Speaker
Description
This study investigates the use of iCorrect, an AI-generated application, in supporting automated speaking assessment at ULIS Middle School. In the context of increasing attention to English-mediated learning and functional language assessment in Vietnamese schools, speaking assessment remains a demanding area due to its reliance on teacher judgement, time-intensive scoring procedures, and the need for consistent feedback. The study aims to examine the extent to which iCorrect can provide reliable, timely, and pedagogically useful feedback on students’ English speaking performance. Adopting a mixed-methods design, the research analyses students’ recorded speaking tasks, AI-generated scores and feedback, teacher ratings, and perceptions from teachers and learners regarding the usability and educational value of the application. The findings are expected to clarify the alignment between AI-based assessment and human evaluation, particularly in relation to fluency, pronunciation, lexical range, grammatical accuracy, and task achievement. The study also explores how automated feedback may support learner autonomy, reduce teachers’ assessment workload, and contribute to a more data-informed approach to speaking instruction. At the same time, it critically considers limitations related to scoring transparency, contextual appropriacy, and the continued need for teacher mediation. The paper argues that iCorrect should not replace teacher assessment, but may function as a complementary tool within a broader formative assessment framework for English language education at the secondary level.
Keywords: automated speaking assessment; artificial intelligence; iCorrect; English language teaching; formative assessment; ULIS Middle School
