| AI-Assisted Speaking Assessment and Traditional Assessment of Speaking and their effects on pronunciation and fluency of Iranian EFL teacher trainees |
| کد مقاله : 1079-ELTCONF6 |
| نویسندگان |
|
اکبر مولایی * English Department, Farhangian University, P.O. BOX: 14665-889 Tehran Iran |
| چکیده مقاله |
| This study explored the comparative effect of AI-assisted oral production assessment and traditional teacher-based valuation on the speaking performance of Iranian EFL teacher trainees. Sixty-eight male freshmen TEFL teacher trainees at Allameh Tabatabaei University contributed. They were classified equally into two groups. Experimental group received AI-assisted assessment treatment using Speech Analyzer and Grammarly with immediate feedback; the control group received traditional teacher feedback using rubric-based evaluation from Farhadi, Jafarpoor, and Birjandi’s speaking assessment model. At the end of the semester, all participants delivered a speech of five-minute followed by an interview, rated by two examiners based on a 20-point rubric covering accuracy, fluency, pronunciation, grammar, and accent. Results from independent samples t-tests showed that the AI-assisted group significantly outperformed the traditional group overall. However, the traditional group confirmed significantly better performance in pronunciation and fluency. The findings suggest that AI tools can improve certain speaking capabilities but may not substitute human-mediated pronunciation and fluency growth. It can be suggested that educators enjoy a hybrid method of assessment including AI assisted feedback and teacher traditional feed back in order to improve both fluency and accuracy in pronunciation and speaking. |
| کلیدواژه ها |
| : AI-assisted assessment, traditional assessment, speaking performance, EFL, teacher trainees, Iran, oral presentation, pronunciation, fluency |
| وضعیت: پذیرفته شده برای ارائه شفاهی |
