Daewoo Jin
Assignments & Activities Archive
Activity Description
This activity aims at developing students’ ability to distinguish statistical text generation from human meaning-making by observing how two AI systems exchange directed arguments in response to short, human-written prose. Students analyze rhetorical clarity, disruption, and absence in AI-to-AI debate transcripts, within one class meeting, and in the context of first year composition. Two AI systems debate anonymized sentences drawn from students’ earlier reflective writing while students watch and take notes. The debate uses reflective writing because reflective sentences carry a situated speaker and a lived stake, so when models attempt to sustain arguments about personal experience, their breakdowns in rhetorical momentum become more legible to students than they would be with thesis-driven prompts where academic conventions can mask limits of intentionality (Kuhn and Simpson). This design foregrounds the constitutive dimensions of discourse as relational practice that exceeds pattern recognition.