Teacher support enhances self-efficacy and learning outcomes in the age of AI

With the increasing acceptance and adoption of generative AI in higher education, it remains unclear how this integration can improve the learning outcomes of college students. This study aims to gather empirical evidence on the impact of teacher support and student academic self-efficacy on learning outcomes in EFL (English as a Foreign Language) learning, utilizing generative AI. This quantitative study is based on the Social Cognitive Theory to assess students' attitudes, intentions, and behaviors toward employing generative AI in EFL learning. A questionnaire measuring teachers' affective support, teachers' capacity support, teachers' behavior support, students' academic self-efficacy, and learning outcomes was completed by 906 EFL students in a Chinese higher vocational college. A structural equation modelling study reveals that teachers' capacity support and behavior support can directly predict learners' academic self-efficacy and indirectly affect the students' learning outcomes. Meanwhile, higher levels of students' academic self-efficacy are associated with better academic learning outcomes. This study demonstrates the effectiveness of generative AI as a language-learning tool for EFL learners. Theoretically, it proves that the supportive environment and learners' positive internal psychological factors combine to produce improved learning achievement. Practically, students can make improvements in their English learning with more support from teachers and enhancement of their academic self-efficacy. Consequently, there should be teacher training guiding how to support students emotionally, skillfully, and practically in their English learning, whereas AI-specific pedagogy needs to be further explored to help students enhance their self-efficacy to achieve better learning outcomes.