2026 ASEE Annual Conference & Exposition

The use of generative artificial intelligence (AI) tools has become increasingly common in many fields of work, including engineering. Engineering students are likely aware of these new tools and may be using them to different degrees for their college coursework. In this study, the hypothesis was that engineering students have become familiar with free tools like ChatGPT and Gemini, but they do not fully comprehend how to best use the tools and interpret the results. Therefore, this study was designed to answer two research questions: (i) Can students correctly solve typical structural engineering homework questions with only generative AI tools? (ii) Are students comfortable with and do they have confidence in using these tools to solve structural engineering problems?

Several homework-style technical questions and perception surveys were given to students at two universities in four typical structural engineering courses (statics, mechanics of materials, structural analysis, and reinforced concrete). Students were tasked with solving the homework-style technical questions twice, first by hand and then only using generative AI tools, with approximately one week between attempts. Before and after perception surveys were deployed to collect data related to the students’ experiences completing these assignments and the tools they chose to use.

The results indicated that students were aware of AI tools (particularly popular versions such as ChatGPT, Google Gemini, and Microsoft Copilot) but less than 50% currently used them in technical engineering coursework. For those that have used the tools, the primary use was an assistant to learn the material. After completing the exercises in this study, more students were interested in using the tools to check their work, but they still showed a hesitance in full adoption of generative AI tools to solve technical problems. In this study, on 75% of the technical questions, students arrived at the correct answer using generative AI tools more often than without the tools. The biggest issue encountered by students using the generative AI tools was related to solving problems with visual pictures, like shear and moment diagrams. There was a gap in students’ trust in the solutions and ability to use engineering judgement to check the results, particularly among students in second-year courses. Those in fourth-year design courses checked their work often, typically using handwritten work or engineering judgement. Overall, students had a neutral view (compared to definitely or definitely not) of using these tools moving forward in school and in a job, but students who had used the AI tools during an internship or job reported high confidence in their results. Overall, it was determined that there is a need to discuss the uses of AI tools with students, continue to provide them with a solid foundation of fundamental concepts, and specifically teach them how to apply engineering judgement in the context of interpreting generative AI output, particularly in lower-division statics and mechanics classes.