We examine the evolving role of human designers in the architectural field amid the rise of artificial intelligence and diffusion models, particularly through a series of design experiments conducted at the University of Antwerp from 2023 to 2025. The introduction of diffusion models, which became widely accessible with technologies like DALL-E, has led to a proliferation of design generation tools. While these models can create compelling images through various methods, such as text-to-image and sketch-to-image, the essence of architectural design remains dependent on human interpretation and critical analysis of the generated outputs. The methodology outlines five experiments aimed at refining diffusion-driven architectural workflows, emphasising the importance of human–machine interaction and the critical design skills necessary for leveraging AI effectively. The results suggest that while diffusion-generated images can aid the design process by facilitating quick iterations and ideations, the role of the designer transitions from an autonomous creator to an assisted curator-creator. The findings indicate a need for mixed-media workflows that integrate traditional design tools alongside AI technologies to achieve a holistic architectural design. The conclusions assert that despite the advancements in AI, the interpretation of design values remains a crucial task for human designers, underscoring their irreplaceable contribution to the architectural process. This recognition invites further exploration of how designers can adapt and innovate in an increasingly AI-driven landscape of architectural design.