Abstract:

The rapid evolution of Generative Artificial Intelligence (AI) is redefining professional workflows in architecture, particularly in the interpretation and application of complex technical regulations. This study reports the results of a teaching innovation project on AI in architectural education, specifically developed within the course Conditioning and Services 2 for fourth-year students in the Bachelor’s Degree in Architecture of the University of Zaragoza (Spain) in 2025/26. The primary objective is to transform students from passive recipients of technology into critical, informed users capable of leveraging AI to navigate technical building codes and design service installations.

The pedagogical strategy was structured into three distinct phases totaling 6 hours (2 theory, 4 practice):
1. Foundational Knowledge and Ethics: Students were introduced to the mathematical and statistical nature of Large Language Models (LLMs), moving away from the "magic" perception of AI. Crucially, this phase also addressed the environmental impact of AI, highlighting the carbon footprint and water consumption of data centers to promote sustainable and responsible usage.
2. Tool Exploration (NotebookLM and IAStudio):
o NotebookLM: Students used this tool to upload mandatory technical regulations (PDF documents). By grounding the AI solely on provided documents, students learned to extract precise technical information and understand complex texts through source-based queries.
o AIStudio: Students with no prior programming knowledge designed custom applications by describing functional requirements in natural language. This was specifically applied to the calculation of plumbing systems for residential buildings.
3. Collaborative Application: In the final phase, student groups autonomously integrated these tools into their design projects. They were required to identify specific challenges, apply the chosen AI tool, and critically present the results, reflecting on both the benefits and the limitations of the approach

Preliminary Results and Conclusions:
The experience suggests that AI can significantly reduce barriers to accessing and interpreting technical knowledge. Students particularly valued the ability to create personalized tools rather than relying exclusively on commercial software. The results also highlight the importance of educating students to critically evaluate AI-generated outputs and understand their limitations. As AI technologies evolve, architectural education must adapt to prepare future professionals to manage increasing technical complexity while maintaining responsible and sustainable technical practices.

Keywords:

Artificial Intelligence, Generative Artificial Intelligence, architecture.