Abstract:

The integration of generative artificial intelligence (GenAI) into architectural education calls for a redefinition of studio pedagogy, particularly in relation to creativity and design thinking. This paper presents a comparative reflection on two academic experiences: the Architecture and Environmental Design Studio at Politecnico di Milano (Prof. Luca M.F. Fabris) and the Industrial Heritage Conservation and Reuse Studio at the Beijing University of Civil Engineering and Architecture (BUCEA) led by Prof. Fanlei Meng. Both courses address the transformation of disused urban artefacts—the former Casoretto Cinema in Milan and the former Shougang Steelworks in Beijing—framing them as latent civic infrastructures whose cultural and spatial value requires critical rediscovery.

Rather than approaching adaptive reuse as a purely technical exercise, both studios position the project as a speculative inquiry into the future role of these abandoned structures within evolving urban ecosystems. In this context, AI is not treated as a generative shortcut but as a pedagogical interface mediating between observation, imagination, and verification.

In Milan, students operate within a controlled design framework grounded in measured drawings, environmental constraints, and material logic. AI is integrated through constrained image-to-image workflows that preserve geometric fidelity and spatial coherence. Here, Human–AI interaction functions as an iterative feedback loop, reinforcing typological reasoning and atmospheric calibration while maintaining architectural authorship. In Beijing, the process begins with site photography and empirical observation of the industrial landscape. AI becomes a speculative mediator, enabling the reinterpretation of infrastructural remnants as ecological and collective spaces. The industrial morphology remains legible, yet is reframed through imaginative projection.

Despite methodological differences, both pedagogies share structural principles. First, AI is embedded within cyclical design thinking processes—analysis, prototyping, evaluation, refinement—rather than displacing critical judgment. Second, students are required to assess algorithmic outputs, identifying spatial inconsistencies and semantic shifts. Third, Human–AI co-creation fosters a reflective awareness of authorship, echoing emerging debates on the transformation of architectural agency in the age of intelligent systems.

The comparison reveals complementary trajectories: in Milan, AI supports disciplined transformation anchored in architectural constraints; in Beijing, it amplifies speculative reinterpretation rooted in urban memory. Together, these cases suggest that pedagogical innovation lies not in technological adoption per se, but in structuring AI as a cognitive and creative partner. When critically framed, Human–AI collaboration strengthens design thinking, enabling students to imagine the future of disused urban heritage while reaffirming architecture’s cultural and civic responsibility.

Keywords:

GenAI, Architectural Design Teaching, Adaptive Reuse, Italy, China.