PRIIT TAMMETS

AI-Assisted Assessment of Legal Case Solutions: Building a RAG-Based Feedback Tool for Legal Education

Legal education depends on case-solving practice, yet providing timely, structured feedback on dozens of student submissions is one of the most time-consuming tasks an instructor faces. This presentation introduces Ragistrat, a Retrieval-Augmented Generation (RAG) tool that uses large language models grounded in the Estonian legal corpus to evaluate student case solutions according to the IRAC method — Issue, Rule, Application, Conclusion.

During the presentation, the audience will see how the system ingests statutes and legal materials into a vector database, retrieves the most relevant norms for each submission, and produces structured, norm-referenced feedback that instructors can review, amend, and publish. Equal attention is given to the harder questions: hallucination control, pedagogical framing of AI feedback, audit trails, and the boundary between automated assessment and instructor judgement.

The goal is not to replace the legal educator, but to give students faster formative feedback and instructors a transparent, auditable starting point for grading and to open a discussion about what responsible AI assistance in legal education should look like.

Priit Tammets is an analyst at Tallinn University's School of Digital Technologies, where he has worked for nearly two decades at the intersection of digital technologies, education, and learning analytics. His work focuses on educational innovation, digital learning environments, and AI-supported teaching and learning. He currently contributes to projects on Teacher-AI complementarity and the TLU AI Teaching and Learning Lab. Priit has led and contributed to numerous national and EU-funded research and development projects, supervised Master's theses, and co-authored peer-reviewed publications in technology-enhanced learning.