Conversations, Not Commands: Challenging Current Models of AI Attribution

Current AI citation guidelines assume simple, one-shot interactions, but academic work involves complex, multi-turn conversations with rich context. Through examples- including the AI conversations that created this presentation- we'll examine how current attribution models fall short in both scholarly work and student assignments. When identical prompts yield different responses, what does meaningful AI documentation look like? Rather than presenting solutions, this interactive session invites participants to challenge existing citation models and explore how we might better document AI's contributions in academic settings, even when traditional goals of replicability are impossible to achieve. [Title and abstract written with help of Claude AI]