Voices from the margins: EFL lecturers’ responses to AI integration in regional Indonesia

Artificial intelligence is transforming education worldwide, yet its application in EFL classrooms, particularly in regional Indonesian universities, remains underexplored. This study examined how EFL lecturers perceived, practiced, and experienced challenges in integrating AI at four universities in Bengkulu, Indonesia. Guided by the Technology Acceptance Model (TAM) and the concept of Teacher Agency, the study explored how lecturers’ perceptions, beliefs, and professional autonomy influenced their engagement with AI tools. A mixed-methods design was employed, combining survey data from 35 lecturers and interviews with 10 participants to capture both broad patterns and contextual experiences. Quantitative data were analyzed using descriptive and correlational statistics in SPSS, while qualitative data were analyzed thematically. The results showed that 74% of lecturers had used AI tools, mainly for lesson preparation, grading, and feedback. Positive attitudes toward AI were strongly associated with pedagogical innovation (r = 0.68, p < .01), reflecting the TAM dimensions of perceived usefulness and behavioral intention. The qualitative findings revealed three themes: enhanced student engagement through interactive AI activities, evolving lecturer roles as facilitators and evaluators, and increased ethical awareness regarding data privacy and academic integrity. These themes also reflected lecturers’ agency in adapting AI use to institutional constraints and learner needs. This study fills a gap in understanding AI integration in regional EFL contexts and extends the application of TAM and Teacher Agency to language education. The findings indicate that sustainable AI adoption depends on positive perceptions, professional agency, and institutional support that encourage reflective and ethical teaching practices.