Latent Space Meandering

This chapter examines how generative AI reshapes creative cognition by altering how designers and artists think, explore ideas, and remain mentally engaged during creative work. In this study, the term ‘artist-designer’ refers to practitioners whose works bridge aesthetic inquiry and functional problem-solving recognizing that while artists and designers differ in their primary orientations (expression vs. function), they share fundamental creative processes, including ideation, iteration, and material engagement. The study participants were design students and recent graduates, but the findings speak to creative practices that cross disciplinary boundaries. Rather than treating creativity as a purely goal-driven or problem-solving activity, the research approaches mind wandering as a meaningful and productive component of creative practice, even within design's functional contexts. The study adopts an exploratory mixed-methods approach, combining qualitative process observation with descriptive self-report measures, to investigate how novel tools and process structures within AI-mediated creative environments shape attention, exploration, authorship, and emotional state. Creative arts participants took part in a controlled experimental design comparing two tool conditions, generative AI systems and traditional digital design tools, across two modes of practice: (a) structured, step-by-step processes and (b) informed by Design Thinking; and open exploratory work. Participants completed creative tasks while documenting their process and reflecting on experiences of focus, distraction, surprise, and authorship. The findings indicate that process structure exerts a stronger influence on cognitive engagement than tool type. Open exploration produced significantly higher levels of mind wandering than Design Thinking, without being experienced as loss of focus. Instead, wandering functioned as an engaged exploratory state, appearing in AI-based work as latent space meandering ( techno-flânerie ). Generative AI increased surprise and reduced transparency, while traditional digital tools supported stronger authorship and collaboration, likely due to continuous feedback during making. These effects were systematically modulated by structure: non-sequential conditions maximized authorship and collaboration, whereas sequential AI use minimized them. Overall, the results suggest that productive mind wandering happens through a deliberate balance between structure and freedom, pointing toward the need of pedagogical approaches that alternate guidance with exploratory drift across both AI and non-AI creative tools.