Newest papers

CoWriter: Guiding Methodological and Synergetic Human-LLM Co-Authoring

Shmuel Amar, Benjamin Lieber, Ori Shapira, Debanjana Kar, Sebastian Dennerlein, Yufang Hou, Ido Dagan · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31277 · cited by 0 works (OpenAlex) · open-access PDF

Portuguese

Enquanto os LLMs transformaram a geração de texto, as interfaces padrão de bate-papo e canvas freqüentemente reduzem a autoria humana a supervisão bastante passiva, causando fadiga de revisão, dependência excessiva de IA e o afastamento do papel humano na escrita de forma longa, baseada em fontes. Os sistemas prévios geralmente abordam fases de escrita isoladas, deixando a colaboração humana LLM abrangente de ponta a ponta em grande parte inexplorada. Neste artigo, apresentamos o CoWriter, um esquema proativo de agentes que adapta metodologias humanas (co-)autoriais experientes da ciência cognitiva para orientar flexivelmente todo o fluxo de trabalho de escrita, ao mesmo tempo que incentiva ativamente o usuário a se envolver. Em um estudo exploratório

Machine translation (Meta M2M100); the English original follows.

Abstract

While LLMs have transformed text generation, standard chat-and-canvas interfaces often reduce human authoring to quite passive supervision, causing review fatigue, over-reliance on AI, and the sidelining of the human role in long-form, source-grounded writing. Prior systems typically address isolated writing phases, leaving comprehensive end-to-end human-LLM collaboration largely unexplored. In this paper, we introduce CoWriter, a proactive agentic scheme that adapts expert human (co-)authoring methodologies from cognitive science to flexibly steer the entire writing workflow while actively encouraging the user to engage. In an exploratory user study, compared to a standard LLM baseline, authors using CoWriter better expressed their perspectives and maintained a higher sense of document ownership, supported quantitatively by $2.2\times$ more successful agent engagements. Furthermore, external reader judges rated the resulting texts higher in "human touch" qualities like storytelling and integrating unique personal experiences. Overall, our results demonstrate that proactively engaging users through established human practices is beneficial for effective LLM-supported co-authoring, providing useful methodologies, techniques and insights for future research on collaborative agents.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Workflow, Computer science, Storytelling, Human–computer interaction, Knowledge management, Scheme (mathematics)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange