With the widespread application of generative artificial intelligence (AIGC) in design, its impact on design conception across different user groups is a significant research issue in design education. This study explores the differences in design conception between designer users and ordinary users using AIGC for poster design, and evaluates the application of multimodal discourse theory and SEM models. The study defines four variables in the poster design concept: representational meaning, interactive meaning, compositional meaning, and color meaning. A questionnaire with 22 items was distributed to 120 respondents. SEM analysis showed a good fit and significant inter-group differences. Designers’ design ideation is influenced by all dimensions, while compositional meaning does not affect ordinary users. Designers’ path coefficients were balanced across dimensions, while ordinary users showed significant differences, particularly in representational and interactive meaning. This study verifies the applicability of multimodal discourse theory and offers insights into design conception differences in the AIGC context. It emphasizes the need to improve non-professional users’ design skills and provides guidance for developing more user-friendly AIGC tools.