Infrastructural AI and the Redistribution of Agency in Writing Pedagogy

This article examines how the integration of generative artificial intelligence (AI) into writing instruction redistributes agency across students, tools, instructors, and institutional systems. Rather than treating AI as an external aid or a discrete instance of automation, the article frames contemporary language models as infrastructural technologies that condition the space of possible utterances before writers make conscious choices. Drawing on rhetoric and composition studies, philosophy of technology, and classroom-based observation, it argues that many current pedagogical debates mischaracterize the problem as one of authorship loss or detection, when the more consequential shift concerns responsibility, interpretive transparency, and decision-making within human–machine assemblages. The article proposes an analytic orientation for documenting how agency concentrates, diffuses, or cools across writing processes under conditions of persistent AI use, and suggests pedagogical practices that foreground reflection, traceability, and accountability rather than prohibition or automation. By reframing AI as infrastructure rather than tool or agent, the article offers a pragmatic foundation for teaching writing in environments where human judgment remains essential but no longer operates alone.