, as we have seen – should they? Who will even the technological playing field?