ウメモト タカトヨ   Takatoyo UMEMOTO   梅本　貴豊    所属   京都外国語大学  共通教育機構    職種   教授
発表年月日	2026/09/10
発表テーマ	From accent variability to phonological consolidation: AI-driven TTS input in English as an international language learning in an EFL context
会議名	EUROCALL2026
学会区分	国際学会
発表形式	口頭（一般）
単独共同区分	共同
国名	アイルランド
開催地名	Ulster University, Belfast
発表者・共同発表者	Kasamaki, T., Takamichi, S., Yoshikawa, Y., Yui, K., Fujiwara, S., Ishikawa, Y., & Umemoto, T.
概要	This study examined the pedagogical use of AI-driven text-to-speech (TTS) in a year-long university English as a foreign language (EFL) course. In the first semester, learners engaged in listening activities using non-native English TTS, while the second semester combined accent-diverse exposure with native English TTS-supported reading-aloud practice. Twenty-two Japanese university EFL learners participated, with primary analyses based on sixteen who completed assessments in July and January. Listening performance and English as an international language (EIL) perceptions were assessed through listening tests, an EIL Perception Scale, and reflective responses. Listening performance improved significantly with a large effect size, although the EIL Perception Scale showed no significant changes. Qualitative findings indicated a progression from perceptual difficulty to adaptation, strategic adjustment, and increased awareness of English diversity. These findings suggest complementary roles for accent-diverse exposure and native English TTS-supported reading-aloud practice, informing a proposed dual-layer instructional model for listening development.