A couple of weeks ago, I had the opportunity to help lead Computational Law workshops at the University of Michigan Law School and Michigan State University College of Law, working with Jeffrey Sharer, co-chair of Akerman’s Data Law Practice and Dan Linna, Director of LegalRnD and Professor of Law in Residence at Michigan State and Adjunct Professor of Law at Michigan. Professor Linna encourages students in his classes at Michigan and Michigan State working on innovation projects to tackle problems using a “people, process, data, and technology” approach. In our Michigan State capstone class, Professor Linna and I trained students in lean thinking, the Toyota Way, and the Improvement Kata during a workshop in early January. Students learned about design thinking during a February workshop with Margaret Hagan, the Michigan State version of which was attended by leading practitioners from around the country and Canada. The Computational Law workshop provided the opportunity to demonstrate how lawyers can innovate by breaking down legal tasks and reasoning into rules.
Rules- versus Data- Driven Artificial Intelligence
Professor Linna began the workshop with a discussion of rules- versus data-driven artificial intelligence. Rules-based artificial intelligence automates advice by applying a set of predefined rules or logic to user input. In a data-driven system, on the other hand, some subject matter expert (e.g., a lawyer) “trains” the system to recognize patterns by feeding it data with “labeled” outcomes. A rules-based artificial intelligence system, such as Neota Logic, works well to centralize knowledge and automate legal advice in an area that is heavily codified (e.g., bankruptcy or most other areas of civil law). Data-driven artificial intelligence systems, such as Kira, have the ability to automatically organize, flag, and prioritize certain document clauses, for example, greatly assisting in cumbersome tasks such as contract review and due diligence.
Akerman Data Law Center
Following Professor Linna’s discussion, Mr. Sharer introduced the students to the Akerman Data Law Center–an expert system he built using Neota Logic. The students in our LegalRnD capstone course at Michigan State are learning Neota Logic, and some teams are using it for their innovation projects with Akerman, Perkins Coie, Davis Wright Tremaine, and Michigan Legal Help. It was inspiring for them to see a finished product and how it is being used by one of the country’s most forward-thinking law firms. For me, the most important takeaway was Mr. Sharer’s point that it is easy to get swept up in the possibilities of legal technology; but for a product to truly improve the way legal services are delivered, its “legal inventory” must be “complete, current, actionable, and affordable.” To learn more about the Akerman Data Law Center, I encourage you to visit the Akerman Data Law Center website, and to read Professor Linna’s blog post detailing his interaction with the platform.
Comparing Legal Tech Use-Cases
As the final component to the lecture portion of the workshop, I discussed the differences between Neota Logic, ThinkSmart, and QnA Markup. Created by David Colarusso, Director of Suffolk LIT Lab, QnA Markup is an open-source platform that can be used for simple document construction and building rules-based question-and-answer systems. Mr. Sharer gave a great example of how Neota Logic can be used, so I focused most of my time on demonstrating use cases for QnA Markup and ThinkSmart, a workflow automation platform. To help prepare the students for the second half of the workshop, when they would build their own applications in QnA Markup, I demonstrated an application that I built that helps a user determine when a contract has been formed under UCC Article 2-206 or 2-207.
During my presentation, I emphasized how easy it had been to create this application because I first took the time to create a flowchart of the law. Once you map all of the possible outcomes and pathways a legal issue or process can take, building the expert system or automated workflow is relatively straightforward–even with sophisticated tools like Neota Logic and ThinkSmart.
I really enjoyed being a part of this Computational Law workshop. Many possibilities exists to better understanding legal processes, automate them, and improve legal-services delivery. I look forward to finding new ways to make legal advice more accurate, accessible, and actionable.