Natural Language Programming (NLP) is a field of computer science that employs predictive analytics and machine learning with a focus on the interaction between computers and both written and spoken language.  Over the past decade, courts have successfully employed NLP techniques for routine case management tasks such as redaction, data extraction, document classification, and automated data population into databases.  With generous funding from the State Justice Institute, the National Center for State Courts (NCSC) explored the feasibility of NLP to support civil case management.  The project focused on two stages of civil case management — assigning cases at filing to an appropriate case processing track and reviewing case filings to ensure that fundamental guarantees of procedural due process are satisfied before entering final judgment.  The Conference of Chief Justices and the Conference of State Court Administrators identified both stages as critical to effective civil case management in their recommendations for the Civil Justice Initiative. 

The studies employed NLP techniques for data extraction, relational matching, and document classification, comparing its performance with the same tasks performed by trained staff.  On data extraction tests, NLP performed at 90% or higher on most data elements needed for case triage, and likely would have performed even better with additional human direction during the machine learning phase of the study.  It was less successful on the document classification test, but largely due to poor image quality on the documents themselves rather than inability to accurately extract the necessary data for analysis.  Based on the findings, NCSC emphasized the need for adequate human involvement in the machine learning phase to ensure consistent and accurate performance as well as high quality data on which to begin learning.  In addition to the final report, the NCSC developed instructional videos on NLP and other artificial intelligence technologies, strategies for implementing these technologies, and a resource guide, all of which are available at https://www.ncsc.org/cji.