Abstract
Generative AI has unsettled a long-standing assumption in higher education: that a written submission is evidence of the thinking behind it. This is a particular problem in marketing communications education, where the capabilities that matter most, reading persuasion critically and arguing for a stronger alternative, are those a language model can now imitate. This paper offers a reflective account of an assessment designed in response, currently running in a postgraduate marketing communications subject at an Australian university. Students work in self-selected groups to analyse a real advertisement, then propose and present an improved version as a Pecha Kucha, before defending their decisions in a live discussion with the class and tutor. Generative AI is permitted for producing slide visuals but not for the analysis or argument. The paper argues that a rule of this kind cannot secure an assessment on its own, since compliance cannot be verified, and that what secures the task is the oral defence: a student who has not done the thinking cannot reconstruct it under questioning. This reframing draws on recent work distinguishing discursive from structural assessment change.
Keywords
Assessment Design, Generative AI, Pecha Kucha, Interactive Oral Assessment, Academic Integrity, Reflective Practice.
