CRITICAL REFLECTION

Throughout this project, AI played an important role in helping me revisit and improve my original Studio Design 3A proposal, NEXUS Cultural Centre. Instead of using AI only to generate images, I used it as a tool to analyse my design, identify weaknesses, explore alternatives, and communicate ideas more effectively. Tools such as ChatGPT, Gemini, Leonardo AI, and Canva AI helped me throughout different stages of the redesign process.

One of the biggest contributions of AI was helping me critically evaluate my original design. The original NEXUS proposal successfully communicated the idea of a wormhole connecting modern Kuala Lumpur and traditional Kampung Baru, but after analysing the project with AI, I realised that the concept was more symbolic than functional. AI highlighted issues such as unclear circulation between spaces and the lack of a strong relationship between the wormhole concept and the actual user experience. This encouraged me to rethink the design and transform the wormhole into a circulation spine, environmental feature, and social gathering space. As a result, I reorganised the building into three clear levels: Modernity, Connection, and Tradition.

AI also helped me explore different design possibilities quickly. By generating multiple diagrams, renderings, and design options, I was able to compare different approaches before deciding which ideas worked best for the project. It helped me visualise how the rainwater harvesting system, waterfall feature, circulation routes, and community spaces could be integrated into the overall concept. This made the redesign process more efficient and allowed me to test ideas that I may not have considered initially.

However, I also discovered several limitations when using AI. One of the biggest challenges was that AI-generated drawings often lacked consistency in scale and proportion. For example, when generating plans, sections, or perspectives, the sizes of rooms, floor heights, columns, and circulation spaces would often change between different outputs. While the images looked visually impressive, they were not always architecturally accurate. This meant that I could not rely on AI outputs as final technical drawings and had to use my own judgement to determine what was realistic and buildable.

Another limitation was AI's inability to create professional technical drawings from scratch. Although it could generate floor plans and sections that appeared convincing at first glance, many of them contained unrealistic structures, incorrect dimensions, or impractical layouts. I found that AI performed much better when it was given an existing drawing or design as a reference. It was useful for enhancing, visualising, or refining a design, but not for producing a complete architectural drawing that could be used for construction documentation. This highlighted the importance of architectural knowledge and technical understanding when working with AI tools.

There were also ethical considerations during the project. AI can generate a large amount of content very quickly, which may encourage users to accept outputs without questioning them. I realised that not every AI-generated solution was appropriate for my project. Some suggestions changed the identity of the design, while others ignored site conditions, cultural context, or practical architectural requirements. Because of this, I treated AI as a design assistant rather than a designer. Every output was reviewed, evaluated, and modified before being incorporated into the project.

Professional responsibility was another important aspect of this process. As architecture students and future professionals, we are responsible for ensuring that our designs are functional, safe, culturally sensitive, and technically feasible. AI does not fully understand local context, user needs, building regulations, or construction methods. While it can generate ideas quickly, it cannot replace the judgement required to make informed design decisions. Throughout this project, I had to decide which AI suggestions were useful and which should be rejected.

Overall, this project showed me that AI is most valuable when used as a tool to support design thinking rather than replace it. It helped me analyse my original proposal, generate new ideas, and improve the overall quality of the project. At the same time, it revealed the importance of human judgement, critical thinking, and architectural knowledge. The final redesign of NEXUS was not created by AI alone; it was the result of combining AI-generated insights with my own design decisions, understanding of the site, and professional judgement. This experience has shown me that the future role of AI in architecture is not to replace designers, but to support a more informed, efficient, and reflective design process.