Who Owns the Text? Design Patterns for Preserving Authorship in AI-Assisted Writing

TL;DR

This paper investigates how AI-assisted writing affects writers' sense of authorship and presents an ownership-aware co-writing editor designed to preserve authorial voice while delivering efficiency gains. Using a within-subject study (N=176) across three professional genres, the authors compare generic AI suggestions, persona-framed coaching, and a lightweight style-personalization pipeline that conditions suggestions on a writer’s prior text. The results show that AI assistance generally reduces psychological ownership, even when writers actively control input, but style personalization partially mitigates this loss and can boost adoption, especially when paired with persona framing. The authors distill five design patterns—on-demand initiation, micro-suggestions, voice anchoring, audience scaffolding, and provenance cues—that offer practical guidance for building authorship-preserving AI writing tools and discuss implications for provenance, user control, and future features like persistent AI contribution indicators. Overall, the work highlights an ownership–efficiency paradox: AI can reduce cognitive burden with broadly intact output, yet erode the personal sense of ownership unless voice alignment is explicitly supported.

Abstract

AI writing assistants can reduce effort and improve fluency, but they may also weaken writers' sense of authorship. We study this tension with an ownership-aware co-writing editor that offers on-demand, sentence-level suggestions and tests two common design choices: persona-based coaching and style personalization. In an online study (N=176), participants completed three professional writing tasks: an email without AI help, a proposal with generic AI suggestions, and a cover letter with persona-based coaching, while half received suggestions tailored to a brief sample of their prior writing. Across the two AI-assisted tasks, psychological ownership dropped relative to unassisted writing (about 0.85-1.0 points on a 7-point scale), even as cognitive load decreased (about 0.9 points) and quality ratings stayed broadly similar overall. Persona coaching did not prevent the ownership decline. Style personalization partially restored ownership (about +0.43) and increased AI incorporation in text (+5 percentage points). We distill five design patterns: on-demand initiation, micro-suggestions, voice anchoring, audience scaffolds, and point-of-decision provenance, to guide authorship-preserving writing tools.

Figures (6)

• Figure 1: The ownership--efficiency paradox and our ownership-aware editor. (a) AI writing assistance can increase efficiency (speed, fluency/quality) while reducing human ownership (agency, voice), creating an ownership--efficiency tension. (b) Our editor operationalizes two levers within an on-demand micro-suggestion loop: writers click Ask Emma (persona-framed coaching) or Ask AI (generic suggestions), a gray inline suggestion appears, and writers explicitly accept (Tab/button) or ignore it; accepted text turns black and remains editable.
• Figure 2: Evaluation design overview. The study uses a mixed design: participants are first randomly assigned to a personalized or non-personalized group (between subjects). Within each group, participants complete three genre-specific writing scenarios (within subjects): an unassisted email baseline (no AI), a generic suggestion mode for proposals (contextual next-sentence suggestions without persona framing), and a persona-framed coaching mode for cover letters ("Emma" as a style-and-tone coach). Scenario order is counterbalanced using a Latin square. In the personalized group, suggestions are additionally conditioned on a brief writing sample provided by the participant.
• Figure 3: Experimental interface for uploading participants' writing samples.
• Figure 4: Experimental interface for persona-framed coaching (style-enhancing persona) condition.
• Figure 5: Experimental interface for Control (No AI) condition.