Not-So-Smart Blockchain Contracts and Artificial Responsibility


The first high-profile decentralized autonomous organization formed in 2016. Called “TheDAO,” it used smart contracts on a bitcoin-style blockchain to allow strangers to come together online to vote on and invest in venture cap-ital proposals. Newspapers raved about the $160 million it quickly raised, even though it purported to have no central human authority, including no managers, executives, or board of directors.

Technologists have grand plans for smart contracts and autonomous organizations. Rather than staying at traditional hotels with elaborate human staff, we may pay for hotel rooms using bitcoin (or another cryptocurrency) which will automatically unlock the room door. If the toilet breaks, the room itself will contract with a plumber to fix it. Similarly, a smart contract may allow us to hire a self-driving car. The car will not only drive passengers around but arrange for its own routine maintenance.

TheDAO itself, however, is now a cautionary tale. A bug in its smart con-tract code was exploited to drain more than $50 million in value. Some purists denounced efforts to mitigate the problem, arguing that the alleged hacker simply withdrew money in accordance with the organization’s agreed-upon contractual terms in the form of computer code. Since the “code is the contract” in their minds, the alleged hacker did nothing wrong.

I defend two related claims. First, contra the purists, I argue that the code does not reflect the entirety of the parties’ agreement, and so the “code is the contract” slogan does not resolve whether TheDAO exploitation should have been mitigated. I take no position on whether mitigation was appropriate except to say that the matter depends on many considerations aside from smart con-tract code itself.

Second, I point to a broader danger lurking in the code-is-the-contract view. TheDAO had tremendous “artificial responsibility” in that we gave it considerable control that couldn’t be easily revoked or reined in. Not-so-smart con-tracts in the future may prove even more dangerous: hotel guests might be locked out of their rooms, and self-driving cars might drive off bridges. I argue that unadulterated commitment to the code-is-the-contract slogan increases artifi-cial responsibility and its associated risks.


Stanford University Stanford, California
  • 21 Stan. Tech. L. Rev. 198 (2018)
Related Organization(s):