English teachers in Indonesian vocational high schools are expected to teach English for Specific Purposes (ESP) with little field knowledge, material or curriculum guidance, and generative AI such as ChatGPT is quickly filling this gap. Existing studies report vocational teachers’ AI perceptions one site at a time, mostly through acceptance surveys, and none has synthesised what these teachers actually say. This literature review applied Braun and Clarke’s reflexive thematic analysis to six freely accessible empirical articles (2024–2026) reporting the views of up to 88 in-service and pre-service SMK English teachers. Four themes were constructed: AI as a stand-in for the subject specialist; AI kept behind the classroom door, used mainly for preparation under constraints of paid features, devices and school rules; checking AI output against vocational reality, sometimes through students’ workshop knowledge; and guarding creativity, integrity and moral guidance as human work. The review shows that AI currently absorbs a structural ESP problem at the level of the individual teacher. It contributes a cross-study reading of vocational teachers’ voices, identifies a verification paradox in AI-assisted ESP, and informs ESP-oriented AI literacy, collaboration with vocational-subject teachers, and institutional guidance.
Copyrights © 2026