English Medium Instruction Educator Language Assessment Literacy and the Test of Generative AI in Online Higher Education

Abstract

Generative Artificial Intelligence (GenAI) is often portrayed as disruptive. Higher Education (HE) assessment is not exempt from this, although the implications for multilingual settings remain an area of limited exploration. Drawing on the scarce literature in English as a Medium of Instruction (EMI) assessment and EMI educator language assessment literacy (LAL), this study sought to explore EMI online HE educator LAL and awareness of GenAI’s potential impact on established language assessment praxis. A sequential explanatory mixed-methods approach was used, comprising a survey on LAL self-perceptions (n=174) and semi-structured interviews (n=12). Findings illustrate general tendencies of low levels of LAL and practitioner unease towards GenAI-assisted academic misconduct. A heightened lack of confidence in GenAI tool usage detection efficacy and HE institutional capacity to respond to their evolving capacities in a timely manner was also found. It is therefore suggested that, given the complexity and continuing swift development of GenAI tools, the implementation of continuous professional development programmes focused on enhancing EMI educators' language assessment literacy and competence in using GenAI technologies is prioritised. These findings underscore the need for initiatives that not only improve technical skills but also address ethical considerations and strategies to uphold academic integrity in the face of emerging GenAI capabilities.