Developing An AI-Enhanced Authentic Oral Assessment Model (AI-AOAM) For Evaluating EFL Speaking Performance among ESP Hotel Management Majors at a Public University in China: A Proposal

This paper proposes a conceptual framework for developing an AI-Enhanced Authentic Oral Assessment Model (AI-AOAM) to assess English oral performance among hotel management students enrolled in English for Specific Purposes (ESP) courses at a public university in China. Traditional oral assessments in Chinese higher education are often summative, teacher-centred, and disconnected from authentic communicative contexts, particularly those relevant to professional fields such as hospitality. Such limitations raise concerns about the validity, reliability, and pedagogical value of current practices in assessing EFL speaking performance. Grounded in a pragmatic research paradigm, this study employs a mixed-method sequential explanatory design. The qualitative phase involves structured interviews with ESP instructors to explore current assessment methods and challenges, while the quantitative phase collects data through authentic speaking proficiency tests administered to students. The AI-AOAM is constructed around three key assessment dimensions- accuracy, coherence, and appropriateness of spoken English, each reflecting both theoretical foundations and industry-specific communication demands. To ensure validity, the model incorporates expert feedback from both the language education sector and the hospitality industry and undergoes correlational analysis with standardised English test scores. The proposed model aims to offer a more practical, context-sensitive, and occupation-oriented alternative to traditional assessment practices. Findings from this study are expected to provide valuable insights for language educators, curriculum developers, and policymakers, contributing to more effective and authentic oral assessment practices in ESP programs and improving the overall quality of EFL instruction in China’s tertiary education system.