Virtual Coworker and Intelligent Tutor: Research on AI Agent-Driven Innovation in Contextualized Teaching Models for Higher Vocational Education

Abstract

Higher vocational education has long grappled with the structural dilemma of the schism between the “school field” and the “work field.” Traditional skills training often prioritizes the replication of operational procedures while neglecting the complexity of social interactions inherent in professional scenarios. With the maturation of AI Agent technology empowered by Large Language Models (LLMs), a digital subject capable of simulating autonomous behavior, memory, and planning has become feasible. Grounded in social constructivism—specifically Situated Learning and Cognitive Apprenticeship theories—this paper proposes a dual-agent pedagogical model comprising a “Virtual Coworker” and an “Intelligent Tutor.” This model aims to re-”contextualize” the learner’s cognitive process by constructing a high-fidelity network of professional social relations. Through horizontal collaborative interaction and vertical expert guidance, it facilitates the transformation from mere knowledge acquisition to the generation of professional competency.