Solving for X: Exploring Data and AI Solutions for Justice Problems

Each session is organized around a set of related problems and potential solutions involving court data and AI. Presenters will briefly describe the problem and solution; the remainder of each session will be devoted to discussion and brainstorming about potential applications, extensions, risks, and limitations.

8:00 - 9:00 a.m.: Breakfast (optional)

9:00 - 9:15 a.m. : Welcome and Run of Show 

• Deven Desai and Charlotte Alexander, Georgia Tech

9:15-10:15: Data Standards and Court Statistics: Creation, Sharing, Integration, Analysis

This session establishes a foundation for the day’s discussions of court data and data-driven AI tools. Presenters will cover the Court Statistics Project and National Open Court Data Standards (NODS), developed by the National Center for State Courts (U.S.), as well as courts’ on-the-ground use of standardized court data to track and analyze productivity, time lags, case activity, and judicial resource management.

Presenters:

• Nicole Waters, National Center for State Courts
• Diane Robinson, National Center for State Courts
• Jarrett Perlow, Circuit Executive and Clerk of Court, U.S. Court of Appeals for the Federal Circuit

10:15 - 11:15 a.m.: Data Availability: Creating Structured Data from Court Records Using AI

Not every court system generates clean, structured data that is amenable to analysis. Valuable information is often buried within the text of court documents and docket sheets. Unlocking that information requires transforming unstructured text into structured data that can support analysis and innovation. This session explores the use of Large Language Models (LLMs) and other Natural Language Processing (NLP) techniques to create structured court datasets.

Presenters:

• Adam Pah, Georgia State University, SCALES-OKN
• Dave Schwartz, Northwestern Law School, SCALES-OKN
• Raphaël Gyori, Université Libre de Bruxelles, Smart Law Hub [Zoom]

11:15 - 11:30 a.m.: Break

11:30 a.m. - 12:15 p.m.: Court Congestion and Delay: Process Mining and Simulation

Court congestion makes headlines on a regular basis. AI-based simulations, running on structured data extracted from court documents, could help courts identify litigation bottlenecks and test the impact of staffing and process changes on court backlogs. Similar approaches have been developed in the hospital emergency room context, where leveraging event logs of patients’ arrival times and progress through various stages of care allows researchers and administrators to experiment with simulating process changes and test the impact on patient wait times. This session will explore the potential of such tools in the justice system context.

Presenters:

• Shany Azaria, University of Toronto, SiMLQ
• Opher Baron, University of Toronto, SiMLQ

12:15 - 12:30 p.m.: Break – grab food and return to main room

12:30 - 1:30 p.m.: Working Lunch, Improving Access Through Court Process Modernization

Courts are exploring methods to enhance access, particularly for self-represented and low-resourced litigants. This session explores the potential for technological and digital innovation in this space, with examples from the Los Angeles Superior Court’s partnership with Stanford Law School. The session will also cover the Pew Charitable Trust’s work on court modernization, guided by the principles of openness, effectiveness, and equity.

Presenters:

• Daniel Bernal, Stanford Law School
• Brian Borys, Los Angeles Superior Court [Zoom]
• Darcy White, Pew Charitable Trusts Courts and Communities Project

1:30 - 2:30 p.m.: Data Linkage: Knowledge Graphs and Person-Centered Court Data

What can data tell us? Structuring unstructured text is just the first step in the process of sense-making with court data. This session explores the potential for knowledge graphs—a method for organizing court data and linking it with external data sources—to create a system-wide picture of the courts and generate insights. It also explores new ways to conceptualize court data, reorganizing around the litigant and tracking each litigant’s multiple touchpoints with the courts, for example, rather than around the case as the unit of analysis.

Presenters:

• Kathryn Albrecht, Georgia State University, Integrated Justice Platform
• Lauren Sudeall, Vanderbilt Law School

2:30 - 2:45 p.m.: Break

2:45 - 3:45 p.m.: Data Governance and Responsible AI: Privacy, Security, Ethics, and Safety

A theme that cuts across all others is the responsible use of data and AI in a court context. This session explores issues of privacy, security, ethics, and safety as they pertain to AI, data, and court operations.

Presenters:

• Robert Deyling, Administrative Office of the U.S. Courts
• Amy Salyzyn, University of Ottawa [Zoom]
• Leo You Li, Stanford University [in person or Zoom]

3:45 - 4:45 p.m.: Decoding Legalese: Summarization of and Information Extraction from Legal Documents

Many jokes highlight lawyers’ wordiness and reliance on “legalese.” At these jokes’ root is the real-world complexity and volume of legal language, which can make the legal system opaque and inaccessible to the general public and even to judges themselves. This session highlights researchers’ work on AI-enabled summarization of and information extraction from legal text, to promote transparency and aid judges’ own work in analyzing precedent and crafting decisions.

Presenters:

• Aniket Kesari, Fordham Law School
• André Lage-Freitas, Universidade Federal de Alagoas, Brazil
• Student Researchers, Law, Data & Design Lab @ Georgia Tech [Zoom]

4:45 - 5:00 p.m.: Concluding Remarks and Farewell

• Charlotte Alexander, Georgia Tech