AI-Assisted Assessment of Speaking Competence: A Methodological Framework

This article examines the use of artificial intelligence (AI) in assessing students’ speaking competence in English for Specific Purposes (ESP), particularly in Business English courses. The research problem concerns the need to combine automated speech analysis with the assessment of contextual, interactional, and professional-pragmatic aspects of oral communication. The aim of the study is to develop a methodological framework that integrates AI-supported analysis with professional teacher judgment. The study adopts a conceptual and methodological approach based on literature analysis, comparison of assessment principles, and theoretical synthesis. Its principal outcome is the proposed six-stage AAAS (AI-Assisted Assessment of Speaking) Framework, which comprises speaking task design, AI-supported speech analysis, rubric-based assessment, teacher verification, individualized feedback, and targeted practice followed by reassessment. The framework identifies six assessment criteria: fluency and coherence, pronunciation and intelligibility, grammatical accuracy, vocabulary and Business English terminology, interactional competence, and professional-pragmatic competence. These criteria are intended to represent complementary dimensions of speaking performance in professional communication. The article concludes that AI should serve as an analytical support tool rather than replace teachers’ professional judgment. Teacher verification remains essential for interpreting context, checking automated outputs, and evaluating communicative appropriateness. The proposed framework is conceptual; its reliability, validity, fairness, and practical effectiveness have not yet been empirically established and require further investigation.