AI text generators are not simply a new writing tool—they are a shift in how meaning is made. Built on statistical probability, AI output strips away the lived context, dialogic depth, and poetic function that make language human. If text shapes experience and agency, as Gadamer and Habermas argue, then the expansion of AI demands a renewed commitment to hermeneutics: teaching students to read AI critically, interrogate its output, and deploy it with purpose. The literature classroom is uniquely equipped for this work. In spring 2026, I developed a sequence of three scaled online courses embedding AI tools within an American literature curriculum. Grounded in three principles—preparing students for their present reality, building analytical assignments, and requiring metacognitive reflection—the courses ask students to engage AI as an intellectual interlocutor rather than a shortcut. Assignments require students to contemplate the revision process through AI-powered dialogue, compare AI-generated and human-authored texts, and defend interpretations against AI counterarguments. This presentation shares the assignment designs and pedagogical framework behind this approach, offering humanities educators a tested model for teaching the literacy our moment demands.