Throughout the development of The Light Sequence Retreat, artificial intelligence played an important role in supporting analysis, design exploration, visualization, and presentation. However, the final proposal demonstrates that AI functioned as a collaborator rather than a substitute for architectural judgement. The design process highlighted both the strengths and limitations of AI and reinforced the importance of critical thinking and professional responsibility within architectural practice.
AI significantly accelerated the design process by assisting with research, analysis, concept development, visualization, and presentation. ChatGPT was used to critically evaluate the original studio project, identify weaknesses, and suggest opportunities for improvement. It helped examine issues related to circulation, daylighting, ventilation, spatial transitions, and environmental performance. This analytical process provided a structured framework for revisiting the original proposal.
Claude AI contributed to the conceptual development of the project. It assisted in generating narratives and spatial ideas that eventually led to the concept of The Light Sequence Retreat, where a sequence of colour, light, and emotional transitions guides the user from creativity towards recovery and reflection. AI also proposed interventions such as transition pockets, the semi-open social terrace, and strategies for environmental performance.
DALL·E was used primarily for visualization. It enabled the exploration of perspectives, atmosphere, sections, elevations, and presentation boards. AI-generated imagery made it easier to communicate spatial qualities and helped visualize ideas that would have been difficult and time-consuming to represent manually. Canva AI further assisted in organizing information and producing presentation boards and diagrams.
Overall, AI enhanced productivity, accelerated iteration, and provided a wider range of possibilities that enriched the design process.
Despite its advantages, AI could not replace architectural understanding and professional judgement. AI often generated images and layouts that ignored the actual dimensions, spatial relationships, and site boundaries of the original project. Several outputs created entirely new buildings rather than improving the existing proposal. The sections, elevations, and plans produced by AI frequently contradicted the original drawings and required extensive correction.
AI was also unable to understand emotional intentions, contextual priorities, and practical constraints with complete accuracy. While it generated numerous ideas, it could not determine which solutions were appropriate, feasible, or architecturally meaningful. Decisions regarding proportions, spatial organization, environmental performance, and user experience still depended on human evaluation.
Most importantly, AI lacked an understanding of the original design intentions and often prioritized visual appeal over architectural coherence.
Several limitations became apparent during the process. One of the most significant challenges was maintaining consistency between drawings and visualizations. AI repeatedly altered the site, building form, and spatial arrangement, requiring detailed prompts and multiple iterations to achieve acceptable results.
Another limitation was the accuracy of technical drawings. Sections and elevations produced by AI did not correspond precisely with the original plans and therefore had to be manually checked and revised. AI also struggled to preserve the exact relationship between the mangrove landscape, sea orientation, and spatial sequence.
The quality of results depended heavily on prompt writing. Considerable time was spent refining instructions and correcting outputs rather than accepting the first generated result. Consequently, AI accelerated some stages of the project while introducing new forms of trial and error.
The use of AI raises important ethical questions concerning originality, authorship, and transparency. AI-generated content should not be presented as entirely independent work without acknowledging the role of the tools used. Architectural education requires honesty regarding how ideas are developed and communicated.
There is also a risk that excessive reliance on AI may reduce critical thinking and encourage superficial solutions driven by aesthetics rather than function and context. AI-generated images can create unrealistic expectations and may present designs that are difficult or impossible to construct. Therefore, architects must ensure that visual outputs remain grounded in reality and supported by technical knowledge.
Furthermore, AI systems are trained on large amounts of existing information, raising concerns regarding intellectual property and the ownership of generated content. Designers must use AI responsibly and treat it as an aid rather than a replacement for creativity and expertise.
Architects are responsible for ensuring that designs are functional, safe, sustainable, and responsive to users and context. These responsibilities cannot be delegated to AI. Environmental performance, structural feasibility, material selection, and user well-being require professional knowledge and accountability.
During this project, every AI-generated proposal was critically evaluated before being incorporated into the final design. Ideas were accepted, modified, or rejected based on architectural principles rather than on visual quality alone. This process reinforced the importance of maintaining professional standards and ensuring that technology serves the design rather than controlling it.
The most important lesson learned from this project is that architecture remains fundamentally a human discipline. AI contributed speed, alternative ideas, and visual communication, but the final decisions were made through human judgement. The concept, spatial sequence, environmental strategies, and overall identity of The Light Sequence Retreat emerged through continuous evaluation and refinement rather than automatic generation.
Human judgement determined which ideas aligned with Andy Warhol's story, which environmental strategies were appropriate for Port Dickson, and how the existing design could evolve without losing its original character. AI acted as a creative partner, but architectural thinking, critical reflection, and responsibility remained with the designer.
Ultimately, this experience demonstrated that the future of architecture is not a competition between humans and AI, but a collaboration in which technology supports creativity while human judgement provides meaning, responsibility, and design integrity.