2025 ASEE Annual Conference & Exposition

The generation of engineering problems is an essential step in an effective problem-solving process. However, interpreting it is difficult due to differences in individuals’ knowledge and expertise. This step stimulates the recognition of problematic system components and eliminates them during the solving process. Additionally, the generation of problems is a crucial stage in the engineering design process, which directly impacts the final design. A poor approach to problem generation may lead to an ambiguous or incorrect problem solution. Despite its significance in pedagogy, the problem formulation is often overlooked particularly, in the context of traditional engineering textbook-based problems, which are well-defined but lack the complexity of real-world problems. This gap restricts students’ experience to the uncertainty and complexity present in real-world problems. This study uncovers the potential of using generative AI (Gen-AI) to redefine engineering problems and overcome their limitations. Chat-GPT, a user-friendly Gen-AI, is utilized as a problem-generation tool. The study hypothesizes that by integrating AI-generated problems with conventional settings, students will gain a deeper understanding of engineering content and improve their performance. It adopts a mixed-method approach with 14 participants to investigate this hypothesis. Students' performance is evaluated using structured analytical rubrics and a deductive coding scheme. Additionally, a coding scheme designed by Grigg and Benson is used to analyze students' responses. The study identifies problem-processing elements, error execution, and solution accuracy as the major experiences of students. Ultimately, the study concludes that there are significant impacts of the generation of problems on student performance compared with conventional textbook problems. The insights of this research offer valuable guidance for redefining traditional engineering problems.