Constructing acceptable use: Instructor boundary‑setting with generative AI

Abstract

Acknowledging Generative AI’s presence in the higher education classroom, instructors must navigate unclear boundaries around acceptable and unacceptable student use. This quantitative study surveyed 89 instructors across four research‑intensive universities in the United States to examine their perceptions of student Generative AI use and the strategies they adopt to manage it. Findings reveal that instructors construct task-specific acceptability judgments rather than applying university rules, with core writing tasks deemed largely unacceptable while supplementary activities have wider acceptance. A perception-acceptance gap was apparent for writing tasks, showing that instructors suspect widespread AI use for the very activities they find most problematic. Prevention strategies emphasize pedagogical over surveillance approaches, but nonetheless almost 40% of the instructors reported catching students using AI to plagiarize, with most cases resolved informally. Open responses raised themes of an academic integrity crisis, learning concerns, and pragmatic tool use at the course level, with larger concerns about ethical and societal implications of Generative AI and institutional tensions