Li Rong

Murad Abdu Saeed

Universiti Malaya, Malaysia

Abstract

Despite the increasing use of artificial intelligence (AI) tools in informal English-speaking practice, there is still limited understanding of multilingual learners’ subjective perceptions and engagement patterns in this context. To fill this gap, the current study aimed to identify core subjective factors and explore how learners with different first language (L1) backgrounds and English proficiency levels differ in their perception and engagement patterns. The study replicated the Q-methodology of Guo and Xia (2025), who explored learners’ perception of AI-mediated informal digital multilingual learning, in investigating perception of AI-assisted informal speaking practice (AI-IDESP) among multilingual undergraduate students in Malaysia. Using the Extended Unified Theory of Technology Acceptance and Use (UTAUT2), and a Q-Methodology, the present study surveyed and interviewed 30 multilingual undergraduate students in a public Malaysian university who had engaged in AI-IDESP and use at least one other language in daily communication. The findings show that performanceoriented use of AI, emotion-driven motivation and efficiency-driven but anxiety dependence influenced learners’ behavioural intentions and actual engagement with AI-IDESP. Students’ different English proficiencies were more likely to align with these different factors. Overall, by deepening our understanding of how learners’ AI-IDESP is shaped by performance orientation, emotional sensitivity, and efficiency-related anxiety, the study offers useful implications for learner-centred AI tool design in multilingual contexts.

Keywords

English language learning, Q-methodology, AI-IDESP, AI speaking practice, multilingual learners