Reflection

What Did You Learn About AI?

Through exploring ChatGPT, Gemini, and Perplexity, I learned that different AI tools can support different stages of the architectural design process. ChatGPT was useful for generating conceptual ideas and organizing architectural thinking into structured design proposals. Gemini helped transform written prompts into realistic visual renderings, making design ideas easier to communicate visually. Perplexity was effective for research because it provided detailed explanations, technical information, and real-world examples with references. I also learned that the quality of AI outputs depends strongly on the clarity and detail of the prompts provided.

⸻

What Surprised You?

What surprised me most was how quickly AI could generate detailed ideas, realistic visuals, and technical research from a simple prompt. The tools were able to combine sustainability, community-focused design, and architectural language in a professional way. I was also surprised by how each AI tool had different strengths — some were better for creativity and visualisation, while others were stronger in research and information gathering.

⸻

What Are the Limitations?

Although the AI outputs were impressive, there were several limitations. Many responses were still general and lacked site-specific context such as climate data, building regulations, budget, and user needs. AI also cannot fully replace human creativity, architectural judgment, or technical expertise. Some generated ideas may not be practical or buildable without further development and validation. Additionally, the results depended heavily on how well the prompts were written, meaning unclear prompts often produced weaker outputs.