La imagen proyectual: promptografías y agencia visual en el diseño

AI-generated images produced by generative artificial intelligence (GAI) appear to be altering the conditions under which ideas emerge in the design process. However, research on their cognitive effects during ideation remains fragmented and, at times, contradictory. This article does not present original empirical research; rather, its contribution is based on a critical and interpretive review of recent academic literature (2023-2025) on generative images, visual agency, and ideation in design and architecture. These images are examined as external supports of design thinking, in continuity with the graphic representations studied by Schön (1983) and Cross (2006), and read through the lens of distributed and extended cognition (Hutchins, 1995; Clark and Chalmers, 1998). The corpus was selected according to its relevance to three key issues: the cognitive function of GAI-generated images as supports for formal exploration, lateral thinking, and metacognition; the role of unexpected variations and “creative errors” as catalysts, or as sources of visual fixation, in design thinking; and the redistribution of creative agency imposed by these systems, shifting the act of design from execution toward selection, curation, and the critical reading of algorithmic output. The concept of the projective image is proposed to designate those generative images that do not translate a prior idea but instead constitute the site where the idea emerges. The review shows that GAI simultaneously produces thematic expansion and stylistic convergence, and that its effect on creativity depends more on the level of critical judgment exercised by the designer over the output than on the degree of intervention. The article argues that the productive integration of GAI into the design process requires an active critical stance toward the agency of the generated image.