Generative AI in college coursework: a longitudinal investigation of instructor communication and student plagiarism

This longitudinal study investigates students’ beliefs and behaviors regarding generative artificial intelligence (GenAI) utilization for college coursework and the role of instructors’ course policies and messaging in students’ decision-making. Guided by the theory of planned behavior (TPB), we note areas of change and stability in students’ GenAI perceptions and practices within and across semesters using short-term longitudinal data collection over the course of three semesters. Initial findings supported TPB as a viable framework for predicting students’ frequency of submitting AI-generated work as their own, yet more recent results only partially support the model. Results highlight shifts in course policy and instructor message framing and sidedness regarding GenAI and changes in how messaging predicts student perceptions. In particular, the prevalence and impact of two-sided messages changed significantly over time. Theoretical and practical implications for instructional communication are discussed.