Simulating the Confessional Self: AI-Generated Writing, Literary Subjectivity, and the Performance of Confession

Abstract

The dramatic advances in generative artificial intelligence have increasingly foregrounded questions of literary creativity, authorship, and textual authenticity in current debates, echoing the centuries-old concern over the definition of subjectivity in literary writing. Though much recent critique on AI authorship has focused on the fields of creative writing, literary authorship and ethics, little has engaged with its implications for the confessional mode, a literary mode invested in lived experience, emotional truthfulness, and autobiographical self-revelation. This paper explores whether AI-produced confessional writing can authentically engage with the confessional tradition, or whether it will simply define the idiomatic register of confessional writing by reproducing its lexicon. Using qualitative comparative textual analysis, it assesses AI confessional writing on loneliness, depression, motherhood, and identity crisis, juxtaposing it with canonical confessional texts by T. S. Eliot, Anne Sexton, and Adrienne Rich. Drawing on psychoanalytic theory, confessional studies, post structuralism, post-human literary criticism, and the works of Sigmund Freud, Jacques Lacan, Michel Foucault, Roland Barthes, Donna Haraway, and N. Katherine Hayles, it demonstrates how AI reproduces stylistic, idiomatic, and rhetorical confessional features while lacking in the consciousness, memory, and psychological depth attributed to individual literary subjectivity, rather than realizing confession, AI performs a simulation of subjectivity through language. Ultimately, the paper offers a new understanding of how contemporary AI confessional writing disrupts understandings of confession, authorship, truthfulness, and the relationship of language to the self.