Policy Practicum: AI For Legal Help

3 credits · Law

In this hands-on studio class, students from accross the university will work with legal team partners, to build and evaluate new access to justice solutions. The driving question is whether and how public interest technology can be built to make quality, safe legal help more accessible to people in need. Student teams will work with legal aid teams doing frontline work on housing, re-entry, consumer debt, and public benefits law. They will learn about the current ways working, and identify a specific workflow where technology might help improve efficiency, reach, quality, and scale. Students will then learn design, development, and legal analysis skills in order to prototype, refine, and evaluate a new technology workflow that addresses the challenge. The course runs through one development and evaluation cycle between the two quarters. Winter quarter works with a legal aid partner to scope an opportunity specific to the partner's needs, prototype a solution, gather key data and standards, and refine a system with regular testing and stakeholder input. In Spring quarter, teams will turn towards evaluation and refinement of the system, to improve how the system performs vis-a-vis standards for accuracy, safety, usefulness, and sustainability. Students will learn how to determine how AI or rules-based systems can be deployed, with careful safeguards, to impact a high-stakes workflow in a legal setting. Students will also develop skills in real-world evaluation of AI systems, and policy-making around whether and how tools should move towards deployment. Legal, design, and technical work happen together, so teams mix students from the law school, the design school, computer science, and beyond. The work will involve collaboration with many stakeholders, including frontline advocates, end-users, and leaders. Teams must ground their technology work in the realities of nonprofit legal services: strict deadlines, high case volumes, sensitive client data, and ethical and professional obligations. Teams will produce a working prototype, a detailed workflow plan of how the tool will operate in practice, a memo critically documenting and examining their R&D journey and decisionmaking, and a suite of evaluation test sets, rubrics, protocols, and results. Students' work will feed into shared, open standards that other organizations can build on, to advance the common development of public interest technology. No legal or technical background is required. Students from across the university are welcome to apply. After the term begins, students accepted into the course can transfer, with consent of the instructor, from section (01) into section (02), which meets the R requirement. Elements used in grading: Attendance, performance, class participation, final paper. CONSENT APPLICATION: Interested students should submit a Policy Lab application (go to link SLS Registrar https://registrar.law.stanford.edu/). This is a Cardinal Course certified by the Haas Center for Public Service. Cross-listed with the d.school (DESIGN 809E).