Abstract:

The rapid diffusion of generative artificial intelligence (GenAI) is profoundly transforming architectural education, particularly within design studios where representation, spatial reasoning, and critical thinking are central learning outcomes. This paper presents the preliminary results of the Erasmus+ KA220-HED project, The Use of AI Tools in Interior Design to Create Aesthetic, Inclusive and Sustainable Built Environment, which investigates how AI image generators can be meaningfully and responsibly integrated into architectural curricula.

Recent studies emphasise that AI is redefining studio culture and the relationship between architect and tool, shifting from manual representation toward prompt-based, language-driven design mediation. Text-to-image systems are increasingly used by students in early ideation phases, enhancing visual exploration and accelerating iterative processes. However, scholars also warn that without critical frameworks, AI risks reinforcing aesthetic homogenisation, weakening spatial literacy, and obscuring material and functional logic. Architectural education must therefore move beyond passive adoption and instead develop structured pedagogical models for AI integration.

The ongoing research (2024–2027) adopts an experimental didactic approach embedded in interior architecture studios. The methodology includes:
(1) comparative testing of selected text-to-image generators;
(2) parallel production of mood boards, drawings, and perspective visualisations by students and AI systems;
(3) qualitative evaluation based on conceptual coherence, spatial credibility, and functional consistency; and
(4) perception surveys involving architecture students and non-expert users.

International workshops have produced pilot inclusive interior projects, a shared knowledge hub, and a structured catalogue of prompts and diffusion-based pipelines designed as transferable teaching tools.

Preliminary findings indicate that GenAI tools are particularly effective in supporting the conceptual and atmospheric phases of studio work. They foster rapid iteration, expand the range of formal hypotheses, and encourage students to articulate design intentions with greater linguistic precision. In this sense, AI operates as a catalyst for reflective learning, consistent with experiential and dialogic models of architectural pedagogy. At the same time, significant limitations persist in geometric accuracy, material detailing, and functional logic, highlighting the continued necessity of architectural judgment and technical competence.

From a didactic perspective, the research demonstrates that the educational value of AI does not lie in replacing traditional skills, but in reconfiguring them. The integration of GenAI into studio teaching can strengthen critical evaluation, conceptual clarity, and awareness of the relationship between representation and built space—if it is accompanied by explicit methodological guidance. The next phases of the project will refine controlled generative workflows, expand case studies, and develop guidelines for embedding AI within architecture and interior design programmes. The aim is to define a pedagogical framework in which AI enhances, rather than dilutes, architectural thinking and professional responsibility.

Keywords:

AI, Pedagogy, Architecture.