Authorship after automation: originality, attribution, and creative responsibility in AI-assisted literary production

Abstract

Background: Generative artificial intelligence has unsettled conventional literary authorship by separating textual production from creative control, public attribution, and accountability across writers, platforms, publishers, and institutions. Objective: This study examines how originality, credit, disclosure, and responsibility are configured in publicly documented cases and governance materials concerning AI-assisted literary production. Method: A qualitative multiple-case documentary analysis was conducted on 27 verified public records, comprising primary, secondary, and contextual sources coded through matrices of creative control, attribution–disclosure alignment, and responsibility. Results: Stronger authorship claims appeared where human actors directed narrative purpose, selected alternatives, transformed generated material, and authorised publication. Attribution was most credible when disclosure specified the extent, function, timing, and audience of AI involvement, whereas delayed, private, or absent disclosure weakened correspondence between contribution and credit. Responsibility was distributed across authors, publishers, platforms, professional bodies, and model providers according to their control over production, classification, access, remuneration, and enforcement. Implication: Literary governance should align authorship claims and accountability obligations with demonstrable contribution, institutional capacity, and remedial power rather than with bylines or tool use alone. Novelty: This study offers an integrated, process-based framework that distinguishes textual generation, transformative originality, attributional transparency, and layered creative responsibility within one comparative, publicly available corpus.