LawyeR: Exploring the future of legal work with machine learning and comics

2H2K: LawyeR presentation.

“2H2K: LawyeR” is an in-progress project exploring the the future of electronic document discovery and machine learning on the practice of the legal profession. I’m pursuing that topic by prototyping an interactive machine learning interface for document discovery and writing and illustrating a comic telling the story of a sysadmin in a 2050 law firm. This work is a part of a collaborative project I’m pursuing with John Powers, 2H2K, which imagines life in the second half of the 21st century.

Discovery is the legal process of finding and handing over documents in response to a subpoena received in the course of a lawsuit. Currently, discovery is one of the most labor-intensive parts of the work of large law firms. It employs thousands of well-paid lawyers and paralegals. However, the nature of the work makes it especially amenable to recent advances in machine learning. Due to the secretive and competitive nature of the field, much of the work has gone unpublished. In this project, I’m working to create a prototype interactive machine learning system that would enable a lawyer or paralegal to do the work of discovery much more efficiently and effectively than is currently possible. Further, I’m trying to imagine the cultural consequences of the displacement of a large portion of well-paid highly-skilled legal labor in favor of automated systems. What happens when a large portion of the white collar jobs in large corporate legal firms are eliminated through automation? What does a law firm look like in that world? What does the law look like?

For this first stage of the project, I’ve been working with the Enron email dataset to develop a classifier that can detect emails that are relevant to a legal case on insider trading. In the course of developing that classifier, I had to read and label 1028 emails from and to Martin Cuilla, a trader on Enron’s Western Canada desk. While this process might seem dry and technical, in practice it threw me into the midst of the personal details of Cuilla’s life, ranging from his management of the Enron fantasy football league to the planning of his wedding to his heavy gambling to his problems with alcohol and contact with recruiters in the later period of Enron’s decline. This experience is already common to lawyers and paralegals who are immersed in previously personal documents in the course of doing discovery on a case. The introduction of this interactive machine learning system would transform the shared soap opera experience of a large team of lawyers into the personal voyeurism of the individual distributed users of the system.

In addition to this technical work, I’m also writing and drawing a comic telling the story of an individual working in a law firm in the year 2050. The comic is in the early stages, but I included rough versions of the first two pages in the presentation.