Generative Artificial Intelligence and the Reconstruction of Subjectivity in Literary Creation

Generative artificial intelligence can participate in literary composition by proposing plots, producing passages and revising prose. Its participation raises questions about the position from which a work is imagined, judged and presented to readers. This paper examines the reconstruction of literary subjectivity through theoretical argument and analysis of research on assisted writing. It distinguishes subjectivity from observable interaction, the institutional attribution of authorship and evaluations of textual quality. Literary theories of authorship and dialogue are brought into conversation with accounts of writing processes and technological mediation, with attention to their different explanatory purposes. A comparative framework considers intention, lived experience, narrative form and responsibility across conception, drafting and revision. The analysis identifies tensions involving delegated intention, narrative anchoring, simulated experience, stylistic specificity and public attribution. It proposes that agency can be sustained through active reconsideration of creative aims, consequential aesthetic judgement, attention to experiential and cultural particularity, and accountable relations with readers. These are analytical proposals, not a validated measure of subjectivity. Existing studies concern specific interfaces and writing tasks and cannot establish a universal effect on literary identity. The paper reports no new experiment or interviews; its focus is short fiction and novel writing.