A course model for teaching AI-augmented legal research
AI-Augmented Legal Research is a short, practice-oriented course developed at Vanderbilt Law School to help students learn how to use AI tools without losing the habits that make legal research reliable. The course pairs hands-on experimentation with traditional research skills, professional responsibility, confidentiality, drafting judgment, and verification. Students do not learn AI as a replacement for legal research; they learn how to evaluate AI outputs, document their process, and decide when AI belongs in a legal workflow at all.
The course has been taught in different calendar formats, so the structure below is organized by skill rather than by week. The order matters more than the exact schedule: foundations and confidentiality come first, source-specific research follows, drafting and verification build on that base, and the course closes with workflow, communication, and reflection. The model can be compressed, expanded, or adapted for a credit-bearing course, workshop series, orientation, CLE, or firm training.
Course arc: how the pillars become skills
The Start Here page introduces five teaching pillars: foundations, confidentiality and ethics, source-specific research, drafting and communication, and verification and workflow. This page shows how those pillars become a teachable course sequence. The nine skills below are not separate from the pillars; they are the course-level moves that help students practice the pillars over time, from framing research questions and protecting client information to verifying authority, documenting workflow, and communicating findings in practice.
Framing legal research in an AI environment
Establish what AI changes about legal research, and what it does not. Set the baseline norms: evaluation before reliance, professional judgment, and the lawyer's continuing responsibility for the work product.
AI ethics & confidentiality
Competence, confidentiality, supervision, and candor, the duties AI use most often implicates. Sanitization as a daily, defensible practice rather than a one-time review.
Case law research
Traditional case research alongside AI retrieval, summary, and synthesis. Where each performs well; where AI most often misleads with confidence.
Statutes, regulations & administrative materials
Primary law beyond cases. Source provenance, jurisdictional nuance, and the verification problems particular to non-judicial materials.
Secondary source research
How AI summarizes treatises, practice guides, and law-review articles, and where its structural fidelity to authoritative sources breaks down.
Drafting with AI
Memos, communications, and professional work product. Editing, judgment, and accountability for the final document, not for the draft that produced it.
Verification, hallucinations & liability
Identifying errors and unsupported assertions. The lawyer's obligation to independently confirm authority, and the liability landscape when that obligation fails.
AI in law firm workflows
How research decisions sit inside firm operations, billing, and risk management. Adoption shapes culture and professional expectations as much as it does technology.
Communicating research findings
Presenting AI-augmented research to partners, clients, and colleagues. Disclosure, documentation, and the professional defense of research choices.
How the skills fit together
Learning outcomes
The capstone asks students to compare traditional and AI-assisted research workflows, document what changed, and reflect on what each method revealed.
Why pass/fail?
The course is graded pass/fail, on purpose. The goal is error-free attorneys, and producing them requires that students be allowed to practice and fail, without the practice itself becoming the thing they are graded on. A student who is worried about a letter grade will not take the risk of a wrong sanitization choice, a botched workflow, or a critique that turns out to be wrong. Those are exactly the risks the course is designed around.
Pass/fail removes the pressure to produce an A-level answer and replaces it with the work that actually matters: thinking through scenarios, noticing complications, defending a choice, and revising it. It supports experimentation, error detection, and honest reflection. Students should leave the room having tried things that did not work, not having performed correctness for an evaluator.