Algorithmic Learning Foundations for Common Law

TL;DR

This paper models a common law legal system as a learning algorithm, analyzing how court costs and incentives affect the system's ability to learn accurately over time.

Contribution

It introduces a novel framework viewing legal proceedings as learning algorithms, highlighting how costs and incentives influence legal accuracy and efficiency.

Findings

High costs lead to under-claiming and learning failure.

Incentives can improve case filing rates and system accuracy.

Learning improves when cases are brought to court more consistently.

Abstract

This paper looks at a common law legal system as a learning algorithm, models specific features of legal proceedings, and asks whether this system learns efficiently. A particular feature of our model is explicitly viewing various aspects of court proceedings as learning algorithms. This viewpoint enables directly pointing out that when the costs of going to court are not commensurate with the benefits of going to court, there is a failure of learning and inaccurate outcomes will persist in cases that settle. Specifically, cases are brought to court at an insufficient rate. On the other hand, when individuals can be compelled or incentivized to bring their cases to court, the system can learn and inaccuracy vanishes over time.