Abstract
In Norway, about half of upper secondary students pursue vocational education and training (VET), yet dropout rates remain high – especially among boys from migrant backgrounds. This study investigates how large language models (LLMs), such as ChatGPT, may serve as tools to support learning in VET contexts characterized by limited teacher capacity and diverse student needs. Set in a vocational school with high dropout rates, the research explores students’ experiences with LLMs as supplementary learning support. Following the release of ChatGPT in 2022, concerns about academic dishonesty have intensified. However, this study emphasizes the pedagogical potential of LLMs to provide timely, individualized assistance. The project is grounded in sociocultural learning theories, particularly Vygotsky’s Zone of Proximal Development and Lave and Wenger’s theory of situated learning. These perspectives position the LLM as a mediating cognitive tool that supports participation in communities of practice. Billett’s and Eraut’s work on informal learning further informs the understanding of AI-supported scaffolding. This research article builds on the first author’s master’s project, which employed an action research design. Empirical data for this article consists of interviews with ten VET students (age 16-18 yrs) who participated in the master’s project. Thematic analysis (Braun & Clarke) is being used to explore the material. Preliminary findings suggest that students view LLMs positively. They report fewer interruptions in their work, greater independence, and increased confidence when encountering technical or cultural complexity. The study aims to examine how such technologies can foster student engagement and supplement traditional instruction in vocational education.
Presenters
Daniele MattioliTeacher, Technology and Industry, Stovner Videregående Skole, Oslo, Norway Arnfinn Gilberg
PhD Candidate, Vocational Teacher Education, Oslo Metropolitan University, Oslo, Norway
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
KEYWORDS
Vocational education, Generative AI, Learning support, Student engagement, Sociocultural theory
