An Empirical Study on AI-Enabled College English Oral Teaching: Constructing a Dynamic Closed-Loop Mechanism Based on the V-Model

Abstract

This study explores the effectiveness and innovative path of empowering college English oral teaching with artificial intelligence (AI) technology. Aiming at the problems of single feedback and low efficiency in traditional oral teaching, this paper studies the V-shaped model in system engineering, constructs a three-dimensional dynamic closed-loop intervention mechanism of AI error correction – teacher guidance – peer evaluation, and integrates the domestic large model (ERNIE Bot) and professional oral training platform (FIF) to achieve multi-dimensional diagnosis and personalized learning. Through a two-stage comparative experiment (25 students in each experimental group and control group), the study found that AI technology significantly improved students' oral ability. The experimental group made significant progress in pronunciation accuracy (+37%), grammar and vocabulary richness (+42%), and content coherence (+29%). The multi-dimensional feedback mechanism effectively compensates for the shortcomings of traditional teaching, and the accuracy of peer evaluation has increased from 40% to 65%. The study also revealed the potential of AI technology to drive the transformation of teaching models, such as personalized material generation before class, real-time interaction during class, and 24/7 tutoring after class. Finally, the article discusses issues such as technological limitations, teacher role transformation, and ethical fairness, providing replicable practical paradigms for educational intelligence.