AI-Supported Interpreter Training: Designing Multimodal Practice and Assessment Tools

Abstract

Court interpreter training asks students to perform bilingually under cognitive load, ethical pressure, and the specialized terminology of legal settings. Classroom role-play and instructor-led simulations remain valuable, but they are difficult to scale when each student needs individualized, repeated practice. This presentation introduces a set of AI-supported pedagogical tools developed for the community court interpreter training program with Utah State University and Utah Courts, which discusses best practices for integrating generative AI into specialized language instruction. The toolkit was designed to give students frequent, low-stakes practice while reducing the time instructors spend producing custom materials. The course incorporated four kinds of activity: AI-generated legal texts for sight translation; conversational AI agents simulating courtroom interactions for consecutive interpreting; multi-speaker audio for simultaneous-interpreting exercises; and interactive mock certification scenarios aligned with the criteria used by professional certifying bodies. All four were delivered through Canvas and required no specialized infrastructure. The presentation will demonstrate how these tools scaffold terminology, ethical reasoning, working memory, and interpreting performance, and will share preliminary self-efficacy findings along with practical questions about AI reliability, prompt design, and responsible classroom use. Keywords: generative AI; interpreter education; instructional technology; language pedagogy; experiential learning; AI-assisted practice

Presenters