22 November, 2016
In a thought-provoking paper presented at this year’s TPRC Research Conference on Communications, Information and Internet Policy, Internet scholar Kevin Werbach explains how the use of blockchain technology creates a new kind of trust architecture, which Werbach calls “trustless trust.”
Most existing trust relationships are based on a centralized “Leviathan” trust architecture where the state or another centralized authority (for example a central bank) acts as the trusted party of last resort. Other trust architectures are fully decentralized, operating on a peer-to-peer basis. Many human relationships are based on peer-to-peer trust, such as trusting a family member to feed your pet while you’re gone. Werbach argues that blockchain introduces a third kind of trust architecture, “trustless trust,” which is a system in which “it is possible to trust the outputs of a system without trusting any actor within it”. Yet Werbach warns that blockchain’s technical reliability should not be confused with trust. “Trust is a manifestation of goodwill of the one being trusted” – it involves a “positive expectation about the other party and a willingness to be vulnerable”, yet blockchain has neither of these attributes.
If a blockchain is reliable (i.e. the data stored in the ledger is immutable and produces accurate outputs), does it matter whether it can be trusted? Werbach suggests that where there is no other legal alternative, such as in situations where there are no existing property laws, courts or land registries to rely on, blockchain could act as a substitute for traditional legal enforcement mechanisms, thereby totally supplanting the very notion of trust. But this situation will be rare. In most cases, economic agents will prefer to have the ability to enforce legal rights in court, as opposed to relying solely on the output of the blockchain or any dispute resolution mechanisms embedded in the underlying code.
Werbach warns that smart contracts, no matter how well coded, will not be able to discern the subjective intent of the parties, or fill gaps in contractual or code language when new circumstances arise that were not anticipated in the code. The outputs may lead to results that are technically right but morally wrong. The DAO hack, for example, led to theft of investors’ “Ether” by exploiting a loophole in the code. To correct what was plainly an unacceptable result, the Ethereum Foundation implemented a “hard fork” to return stolen Ether to investors. In doing so, Ethereum broke the blockchain rule of immutability, but did so for a good reason. In that instance, reliability of output was sacrificed to higher principles.
Wright and de Filippi have predicted that blockchain will generate a body of rules, “Lex Cryptographia”, that is independent from state-created legal rules. Werbach admits that blockchain may supplant legal rules in societies without robust legal systems. Without a trustworthy government, “the availability of a financial infrastructure that doesn’t rely at its root on sovereign authority is a powerful opportunity”.
However, in countries with a longstanding respect for the rule of law, blockchain will not escape national laws any more than activities on the Internet escape national laws. As blockchain matures and becomes more widely adopted, it will attract attention from lawmakers, regulators and courts, who will apply established legal principles to blockchain transactions. The use of blockchain technology will not obviate the need for courts to untangle dumb contracts, discern the intent of the parties, and identify liable parties. Blockchain may actually help regulators apply existing law. Being transparent and immutable, blockchain could allow regulators (and regulated entities) to better detect anomalies. Blockchain may also reduce governments’ costs in maintaining registries for land, corporate or intellectual property registries.
Werbach’s paper is a helpful reminder that in countries with robust legal systems, blockchain is unlikely to change the relationship between law and code, or between code and trust.