Reflection

Reflection Summary

The purpose of this exercise was to explore how artificial intelligence can support architectural design workflows while maintaining the architect's role in critical thinking and decision-making. Throughout this project, several AI tools were used, including ChatGPT, Gemini AI, and Perplexity AI. Each tool contributed differently to the design process and demonstrated unique strengths and limitations.

ChatGPT was particularly useful during the early stages of concept development. It helped generate design ideas, architectural narratives, and refined prompts that provided clearer design directions. Gemini AI was effective in producing visual interpretations of the design concepts, allowing rapid exploration of different architectural possibilities. Perplexity AI supported research, information gathering, and fact verification, helping to strengthen the design process with reliable references and contextual information.

One of the most important lessons learned was that the quality of AI outputs depends heavily on the quality of prompts. Initial prompts produced generic and less detailed results. However, as more context, architectural references, environmental considerations, and design intentions were added, the outputs became significantly more relevant and meaningful. This demonstrated that prompt development is itself a form of design thinking and plays a critical role in guiding AI-generated outcomes.

The project also highlighted the limitations of AI. Although AI tools can quickly generate ideas, visuals, and research summaries, they cannot fully understand architectural values such as human experience, emotional impact, cultural context, functionality, and spatial quality. Many outputs required evaluation, refinement, and comparison before they could be used effectively. Human judgment remained essential throughout the workflow.

The final design direction was selected based on both conceptual and functional considerations rather than visual appearance alone. The chosen proposal best reflected the objective of creating a therapeutic visitor centre within a tropical rainforest environment, emphasizing emotional healing, connection with nature, natural light, and peaceful spatial experiences. This decision required careful analysis of multiple AI-generated alternatives and architectural considerations.

Overall, this exercise demonstrated that AI should be viewed as a powerful design assistant rather than a replacement for architects. The most valuable outcome was learning how different AI tools can be strategically integrated into a workflow to improve creativity, efficiency, research, and design exploration while ensuring that final decisions remain under human control. This experience has enhanced my understanding of AI-assisted architectural design and its future applications in professional practice.