As AIGC technologies become deeply integrated into higher education teaching and learning, a critical challenge has quickly emerged: many pedagogical approaches remain constrained by traditional linear thinking, particularly evident in the persistence of unidimensional instructional models. To address this impasse, this study introduces quantum cognition theory to construct a human–AI collaborative creative model. The model operates through three phases: latent state generation, observational collapse, and iterative evolution. An empirical investigation was conducted with three cohorts of students from the AIGC Film and Television Production Innovation Program at Sichuan Film and Television College. The findings indicate that under the influence of quantum cognition, a nonlinear pedagogical paradigm significantly enhances students’ cross-path exploration capabilities. Moreover, human–AI co-created works demonstrated notable “emergence” in provincial- and national-level competitions. This research substantiates how this approach reshapes cognitive pathways and teaching evaluation systems in sound art, offering theoretical grounding and practical pathways for the paradigm reconstruction of arts education in the era of artificial intelligence.