Artificial intelligence (AI) tools are increasingly used in engineering design, yet their recommendations often require human interpretation, contextualization, and validation—especially in safety-critical domains. This paper presents a case study of a dual-zone residential heat pump installation where AI-generated electrical design recommendations were plausible but incomplete, omitting key National Electrical Code (NEC) requirements and manufacturer constraints. Through iterative human–AI dialogue, the designer identified hidden assumptions, corrected non-compliant outputs, and refined the design to ensure full code compliance. The analysis shows how develops through questioning, how human-in-the-loop oversight enhances safety and accountability, and how collaborative reasoning improves technical results. Drawing from research in AI governance, explainability, and STEM education, the paper proposes a replicable instructional model for Engineering Technology programs that uses AI-assisted design tasks to teach students how to critically examine automated recommendations, apply codes and standards, and communicate technical rationale. The case illustrates how AI can serve as a catalyst for inquiry-based learning and responsible engineering practice.
http://orcid.org/0000-0003-0053-753X
Purdue Polytechnic Institute, Purdue University – West Lafayette
[biography]
Are you a researcher? Would you like to cite this paper? Visit the ASEE document repository at peer.asee.org for more tools and easy citations.