🤖 AI Summary
Existing LLM-based writing assistance systems struggle to balance personalized expression with user controllability. This paper introduces GhostWriter, an AI-augmented writing probe featuring a novel dual-path mechanism: “implicit style modeling” enables seamless, zero-intervention stylistic adaptation via LLMs, while “explicit instructional feedback” supports interpretable, user-intervenable style refinement through interactive prompting and multimodal editing feedback. This design preserves fluid personalization while strengthening authorial agency and reflective capacity. An 18-participant user study demonstrates that GhostWriter significantly enhances textual stylistic distinctiveness (p < 0.01) and subjective sense of control (+42%), validates the efficacy of diverse guidance strategies, and yields transferable human–AI co-creation design principles.
📝 Abstract
Writing is a well-established practice to support ideation and creativity. While Large Language Models (LLMs) have become ubiquitous in providing different kinds of writing assistance to different writers, LLM-powered writing systems often fall short in capturing the nuanced personalization and control necessary for effective support and creative exploration. To address these challenges, we introduce GhostWriter, an AI-enhanced writing design probe that enables users to exercise enhanced agency and personalization. GhostWriter leverages LLMs to implicitly learn the user's intended writing style for seamless personalization, while exposing explicit teaching moments for style refinement and reflection. We study 18 participants who use GhostWriter for editing and creative tasks, observing that it helps users craft personalized text and empowers them by providing multiple ways to steer system output. Based on this study, we present insights on people's relationships with AI-assisted writing and offer design recommendations to promote user agency in similar co-creative systems.