Exploring the Impact of AI-Assisted College Oral English Teaching on Students’ Oral Proficiency

Abstract

Addressing the three common and prominent practical dilemmas in current college English oral teaching — namely, reluctance to speak, delayed feedback, and lack of personalization—this study closely integrates the widely accepted core theoretical framework of modern foreign language teaching in second language acquisition theory: the comprehensible input hypothesis, the interaction hypothesis, and the affective filter hypothesis. Using 32 non-English major undergraduates from a comprehensive university in Northwest China as participants, a pre-test-post-test quasi-experimental design was employed, supplemented by qualitative research methods such as classroom observation and student interviews. This study systematically explores the application effects, key influencing factors, and optimized implementation paths of artificial intelligence (AI) technology in college English oral teaching. Through a 12-week systematic and standardized teaching intervention, this study rigorously compared and analyzed various data indicators of students before and after the oral proficiency test. It comprehensively integrates real-time classroom observation records, detailed student learning logs, and in-depth one-on-one brief interview results to deeply reveal the specific impact of AI-assisted teaching on students’ oral fluency, pronunciation accuracy, vocabulary and grammar application ability, and expression confidence. During the research process, the common pitfall of “emphasizing form over substance” in the application of artificial intelligence technology was consciously avoided. Ultimately, it provided a scientific and operable practical reference and solid theoretical support for improving the quality and efficiency of college English oral teaching, and helped to break through the development bottleneck that has long plagued the current college English oral teaching.