Generative AI is revolutionizing oral language assessment by providing innovative solutions to alleviate the burden on teachers [1]. In traditional settings, particularly in HE classrooms with large student-to-teacher ratios, assessing each student’s speaking abilities can be time-consuming and challenging, often resulting in inconsistencies and subjectivity. Generative AI addresses these issues by offering scalable solutions that maintain high standards of reliability and objectivity. It can analyze various aspects of spoken language, such as pronunciation, fluency, grammar, and vocabulary usage, providing detailed and immediate feedback [2]. Moreover, it incorporates multimodal cues, such as facial expressions, gestures, and body language, into oral performance assessment tasks, assessing not only linguistic competence but also communicative effectiveness and sociolinguistic appropriateness [3]. A compelling case study – a classroom with 145 students from the University of Padua – exemplifies the transformative impact of generative AI in oral language assessment. In this setting, the application of AI-driven assessment tools significantly improved the lecturer’s effectiveness and efficiency. Previously overwhelmed by the sheer number of students, the lecturer was able to manage assessments more systematically, as the AI system provided consistent evaluations and immediate feedback on students’ spoken language skills. This multimodal approach to oral language assessment offers a more nuanced understanding of learners’ oral communication skills and fosters the development of communicative competence in real-life contexts [4].