AI-Driven Interpreting Training: Students’ Perceptions of ChatGPT as a Spoken Trainer with Human Guidance

Abstract

All disciplines of scientific enquiry have fruitfully deployed artificial intelligence, and language education is no exception. In this study, twenty participants from the Translation Department at Misurata University, including both recent graduates and advanced students who had successfully completed their core interpreting coursework, explored their perception of using ChatGPT Plus as a spoken interpreting trainer within a structured, human-guided framework. The research employed a mixed-methods approach utilizing a comprehensive survey designed to capture both quantitative trends and qualitative perceptions. This instrument assessed the tool's effectiveness in supporting vocabulary acquisition, improving reflexes, translation accuracy, and both consecutive and simultaneous interpreting performance. Findings revealed that students perceived the tool as helpful and engaging, particularly for consecutive interpreting training, as ChatGPT Plus provided participants with flexible, self-paced oral practice. However, technical issues hindered simultaneous interpreting training. The tool always stopped delivering speech when participants began interpreting, mistaking their rendition for a conversational command. Other limitations included early task termination and overly positive automated feedback that occasionally overlooked translation inaccuracies. The findings highlight that while AI tools can improve student motivation and support mechanical skill development, they are most effective when integrated into a framework of consistent human guidance.