Exploring the Development of Chinese College Students' Proficiency in English through ChatGPT: An Experimental Study

A semester-long teaching reform experiment was conducted at a university in China, utilizing ChatGPT and other large language models to support the teaching of College English (Comprehensive English). In the experimental class, students employed a ChatGPT-assisted teaching mode, through a “fragmented mobile use” approach for English acquisition under the guidance of the researchers. This method significantly improved weekly English learning time, English learning interest and learning autonomy, classroom participation, and the final examination results of College English courses. Consequently, English learning anxiety was reduced. In contrast, the control class, which followed a traditional teaching mode, did not exhibit similar improvements in these parameters. The experiment demonstrated that the effective use of generative large language models, such as ChatGPT, can significantly enhance second language acquisition and gain high student approval. However, a few students still showed a lack of interest in English learning, and AI struggled to address issues related to learning initiative.