Comparing and contrasting students' and faculty perceptions and utilization of GenAI

The introduction of generative open AI (GenAI), particularly the release of ChatGPT in November 2022, has (already) impacted the education of physics in both foreseeable and unexpected ways. GenAI presents opportunities for students and faculty to enhance efficiency but poses challenges as well, including plagiarism and overreliance. In this study we collected, compared and contrasted physics university students’ (N=41 Bachelors, N=18 Masters) and faculty (N=17) perceptions and utilization of GenAI – with a focus on physics related tasks including problem solving, coding, writing and doing (experimental) research. Using an online survey, we found that their self-reported ability to determine relevance and accuracy of a resource does not depend on the source of the resource (traditional vs GenAI). Further, while the students and faculty's patterns of GenAI use differ, their views on legitimate uses of GenAI are largely comparable and agree that GenAI is less acceptable as greater agency is transferred to the AI. However, while their perspectives align in general terms, there remain nuances suggesting that specific expectations may need to be adapted based on the role and responsibilities of the user. This study can serve as a starting point for initiating discussions between students and educators about acceptable uses of GenAI and the reasoning behind such boundaries.