From resource constraints to AI-rich learning: a UTAUT–SDT insight into Tibetan L3 learners’ AI adoption in higher education

AI adoption in education has attracted growing scholarly attention over the past decade. However, the specific mechanisms underlying AI adoption in multilingual educational contexts remain underexplored. To address this issue, this study drew on semi-structured interviews with Chinese ethnic minority students learning English as a third language (L3) to examine their adoption of AI-assisted English learning through an integrated UTAUT–SDT lens. Thematic analysis showed that, for minority L3 learners, facilitating conditions were not perceived as direct enablers of AI adoption, given the prior insufficient digital learning resource constrained their autonomous decision making in learning. Social influence manifests itself through co-ethnic peer affiliation and supportive teacher recognition, reinforcing learners’ sense of relatedness. Tibetan L3 learners perceived competence restoration, indicating performance expectancy, as AI makes English learning more accessible, manageable and academically supportive. At the same time, driven by necessity to adopt AI, L3 learners have been exposed to an increased level of vulnerability, including concerns about over-reliance, weakened critical awareness and risks to their multilingual cognitive development. These findings frame AI adoption in multilingual contexts as a complex socio-motivational process and offer implications for supporting minority students’ AI-assisted English learning.