Back to search

University of San Francisco

Justice on Autopilot: Comparing Predictive Litigation and Data Justice in Bay Area Municipal Governments

Abstract

dc:description.abstract

This Capstone project examines how Bay Area municipal governments, specifically the City and County of San Francisco, Alameda County, San Mateo County, and Santa Clara County, are adopting and governing predictive litigation technologies, and what these practices reveal about the possibilities and limitations of data justice in advancing racial equity, transparency, and public accountability. Drawing on a qualitative, comparative case study methodology analyzing procurement contracts, AI governance policies, California Public Records Act (CPRA) disclosures, and semi-structured interviews gathered across fifteen municipal legal institutions, the research applies an interdisciplinary theoretical framework synthesizing John Rawls’s distributive fairness, Nancy Fraser’s recognition justice, Tom R. Tyler’s procedural legitimacy, and Linnet Taylor’s data justice to move beyond narrow debates about algorithmic accuracy toward a structural analysis of power, accountability, and equity. Findings reveal a consistent pattern across all four counties: AI-assisted tools are already deeply embedded in prosecutorial analytics, evidence management, legal research, and pretrial decision-making, yet no institution examined conducted a pre-adoption equity impact assessment, required independent vendor bias testing, established community advisory structures, or implemented proactive public disclosure of AI governance policies; failures exemplified most starkly by the San Mateo County District Attorney’s use of Closure Intelligence, a purpose-built prosecutorial analytics platform procured without competitive bidding, equity review, or any disclosure mechanism for affected defendants. These governance failures are not primarily technical but structural, reflecting institutional choices to prioritize efficiency over the procedural fairness, recognition of marginalized communities, and democratic accountability that legitimate legal administration demands, and the project concludes with jurisdiction-specific recommendations, including mandatory AI inventories, enforceable vendor accountability provisions, defendant notification standards, independent bias audits, and formal community advisory structures, advancing the central argument that artificial intelligence in municipal legal systems must remain subordinate to democratic accountability and the enduring pursuit of equal justice. Keywords: predictive litigation analytics, artificial intelligence governance, data justice, algorithmic bias, municipal government, racial equity, procedural legitimacy, Bay Area legal systems, public accountability, AI ethics

Degree

thesis:*
Name thesis:degree_name
Master of Arts in Urban and Public Affairs
Level thesis:degree_level
Project/Capstone - Global access
Discipline thesis:degree_discipline
Urban and Public Affairs

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Thomas, Darris M.
Contributors dc:contributor
  • Dr. Patrick Murphy
  • Dr. Kresten Froistad-Martin

Subjects

dc:subject × 19

Identifiers

dc:identifier.*
Repository record dc:identifier
https://repository.usfca.edu/capstone/2009

Chain of custody

source
Harvested from
University of San Francisco
Base URL
repository.usfca.edu/do/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Thomas, Darris M.. Justice on Autopilot: Comparing Predictive Litigation and Data Justice in Bay Area Municipal Governments. Project/Capstone - Global access thesis, https://repository.usfca.edu/capstone/2009