Harnessing AI for Oral English Proficiency Enhancement in Non-native Tertiary Teachers

Tertiary teachers in Hong Kong as subject matter experts often receive training in pedagogical training but lack training in presentation skills, which are crucial for helping students understand complex and abstract concepts. This study examines the effectiveness of using AI speech-recognition and generative AI tools to enhance non-native teachers’ oral English skills. Throughout a semester-long business course, the teacher reviewed the AI-generated transcripts after each lecture, focusing on four dimensions of oral skills, namely grammar, vocabulary, phonology and discourse, with weekly feedback from a generative AI tool. After tracking pronunciation for 13 weeks, the teacher’s pronunciation accuracy improved, evidenced by a 25% reduction in the effective Word Error Rate (WER). The teacher also experienced a significant reduction in unnecessary repetitions and long pauses, indicating improvements in vocabulary and discourse skills. Moreover, the teacher demonstrated increased proficiency in adjusting speech rates based on the cognitive demands of the material, as evidenced by variations in words per minute (WPM) in later lectures. However, our results suggested that the teacher’s grammar skills did not improve as much as in other dimensions. This research presents a practical, self-sufficient and embarrassment-free approach for university faculty members to independently improve their English delivery. Drawing on our experiences, this study also explores effective prompting techniques for AI tools in oral English proficiency enhancement and highlights the limitations of AI technologies encountered during the language development process.