Publikations-Information

Download

Rifat Amin, Seyit Baran Oezdemir, Sven Mayer, Andreas Butz
EmoArc: Interactive Emotion Graph for Human-AI Collaborative Writing
In Proceedings of the 18th International Conference on Advanced Visual Interfaces (AVI '26). ACM. New York, NY, USA. 2026. (bib)

With the advent of large language models, creative writers can get support for refinement, co-writing, and text generation. Yet, writers often struggle to understand how model-driven rewrites affect the emotional trajectory of their narratives. We investigate how interactive visualizations can help writers gain awareness, agency, and control over the emotional flow of their texts. First, we designed an interactive visualization-based writing tool that allows writers to adjust emotional pacing, explore different emotional directions, and revise the emotional storyline in real time. Second, we conducted a user study with creative writers (N = 24) to evaluate its effectiveness. Our results show a significantly improved enjoyment, exploration, and perceived value. Writers highlighted increased emotional awareness and narrative coherence; simultaneously, they expressed concerns about authorship. Thus, support can enrich the creative process when designed with transparency, user agency, and adaptability in mind, contributing to the understanding of augmenting human writing creativity.