Notes from AM 19/ Preparing For the Robo-Revolution: Arbitrating Smart Contract Disputes

By Shannon Collins

Following up on last week’s CPR Speaks post on the CPR Institute’s 2019 Annual Meeting—see “The ‘Risky Business’ of NAFTA, BITs, ITAs and Global Trade” (March 21)–blockchain, artificial intelligence and “smart” contracts were the focal points of CPR AM19’s Session 7, “Preparing for the Robo-Revolution: Arbitrating Smart Contract Disputes.”

On Friday, March 1, panelists Daniel E. Gonzalez, a Miami partner in Hogan Lovells; Andrew James Lom, a partner in the New York office of Norton Rose Fulbright; Lee Schneider, a New York-based general counsel at block.one, a Grand Caymans company that publishes protocols enabling blockchain transactions; and moderator David L. Earnest, Shearman & Sterling partner in Washington alerted attendees to the increasing use and necessity of technology in arbitration, as well as the legal professional overall.

The session’s topics included the use of technology in arbitration today, an overview of blockchain and an introduction to Ricardian contracts.

Daniel Gonzalez started the panel off by discussing the arbitration technology. He noted that general counsels are pushing for more use of technology because they strongly believe it will increase efficiency. Software can be used to help appoint arbitrators, conduct document review, aid in document production, manage costs and assist with research.

Picking arbitrators with artificial technology poses several issues, Gonzalez explained. While it can assess a likely outcome of a case using algorithms, these algorithms rely on correlation on a purely mathematical level, while legal disputes are based on causation.

These algorithms can also be biased, which seems counterintuitive, but they are programmed by people. People are inherently biased and the data sets that they use as a base for the algorithm may reflect those biases, which in turn can affect the outcomes.

Lee Schneider echoed this concern, noting that people working in artificial intelligence are grappling with getting the right data set because if the data is skewed, then it can result in biases. Programmers must be careful and judicious about the elements of the data sets.

The discussion noted that even if they are carefully coded, algorithms can still fail to see the human factor in disputes that an arbitrator may see. An arbitrator could be swayed on a deeply personal level by an argument and rule in a way that is inconsistent with their history. The algorithms cannot account for this. In arbitration, this means there will be accurate data available for the arbitrator to consider when ruling on a case.

Gonzalez brought up questions of “robo-arbitrators,” which are not yet replacing arbitrators, but that Gonzalez says will play an increasing role in the process and is something we should all embrace. Not many institutions have rules specifically requiring “natural persons,” Gonzalez informed the room, which brings interesting determinations of whether Due Process is satisfied with the use of a robot as arbitrator. Another concern with robo-arbitrators is their inability to provide reason for a decision. They purely provide an outcome.

Gonzalez also discussed technology specifically in the context of construction disputes. Drone technology collects data that can be useful for monitoring tagging of each piece of equipment or material as well as tagging workers. By using a tagging system, the supply chain and the project’s progress can be tracked entirely from inception to construction, thereby ensuring transparency and efficiency.

The tagging systems are an example of the kind of data that can benefit by storage on a blockchain. As explained by Andrew Lom, the blockchain is a database where data lives in blocks. The chain links these blocks together through a crypto-relationship. These blocks are sealed once a new block is added to the chain and it is incredibly difficult to alter a block once the new block is added, because additions are conspicuous and documented.

Lom emphasized that the blockchain is not automated or perpetual. You need people to pay attention to the database, he said. This doesn’t mean that it would always need to be monitored, but without attention, then it could fall apart, because alterations in blocks of data could go unnoticed.

The blockchain serves as a useful tool on which smart contracts can be stored. A smart contract is a self-executing contract. Lom equates it to when a bank is authorized to automatically pay the minimum on a credit card. It is done automatically without the debtor’s need for immediate implementation.

Smart contracts are written in code like a computer program. Ricardian smart contracts are contracts that are written partially in computer code and partially in human-readable language. (A smart contract usually describes the terms and contains codes that execute electronically under the contract’s terms, while a Ricardian contract is similarly linked to code, but is usually fully realized whether executed or not.) According to Lee Schneider, this introduces the human element into drafting smart contracts. Humans serve as a check for the execution of the contract. They can discover when something goes wrong and resolve the issue.

Andrew Lom posed the problem of bugs in the code. If the code is flawed, could that constitute a drafting error? A person would have to catch it and correct it. Could that correction potentially be modifying the substance of the contract?

These are questions and problems that will be answered only in time. As for now, Schneider noted, we are all participating in a digital world.

* * *

The author is a CPR Institute Spring 2019 intern.