Assessing AI-Generated Online Discussion Assignments in an Asynchronous Bioethics Course: A Comparative Analysis of AI vs Human

One purpose of Generative AI tools in the higher education context is to reduce the workload of course design; however, their ability to provide ethically nuanced content on complex themes remains unproven. This research compares AI-generated discussion assignments with those designed by humans in a Bioethics module on patient consent. An expert evaluated each assignment using Salmon's (2013) Five-Stage Model, focusing on clarity, objectives, interactivity, reflection, and support for knowledge. Results show a significant quality gap: AI outputs are structurally sound but lack true peer discussion, ethical reasoning, and guidance through socialization and collaboration. Conversely, the human assignment used inclusive language, culturally appropriate scenarios, and invitations for debate, with clear instructions, transforming prompts into meaningful learning experiences. Overall, free AI models function better as rapid-draft tools than as autonomous discussion generators. Instructors can use AI to start the process and then add depth, sensitivity, and interactivity through human effort.