This paper examines what has to align before educators can integrate LLM-based ethics scaffolding into ordinary higher-education teaching without increasing marking burden, rule ambiguity, or inappropriate reliance on automated output. The paper develops an educator-facing framework from a bounded cross-case synthesis of recent higher-education literature and two prior design case, one addressing classroom use of an LLM-supported ethics activity, and one addressing assessment redesign for keeping student reasoning inspectable under AI use, rather than reporting a new adoption survey. The paper proposes six design conditions for course integration: pedagogical positioning, workflow fit, visible reasoning, governance clarity, infrastructural reliability, and educator capability-building. Taken together, these conditions suggest that adoption is less a matter of individual willingness to use AI than a matter of coordinated design across pedagogy, assessment, policy, infrastructure, and staff development. The paper contributes a practical framework, a temporal reading of how the conditions surface across the life of a course, and a worked illustration showing how the framework travels beyond computing to other judgment-intensive disciplines. Its central argument is that responsible adoption becomes more plausible when AI is bounded to a clearly defined pedagogical role and when student deliberation remains interpretable, assessable, and governable in everyday teaching conditions.

Authors:
Stefania Zourlidou, University of Koblenz, Germany

About the Presenter(s)
Dr. Ing. Stefania Zourlidou is a postdoctoral researcher and academic advisor at the University of Koblenz, Germany.

See this presentation on the full schedule – On Demand Schedule

Bookmark