AI²: Investigating student use of artificial intelligence and faculty-student perspectives on academic integrity

Candidate Name: Terry Shirley Program: Educational Leadership - Learning, Design, and Technology Committee Chair: Dr. Ji Yae Bong​​​​​​​ Committee Members: Dr. Beth Oyarzun, Dr. Stella Kim, Dr. Casey Davenport Abstract: Artificial intelligence has rapidly advanced in recent years and now impacts many aspects of higher education. Generative artificial intelligence (GenAI) offers significant opportunities for student learning while raising complex questions about academic integrity. A growing body of empirical research is emerging on how students use GenAI, with more limited work on what motivates that use and how students and faculty evaluate various GenAI assisted practices. Using a quantitative correlational and comparative design, this study examined data collected from students and faculty at a large public university. GenAI adoption was widespread: 92% of students reported using GenAI, more than 80% had used it for academic purposes, and 90% identified ChatGPT as the most commonly used GenAI tool. Adoption rates were generally consistent across demographic groups, but differences emerged between STEM and non-STEM students regarding specific uses. Regression analyses supported expectancy-value theory (EVT) as a framework for understanding GenAI adoption. Utility and attainment were the strongest predictors of academic GenAI use, with the model explaining 31.8% of the variance. EVT motivational constructs and perceived cost also significantly predicted students’ dishonesty evaluations of a variety of GenAI use scenarios. Faculty generally rated unacknowledged GenAI use as more dishonest than students did, although both groups shared a near identical hierarchy of ethical severity across uses. Acknowledging GenAI use significantly reduced perceived dishonesty across all scenarios for both groups and substantially narrowed faculty/student perception gaps. These findings underscore the importance of transparent acknowledgment standards and clearly defined institutional guidance for responsible academic GenAI use.