AI helped me improve my project because it made me look at my original design in a more critical way. My original project was called “The Purity,” and it already had a strong idea based on Piet Mondrian, De Stijl, order, balance, vertical and horizontal lines, and primary colours. But after using AI, I realised that my first design was mostly focused on the look and the concept, not enough on the site and environment.
AI helped me connect the design better to Pulau Burong. It helped me think about the mangrove site, tidal movement, sea breeze, humidity, salt spray, rain, storm surge, soft ground, and mangrove roots. Because of this, the revised design became more suitable for the site. It was not only a Mondrian-style building anymore, but a retreat that responds better to nature and climate.
AI also helped me improve my presentation. It helped me test different ways to show the plan, section, elevation, material palette, site response, and diagrams. This made the final work look clearer and more professional.
AI could not replace my own thinking. It can give ideas, images, diagrams, and text, but it does not fully understand my original intention or my studio process. Sometimes AI made the work look nice, but it did not always keep the drawings accurate.
AI also could not make the final decisions for me. For example, it suggested ideas like elevated structure, passive ventilation, and mangrove protection, but I still had to decide what was right for my project. AI cannot really experience the site like a human. It cannot feel the humidity, understand the atmosphere, or judge the feeling of the space properly.
AI also could not always be trusted with technical accuracy. Sometimes it changed the plan, added rooms, changed the roof, or made details that were not realistic. Because of that, I had to keep checking everything and compare it with my original design.
The biggest limitation was accuracy. When I used AI to generate or improve images, it sometimes changed the design even when I asked it not to. It could change the floor plan, roof plan, spaces, or details. This was a problem because architecture is not only about making beautiful images. The drawings must also be correct.
Another limitation was the text. In some AI-generated posters, the text was too small, unclear, or incorrect. So I could not just use the AI result directly without checking it.
AI also struggled to keep the exact same drawings. Even when I said “do not change anything,” it sometimes redrew the drawings in a different way. This showed me that AI is helpful for ideas and presentation, but it is not always reliable for final architectural drawings.
One ethical concern was authorship. Since AI helped me with the visuals and presentation, I need to be honest that AI was part of the process. I should not pretend that everything was done fully by myself without AI support.
Another concern was overusing AI. I had to make sure the project was still mine and not completely created by AI. The final design still needed to come from my original studio work, my concept, and my own judgment.
There was also an ethical concern about the site. Pulau Burong is a sensitive mangrove ecosystem, so the design should not show careless construction or damage to nature. The project should respect the mangrove roots, tidal flow, birds, and natural environment. Nature should not just be used as a background image.
As a student designer, my responsibility is to use AI carefully. I should not accept every AI output just because it looks good. I need to question it, check it, and make sure it follows the brief, the site condition, and the design idea.
I also have a responsibility to protect the site. Since the project is in a mangrove area, the design should have a small footprint, be raised on piles, allow tidal movement, and avoid damaging the roots. The building should work with nature, not against it.
I also need to communicate my process clearly. The poster, QR card, and Google Site should show what changed, why it changed, and how AI helped me improve the project.
Human judgment was the most important part of the whole process. AI helped me test ideas, improve the site response, and organise the presentation, but I still had to choose what to keep and what to reject.
My judgment was important because I had to protect the main idea of the project. The design still needed to feel like Mondrian, with order, balance, vertical and horizontal lines, primary colours, and open flowing spaces. At the same time, it also needed to respond to Pulau Burong and the mangrove ecosystem.
The main lesson I learned is that AI does not replace architectural thinking. It is only a tool. It can make the process faster and help me explore more ideas, but I still need to slow down, think carefully, question the result, and make the final decisions myself. In this project, AI expanded my design process, but the final judgment remained mine.