Speaker
Description
As AI technology gets more embedded in a mobile learning context, it is worth investigating the potential of AI to facilitate oral language development and learner engagement. In this article, we consider the integration of two mobile AI applications, called VoiceGPT and TalkBackAI, to assist English language learners with real-time speech practices and feedback. Based on a mixed-method research design, the study was conducted with 268 students attending EFL speaking classes in Vietnam from three different universities.
Quantitative data were obtained through a structured survey measuring Perceived Usefulness, Speaking Confidence, and Engagement, with qualitative data from focus group interviews providing contextual information. Confirmatory factor analysis corroborated the three-factor structure of the questionnaire, which demonstrated high internal consistency and good construct validity. Students indicated that they were speaking more spontaneously, experiencing less practice anxiety, and becoming more motivated to use English outside the classroom.
Nevertheless, results indicated limitations in voice recognition accuracy for regional accents and variable response time. The study suggests mobile AI could help increase access to speaking practice and offers possibilities for enhancing voice-responsive technologies in language learning.
Keywords: Mobile AI, Speaking Confidence, Language Learning, VoiceGPT, Student Engagement, EFL Education
