Law Student Builds AI Tool to Detect Bias in Court Cases
Marilyn Aguilar-Portillo, a second-year law student at UC Law San Francisco, designed an AI system that identifies patterns of bias in legal cases. The tool, built during the school's AI-Enabled Lawyers Bootcamp, analyzes case files to surface subtle language and assumptions that can skew outcomes.
The bootcamp, offered for the first time this spring, trains students to use AI for core legal work-research, drafting, discovery, and document analysis-while addressing ethics questions around privilege and professional responsibility. Students complete capstone projects by building their own AI tools, often without coding experience.
How the Tool Works
Aguilar-Portillo's capstone project, called the CRT Trial Auditor, applies Critical Race Theory to examine real case files. The system identifies coded language, unfair assumptions, and what she calls "shadow narratives" that can reduce people to stereotypes in court.
"It helps ensure people aren't reduced to stereotypes or stripped of their complexity in the courtroom," Aguilar-Portillo said.
Learning by Building
Aguilar-Portillo enrolled in the bootcamp because she saw AI as the next major shift in how society operates. The eight-session course taught her that the best way to learn AI is through hands-on experimentation.
Working with AI models like Claude gave her insight into both their capabilities and limits. She learned how to use prompt engineering to surface bias patterns and how to think through the ethical stakes of building these systems.
"When AI is used intentionally and creatively, it can expand my reach as an attorney, allowing me to help more people in a more personalized and meaningful way," she said.
Practical Application for Practice
The bootcamp gave Aguilar-Portillo a framework for integrating AI into every stage of her legal work. She views the technology as a tool that lets lawyers work faster and at greater scale without changing what they fundamentally do.
For legal professionals interested in AI for legal work, the bootcamp model shows how to move beyond theory into practical application before entering practice.
The course was taught by Adjunct Professors Luis Villa and Zoe Dolan, both practicing attorneys with experience integrating technology into legal practice.
Your membership also unlocks:
