Why students use or not use generative AI: Student conceptions, concerns, and implications for engineering education

Generative artificial intelligence (GenAI) technologies is believed to transform engineering education. However, it remains underexamined how engineering students choose to use GenAI or not, along with the reasons behind their choices. To fill in this research gap, this study presents a natural experiment that examines student use or non-use of GenAI tools in engineering design tasks in an undergraduate course. In this experiment, the participants (n=403) were provided with an unconstrained access to an GPT 4.0-empowered chatbot and were allowed to use it for their design projects voluntarily. Overall, 59.80% of the students had reported substantial use of GenAI in their design projects, and 40.20% showed limited or no use. Those adopters used GenAI to aid idea generation and brainstorming, to mediate discussions with instructors/TA, to overcome non-technical expertise gaps, and to optimize the solution. Conversely, non-adopters attributed their reluctance and rejection to inherent limitations in GenAI outputs, misalignment between GenAI functionalities and project needs, a lack of adaptation and prompt skills, and unclear benefits of GenAI use for personal development. This study has challenged the popular assumption of naturally active GenAI adoption by uncovering the complexity and multiplicities in student behaviors. The findings highlight the necessity of establish a consensus on the role and value of GenAI across various stakeholders, while suggesting a need for adaptation in engineering education research and practices.