ChatGPT and Text-to-Speech with Instructional Motivation in Online Oral English Learning

ChatGPT features anthropomorphic dialogue generation and supports multiple rounds of dialogue, making it suitable for teaching spoken English online. This study utilizes ChatGPT to reconstruct texts with different CEFR levels based on the Common European Framework of Reference. Based on instructional motivation strategies, the language style of the generated content is adjusted according to the learners' learning material. Different speech styles are then generated using the Single-Stage Text-to-Speech model (Bert VITS2). Compared with traditional online English teaching, this method meets the requirements of improving the emotional experience and promoting cognitive learning in online oral English teaching.