Disaggregating Authorship: A Multidimensional Theory for AI-Assisted Writing

This paper develops a multidimensional account of authorship in response to the rise of generative AI in writing. Rather than treating authorship as a binary status—either preserved or displaced—the analysis conceptualizes it as a cluster concept composed of distinct normative dimensions, including intention, evaluative governance, accountability, origination, identity, labor, and closure.

The paper argues that the relative importance of these dimensions varies across writing practices. In governance-centered domains such as academic, legal, and technical writing, authorship primarily signals responsibility and oversight. In expressive domains, it also functions as an index of identity and stylistic origination. Generative AI expands textual production but does not assume governance, accountability, or closure. Its impact on authorship therefore depends on which dimensions are normatively central within a given practice.