Intelligent personal assistants as ‘scaffolding on demand’: Mechanisms of impact on L2 listening performance and anxiety among EFL learners

This mixed-methods study uses the context of second language (L2) learning to investigate the psychological mechanisms through which technology-mediated scaffolding impacts cognitive performance and affect regulation. A quasi-experimental design was employed, involving 57 intermediate Chinese EFL learners divided into experimental (n = 27) and control (n = 30) groups. The experimental group used Google Assistant for L2 listening practice, while the control group engaged in traditional listening activities. Quantitative data were collected using the Listening Comprehension Subscale of the TOEIC and the Foreign Language Listening Anxiety Scale (FLLAS), with pre- and post-test assessments conducted. Qualitative data were gathered through semi-structured interviews with 12 participants from the experimental group to explore their experiences. The results from paired-samples t-tests and ANCOVA analyses revealed that the experimental group showed significant improvements in L2 listening performance (p < .001, η2 = 0.435) and a substantial reduction in listening anxiety (p < .001, η2 = 0.300) compared to the control group. Thematic analysis of learner narratives suggested key psychological mechanisms: the IPA's 'scaffolding on demand' features appeared to reduce perceived extraneous cognitive load and enhance learners' sense of autonomy. These mechanisms, in turn, appeared to increase engagement and confidence. This study provides ecologically valid evidence for how user-controlled technological tools may facilitate cognitive processing and emotion regulation during a complex task, offering preliminary insights that inform psychological theories of learning, cognition, and motivation.