Critical Reflection & Professional Judgement

1. What AI Contributed to the Redesign

Utilizing generative Artificial Intelligence within the redesign workflow of The Silent Sanctuary fundamentally expanded my analytical and visual exploration boundaries. Rather than treating tools like ChatGPT or LookX as simple optimization engines, they were deployed as dialectic partners to pressure-test the architectural assumptions made during Studio 2A. The key contribution of the Large Language Model was its capacity to immediately bridge two highly polarized concepts: the heavy, uncompromising phenomenological materiality of Peter Zumthor and the hyper-specific, fluid demands of tropical passive sustainable design.

On a visual level, the text-to-image and image-to-image AI pipelines allowed for rapid, high-fidelity spatial detailing. Manually rendering how complex, rough board-formed concrete blocks scatter intense, low-lux linear light rays is an incredibly time-consuming process in traditional software. The AI workflow visualized these complex, atmospheric light physics within seconds, enabling an agile design iteration cycle where I could test how shifting opening depths directly impacted the internal spiritual ambiance.

2. Technical Limitations and What AI Could Not Contribute

Despite its speed, the generative AI tools demonstrated severe technical limitations, reinforcing the fact that AI cannot replace foundational architectural logic. The most prominent deficit encountered was the AI’s total lack of structural and spatial orientation awareness. When prompted with image-to-image scripts, the models frequently hallucinated nonsensical geometry placing decorative glazing within what were explicitly designated as solid load-bearing concrete block structures, or closing off vital wind-permeable brise-soleil layers to prioritize flat, superficial patterns.

Furthermore, the AI has no inherent comprehension of building regulations, structural load paths, or local Malaysian UBBL requirements. It can effortlessly draw a beautiful, evocative perspective of an internal concrete courtyard, but it cannot calculate whether the 1 to Northwest wind path will actually generate sufficient pressure differentials to drop internal temperatures. It treats architecture as a 2D canvas of textures and pixels, entirely divorced from gravity, thermal dynamics, or human scale.

3. Ethical Concerns and Professional Responsibility

Engaging with generative AI introduces heavy ethical questions regarding intellectual property, authorship, and the dilution of the regional identity. Because text-to-image models are trained on massive, scraped global datasets, the initial outputs generated for Dataran Malawati frequently defaulted to generic, eurocentric or hyper-futuristic architectural styles. Uncritically accepting these outputs poses an ethical risk of architectural colonization, where local contextual history, material heritage, and specific cultural requirements are completely erased by a smooth, algorithmically generalized aesthetic.

Additionally, as a professional-in-training, I must uphold the principle that copy-paste is not authorship. Relying on an algorithm to generate complete design solutions represents a direct abdication of professional responsibility. The ethical use of AI demands total transparency in tracking prompt engineering and a strict refusal to use AI to counterfeit design thinking.

4. The Indispensable Role of Human Judgement

Ultimately, architectural wisdom belongs exclusively to human judgment. An AI can simulate an atmosphere, but it cannot feel the emotional relief of escaping the chaotic, hot, monkey-infested outdoor environment of Kuala Selangor to step into a dark, silent stone sanctuary. The human architect must act as the ultimate critical filter.

It required intentional human choices to reject the over-glazed, flashy images the AI proposed, forcing the model back toward a low-key, heavy, minimalist brick aesthetic to protect the project's conceptual integrity. The human mind orchestrates the narrative, establishes the ethical boundaries, coordinates structural feasibility, and ensures that the final building serves the community's physical and spiritual well-being. AI expanded my reflective possibilities, but human intent dictated the architecture.