AI news ·
Law professor builds AI practice tools by starting with pedagogy, not technology
USF law professor Phillips spent two years building AI tools to protect critical thinking. Her five-step framework guides faculty to design tech around learning goals.
At the University of San Francisco School of Law, Professor Nicole Phillips has built a suite of AI-powered practice tools that help students master legal skills without replacing critical thinking. Over two years, she developed a case brief helper, a mediation bot, and an employment law counseling coach, all while openly admitting she has no coding background. Her method - a five-step framework detailed below - gives other law schools a replicable path to integrating AI responsibly.
Phillips, co-director of the legal research, writing, and analysis program, said faculty often make the mistake of starting with the technology rather than the problem. "It can't just be 'Let's use AI!' There has to be a specific learning outcome," she said.
Design before you build
Every tool began with a concrete student frustration. The Bluebook Citation Bot emerged because students wanted real feedback on citation formatting. A Socratic method assistant addressed anxiety about the cold-call technique that was preventing students from showing what they actually knew. Phillips maps out the entire student interaction and desired outcome before touching any platform. "I'm really thinking about what I want the students to get out of it and then working backwards from there," she said.
Constraints that force thinking
None of her tools will give a student the answer. The citation bot never produces a complete citation on demand; instead, it guides students toward understanding why a citation is structured the way it is. The Socratic assistant forces students to drive their own thinking and sit with discomfort in a private, low-stakes space. Before releasing a tool, Phillips tests it exhaustively - feeding it incorrect law to see if it pushes back, probing for any way it might accidentally reveal an answer. "I do a lot of testing and breaking and then rebuilding," she said.
From private rehearsal to public confidence
Students who use the tools arrive in class more willing to participate. Phillips said those who practice with the experiential simulations can point to that work in job interviews as evidence of real skills. For first-generation law students especially, the Socratic method can feel like a barrier. "Competence is often mistaken for confidence," Phillips said. "The opportunity to practice being wrong privately is really important."
Why this matters for legal professionals
The biggest barrier to building such tools is not technical skill, Phillips argues, but the willingness to design carefully around learning goals. For law school faculty, administrators, and legal training directors, the takeaway is clear: start with the problem students face, build in constraints that protect critical thinking, and let students help refine the tool. As more firms and schools turn to AI for Legal applications, the pedagogy-first approach ensures that technology supports rigorous analysis rather than bypassing it.