Abstract:

The rapid integration of Generative Artificial Intelligence (GAI) into professional engineering practice poses new challenges for higher education. While AI systems can generate technically structured outputs, their pedagogical value depends on how they are integrated into disciplinary learning processes. This study presents a multi-course educational experience designed to critically embed GAI within Chemical Engineering education at the University of La Laguna (ULL).

The intervention was implemented in two second-year undergraduate courses (Thermal Engineering and Fluid Mechanics; 60 students each) and in the Master’s course “Analysis and Design of Chemical Processes” (17 students). In the undergraduate courses, students solved classical engineering problems analytically and subsequently submitted the same problems to a GAI system. They documented their initial prompts, evaluated the AI-generated solutions in terms of numerical accuracy, explicit assumptions, logical coherence, and undeclared simplifications, and then refined their prompts to enforce step-by-step reasoning and hypothesis justification. Comparative analysis allowed students to identify epistemic limitations of AI in structured physical modeling.

In the Master’s course, the approach was applied to a capstone activity titled “Integrated Biorefinery Design Project.” Students were initially provided with predefined objectives, competencies, project scope, and assigned AI tools. They first developed the project using conventional research and engineering design methodologies. Subsequently, they replicated the process using GAI support. Final presentations included a collective reflection session analyzing differences between traditional and AI-assisted approaches in terms of rigor, creativity, efficiency, and validation requirements.

Across both levels, the activity repositioned AI from a solution generator to a critical learning partner. Preliminary observations indicate that while GAI facilitated structuring and synthesis tasks, it frequently required domain-based validation and hypothesis clarification. The iterative prompting process in undergraduate courses improved reasoning transparency, while the Master’s experience highlighted the importance of disciplinary expertise in complex system design.

This multi-level implementation suggests that structured and reflective integration of Generative AI can enhance critical engineering thinking without compromising analytical rigor. The model demonstrates scalability across courses and academic levels, offering a transferable framework for AI-enhanced STEM education grounded in active learning and professional simulation.

Acknowledgment:
This work was supported by the University of La Laguna (ULL) through the Vice-Rectorate under the “Convocatoria de Proyectos de Innovación y Transferencia Educativa 2025”, and by the “Convocatoria de Proyectos de Innovación Educativa 2025 Interuniversitarios” through the project “SimuLab-IA: Improving Collaborative Learning in Laboratory Environments through Simulation and Artificial Intelligence”, for the 2025–2026 and 2026–2027 academic years.

Keywords:

Generative Artificial Intelligence, Chemical Engineering Education, Active Learning, Critical Thinking, STEM Innovation, Project-Based Learning.