ChatGPT Voice as a Tool for Improving Pronunciation in English: A Case Study of Spanish-Speaking University Students
摘要
Mastering English pronunciation presents challenges for Spanish-speaking students due to differences in phonetic systems, the variety of vowels, and the articulation of consonants. This study aims to identify the main pronunciation difficulties and to use ChatGPT Voice to enhance the pronunciation skills of Spanish-speaking university students. A three-month quasi-experimental design was implemented, involving 30 Spanish-speaking university students enrolled in an A1-level English course. Data was collected through initial and final pronunciation tests, and a satisfaction questionnaire regarding the tool was also administered. The results show that errors in the production of /θ/ and /ð/ decreased to 15%; recognition and production of vowel contrasts, such as /iː/ vs. /ɪ/, improved to an reduction to 18% in errors; errors in the production of aspirated consonants (/p/, /t/, /k/) decreased to 14%; and improvements in rhythm and fluency in stress-timed patterns resulted in an reduction to 18% in errors. The study concludes that integrating ChatGPT Voice into English language instruction highlights the potential of AI-based tools to provide real-time feedback and offers an innovative approach to pronunciation training. Future research should explore the use, application, and long-term effects of AI tools on fluency and prosody across diverse student contexts.