Designing with Non-Human Things

The Impact of Text-to-Image Generative AI on Human Roles in the Conceptual Stage of Interior Design

Abstract

This paper examines how text-to-image Artificial Intelligence (AI) reshapes the conceptual stage of interior design by reconfiguring relations between humans and non-human technologies. Interior design is framed as an inherently image-based, interdisciplinary practice in which sketches and visualisations mediate between designers and clients. Building on philosophical accounts of humans, artefacts, and media, the paper argues that text-to-image AI should be understood not as a neutral tool but as a non-human generative collaborator that both extends and reorganises human creative capacities. Drawing on literature from design and AI, as well as practice-based experiments with AI platforms, the study traces how these systems expand design imagination through non-human perspectives while simultaneously embedding biases, probabilistic homogenisation, etc. The paper shows that text-to-image AI shifts image-making from direct manipulation in specialised software to a language-driven process in which text prompts become a primary site of design labour. This transition intensifies the linguistic demands placed on designers, who must negotiate meaning, specificity, and spatial intention with an opaque generative system. The central argument is that co-designing with non-humans does not replace human designers; instead, it repositions them as critical mediators whose interpretive, ethical, and contextual judgment remains indispensable to meaningful interior design practice.