Reflection

Reflection

This workflow exercise helped me understand how different AI tools can be integrated into a structured architectural design process. Rather than using AI as a single tool, I learned how multiple AI systems can work together to support research, concept development, visualization, analysis, and presentation. The workflow demonstrated how AI can improve efficiency while still requiring human judgment at every stage.

One aspect that worked well was the speed at which information and design ideas could be generated. Perplexity AI provided useful research on sustainable library design principles, while ChatGPT helped transform the research into a clear architectural concept statement. AI image-generation tools then allowed me to visualize the concept quickly and explore different architectural possibilities. This significantly reduced the time required for early-stage design exploration compared to traditional methods.

The most useful tool in this workflow was ChatGPT because it connected the different stages of the process. It helped organize research findings, generate design concepts, improve prompts, and support reflection. The image-generation tools were also valuable because they transformed abstract ideas into visual representations that were easier to evaluate and communicate.

However, not everything worked perfectly. Some AI-generated images contained unrealistic building details, construction elements, or spatial arrangements. In some cases, the outputs looked visually attractive but were not fully practical or technically feasible. This highlighted one of the main limitations of AI: it can generate possibilities but cannot always determine whether a design is appropriate for a specific site, community, or real-world condition.

Human judgment was required throughout the workflow. I had to evaluate the research, select relevant information, refine prompts, compare outputs, and decide which design ideas best reflected the project goals. The final concept was not produced by AI alone but through a combination of AI assistance and architectural decision-making.

Overall, this exercise showed that AI can be a powerful support tool in architectural practice. It improves productivity, encourages creativity, and helps organize complex workflows. However, architects remain responsible for critical thinking, ethical considerations, and final design decisions. The experience reinforced the idea that AI should assist the workflow, while architects remain responsible for creating meaningful, functional, and sustainable built environments.

Reflection Questions Summary

What worked well?

• Fast research, idea generation, and visualization.

• Improved workflow efficiency.

What failed?

• Some outputs were unrealistic or lacked technical accuracy.

Most useful tool?

• ChatGPT, because it supported multiple workflow stages.

Where was human judgment required?

• Evaluating outputs, refining prompts, and making final design decisions.

How did the workflow improve efficiency?

• Reduced the time needed for research, concept development, and visualization.

Risks of depending too much on AI?

• Inaccurate information, unrealistic designs, reduced creativity, and overreliance on automated suggestions.

Overall Reason

These tools were selected because each supports a different stage of the architectural workflow. Together, they improve efficiency, enhance creativity, provide data-driven insights, and help communicate design ideas, while human judgment ensures that the final outcome remains practical, ethical, and architecturally meaningful.