2026 ASEE Annual Conference & Exposition

This empirical full paper explores the influence of Generative AI on engineering students’ learning experiences. Engineering education is rapidly integrating advanced tools including Generative AI (GenAI) which is based on Large Language Model (LLM) trained to generate text with reasoning capabilities. GenAI adoption is growing notably and so is the research in this field. But there remains a gap in understanding the specific lived experiences of engineering students using these tools. This qualitative study aims to fill this gap by addressing the research question of How do engineering students at different curriculum levels perceive GenAI-assisted learning as support for their academic needs? This study employed a targeted recruitment strategy to select three students who use GenAI as a learning assistant and represent different levels and disciplines of engineering curricula. The three participants are a Data Science graduate student, a junior Mechanical Engineering student, and a sophomore Environmental Engineering student. This multiple-case approach provides initial in-depth insights into GenAI's influence across different engineering disciplines and experience levels. Data was collected in Fall 2024 through ~60-minute semi-structured online interviews and analyzed using systematic open coding techniques. Although not generalizable, findings viewed through the lens of Self-Determination Theory provide critical insights for future research. We found that students used GenAI for specific tasks (e.g., real-time code debugging, brainstorming project ideas, clarifying complex foundational concepts among others). They perceived them as directly supporting their needs for Autonomy (by controlling the pace and providing a non-judgmental space for help-seeking) and Competence (by overcoming immediate obstacles). Furthermore, one student used GenAI to bridge the gap of instructor’s availability, demonstrating that students leverage the tool for relatedness and institutional support. Students were not passive users of GenAI. They demonstrated active Self-Regulated Learning strategies in their experience with GenAI like cross-checking and directive prompting to manage tool-specific risks like inaccuracy and dependency. The study also articulated challenges regarding hallucinations and the systemic fear of academic surveillance. The paper also outlines practical strategies for incorporating GenAI into engineering programs. The insights gained into student usage lay essential groundwork for more extensive future studies.