#725: A Transpositional Grammar Cyber-Social Study of Student–AI Relationships

As AI-mediated feedback becomes embedded in higher education, understanding how learners construe relationships with AI and human reviewers becomes increasingly more important. This pilot applies Transpositional Grammar (TG) to student prompts, AI-generated images, and reflections in a cyber-social learning environment. Graduate students (n = 7; P300–P306) at a Midwestern U.S. research university contributed coursework artifacts coded collaboratively by a human researcher and GPT‑5, yielding Cohen’s κ = 0.761. TG analysis identifies five recurrent transpositions—expansion, visualization, reinterpretation, materialization, and affective deepening—and surfaces relational accountability as a new dimension of learner expectations. The full study (n = 35; P300–P334) will add a second human coder to triangulate interrater reliability and inform the design of affect-sensitive human–AI feedback ecosystems in graduate instruction.