AN EXPLORATION OF SCENARIO-BASED TEACHING IN LEGAL TRAINING FOR ADMINISTRATIVE RECONSIDERATION ASSISTED BY GENERATIVE ARTIFICIAL INTELLIGENCE

Abstract

Practical training in administrative reconsideration has long been mired in formulaic case simulations, making it difficult to replicate the adversarial nature and uncertainty found in real-world cases. The rapid development of generative artificial intelligence offers new possibilities for overcoming this challenge; it can not only create diverse virtual case scenarios but also simulate real-time dialogue between parties and provide rapid feedback. Focusing on the specific context of administrative reconsideration legal training, this paper proposes a “scenario-based teaching” design approach. Under this approach, generative AI serves as a scenario builder, a dynamic interactive partner, and a teaching assistant providing differentiated feedback, thereby supporting students’ learning within highly realistic dispute resolution scenarios rather than replacing their own critical thinking. The teaching practice adopts a three-stage progressive model: “pre-class AI-assisted scenario construction—in-class role-playing and structured adversarial exercises—post-class AI-driven in-depth debriefing.” Before class, students receive randomly generated case scenarios, draft reconsideration applications, and engage in mock dialogues with the AI as parties to the case; during class, they undergo in-group reviews and full-class mock hearings, with the AI introducing plausible “surprises” to force students to adjust their strategies in real time, while instructors identify cognitive blind spots and visualize thought processes; After class, the AI conducts structured debriefing interviews and generates debriefing summaries for instructors to provide detailed feedback, ensuring that reflection truly facilitates a shift in thinking. This process shifts instructors from “demonstrators” to “directors-monitors,” emphasizing critical scrutiny of AI outputs and establishing a process-oriented assessment system centered on thought processes. This initiative aims to unlock the potential of generative AI in simulating professional practice scenarios, making training more closely aligned with the actual logic of dispute resolution, and ultimately fostering students’ legal reasoning and practical skills.