When Disciplines Diverge: An Explanatory Sequential Mixed-Method Study of Humanities and STEM College Students’ Acceptance of GenAI

As generative artificial intelligence (GenAI) increasingly reshapes language education, understanding how students from different academic disciplines engage with these technologies becomes critical. This study explores how disciplinary differences shape university students’ acceptance of GenAI for English-speaking practice. Drawing on the technology acceptance model (TAM), we employed an explanatory sequential mixed-method design with 372 survey responses and 20 follow-up interviews. Bayesian results revealed that Humanities students reported significantly higher overall acceptance and use of GenAI than STEM students. Structural equation modeling confirmed consistent TAM pathways across both groups, with PU strongly predicting IU, while PEU exerted an indirect effect. Qualitative findings suggest that evaluation criteria, epistemological beliefs, and disciplinary literacies shape how students interpret usefulness and engage with GenAI in practice. These insights extend TAM by showing disciplinary culture may act as a contextual moderator of technology acceptance, and support more discipline-sensitive designs for GenAI-mediated learning.