Defining, Operationalizing, and Measuring Generative AI Competence among Vocational Teachers: A Scoping Review
TL;DR
This review charted how generative AI competence among vocational teachers has been defined, operationalized, and measured in the Scopus Q1 indexed literature published between 2021 and 2026 and suggested that generative AI competence for vocational teachers remains comparatively under-theorized and under measured next to its counterpart in general education.
Abstract
Generative artificial intelligence (GenAI) tools such as ChatGPT have spread quickly enough that teachers are now widely expected to develop competence in applying these technologies to instruction. Technical and vocational education and training (TVET) faces this pressure with particular force, since its teachers must simultaneously satisfy pedagogical demands, keep pace with rapidly changing technology, and stay aligned with industry practice. Research on "generative AI competence," however, has not kept these demands in a coherent frame: overlapping labels among them AI literacy, AI self-efficacy, and AI TPACK are applied inconsistently across studies, leaving the construct conceptually fragmented. Using an established scoping-review framework and reporting the process per PRISMA ScR guidance, this review charted how generative AI competence among vocational teachers has been defined, operationalized, and measured in the Scopus Q1 indexed literature published between 2021 and 2026. A structured search of five databases plus citation tracking sources identified 222 records, 28 of which all Scopus Q1 indexed satisfied the eligibility criteria after screening. Definitions in this set span a wide range, from narrowly tool specific notions such as prompt engineering competence to broad, multidimensional constructs embedded in knowledge based frameworks such as Intelligent TPACK. Operationalization efforts cluster around adaptations of three frameworks the UNESCO AI Competency Framework for Teachers, DigCompEdu, and various TPACK derived models while measurement has relied almost exclusively on self report Likert instruments; only a single instrument has been validated specifically on a vocational-teacher sample. Taken together, these patterns indicate that generative AI competence for vocational teachers remains comparatively under-theorized and under measured next to its counterpart in general education. To help close that gap, the review proposes a three layer integrative model encompassing foundational AI literacy, pedagogical technical integration, and an industry aligned contextual layer intended to guide future instrument development and the design of professional training.