Construction and Practice of a Virtual-Real Integrated Practical Teaching Model Driven by AIGC and AI Agents for Higher Vocational Digital Media Education

Abstract

In digital media programs at higher vocational colleges, students usually learn tools in class but rarely carry a project from a client's brief to a delivered work. Generative AI tools and AI agents have entered professional production chains; in most curricula, though, they remain add-on demonstrations, and the virtual-real equipment in the media center is used mainly for demonstration rather than for graded assignments. This paper reports a redesign of the Photography and Videography course at Yibin Industrial Vocational and Technical College, built around a three-loop model that links online AI-assisted planning, offline multi-device shooting and delivery of industrial tasks. The course ran from March to early July 2026 with 180 first-year students in four classes. Evidence came from a pre/post survey with 176 valid responses, a catalogue of student works, and reviews by Sichuan Tourism Investment Group. Self-rated creative skills rose from 2.3 to 4.1 on a five-point scale, the share of works meeting enterprise acceptance standards rose from 52% to 89%, and 91% of students reported that they produced faster. In total the students completed more than 1,000 photographs and 102 videos. The paper describes the model, the course redesign and the empirical findings, and closes with the conditions under which the model can be transferred to comparable colleges.