The conflicted student and the internal battle over generative AI use

The common narrative around university students and generative AI is one of crisis. Underpinning this is an implicit assumption: if students have the capacity to cheat, they will. Our survey reveals two missing factors from that narrative: students have ethical principles, and students want to learn. While a minority of students might fit that implicit assumption of the crisis narrative, a similar proportion refuse to use any genAI tools at all. In between these are the majority: students using genAI tools in what they self-define as a ‘supportive’ way. These students, finding these tools unavoidable, worrying about being disadvantaged or simply finding them valuable, struggle to use them in a way that doesn’t compromise their personal sense of authorship, fits with their environmental and social principles, and doesn’t detract from their ability to learn. Students typically care greatly about these issues and are internally conflicted about this, feeling guilt when using these tools and personally trying to find that invisible line where support becomes cheating. They also use a range of rhetorical techniques to frame their personal usage and preserve their sense of authorship. We argue that we need to work with students, using approaches that enhance these feelings and encourage self-discipline, rather than primarily adopting a ‘criminal’ model of checks, detection and punishment.