Acceptance of AI-assisted English language learning tools in higher education: psychological correlates across disciplinary and proficiency groups

Artificial intelligence (AI)-assisted language learning tools are increasingly used in higher education, yet student acceptance of these tools may vary across academic disciplines and English proficiency levels. Building on the Technology Acceptance Model (TAM) as a baseline framework, this study examined how learning motivation, self-efficacy, anxiety, and risk perception were associated with acceptance outcomes for AI-assisted English learning in a Chinese higher education context. The study focused on perceived usefulness, perceived ease of use, behavioral intention, and satisfaction, and further examined mean-level group differences and descriptive subgroup-specific association patterns by academic discipline and English proficiency. Survey data were collected from 210 undergraduates (STEM = 91, Humanities = 119; English proficiency: Low = 77, Intermediate = 103, High = 30). Data analyses were conducted using IBM SPSS Statistics 27.0 and Stata 18.0, including descriptive statistics, reliability and convergent validity analyses, full-sample adjusted linear regression models using robust standard errors, mean-level group comparisons, post hoc comparisons, and exploratory subgroup-specific regression analyses. Learning motivation and self-efficacy were consistently and positively associated with acceptance-related indicators. Anxiety did not show a uniformly negative pattern; its positive associations with acceptance outcomes were exploratory concurrent patterns and should not be interpreted as evidence that anxiety is beneficial. Risk perception was also positively but generally more weakly associated with acceptance outcomes and should be interpreted cautiously. Mean-level comparisons indicated descriptive differences in several acceptance outcomes across academic discipline and English proficiency groups. Exploratory subgroup-specific regressions further described within-group association patterns; these analyses were not formal tests of between-group differences in regression coefficients. Given the cross-sectional design, all findings should be interpreted as associative rather than causal. Overall, the study provides a learner-centered account of AI-assisted English learning acceptance by highlighting psychological correlates and group-related descriptive patterns in a Chinese higher education context.