Artificial intelligence-based assessment in ELT exam creation: a case study of Van Lang University lecturers

Purpose Focusing on the growth of artificial intelligence (AI) in education, this research reveals that AI can create and improve English language assessments for learners in order to optimize and enhance test questions as a bilateral tool in the traditional way for English Language Teaching (ELT) is possible. Design/methodology/approach The research adopted a qualitative methodology by conducting semi-structured interviews with a varied range of language institutes’ lecturers, revealing new beneficial effects of AI on test time, content and human variables. Findings Several interviewees agreed that AI should be used in ELT exam creations because of its overt advantages in making test items automatically, adaptive testing, enhanced feedback mechanisms and quality assurance and innovative formats. Simultaneously, some disadvantages are recorded, including complexity and nuance of language, technical limitations, ethical and bias concerns and human oversight and validation. Research limitations/implications The study was also limited by the time frame of the research, which may not have fully captured the complex dynamics between the different actors, such as using AI in preparing questions for reading tasks such as automatic creation of pre-reading questions as well as possible answers. Originality/value For future studies, as AI-generated material is becoming more ubiquitous, from music to artwork, it presents crucial legal problems regarding who owns the rights to the work or construct ELT exams. It has also become the next problem that the writers should concentrate on.