This study examines the influence of AI-driven technology on academic English education through a comparison of 100 undergraduates in two groups: an AI-enhanced learning group (n = 50) and a traditional instruction group (n = 50). The results show a statistically significant improvement in academic performance in the AI group (p < 0.01), particularly in standardized assessments. A strong positive correlation (r = 0.81) was found between engagement frequency and academic performance. While writing proficiency improved at a slower rate than reading comprehension, AI’s feedback mechanisms were found to be less effective for more complex tasks. This research highlights AI’s dual role in enhancing student engagement and reducing equity gaps, with high-frequency users demonstrating the most optimal learning patterns. It introduces an adaptive feedback framework and support strategies to address participation gaps. These innovations help advance a hybrid teaching model, showing AI’s potential in academic language education and emphasizing the need for context-sensitive AI tools.