Promoting student engagement in online collaborative writing (OCW) activities has been a critical concern for educators. Previous research has attempted to design conversational agents (CAs) utilizing retrieval-based models to engage students in collaborative learning. However, few studies have yet explored the design of CAs for OCW based on generative AI (GAI) models. Researchers are calling for investigations into how GAI technology can be better utilized to support learning. Addressing this gap, this study integrates advanced AI technologies (i.e. the retrieval-based model, GAI model, and retrieval-augmented generation) to develop a CA aimed at enhancing students' engagement in OCW activities. Furthermore, a quasi-experiment involving 78 undergraduate students was conducted to explore the effects of this CA on students' engagement (including behavioral, cognitive, and emotional engagement) and group writing performance. The results indicate that the CA did not significantly impact behavioral engagement, emotional engagement, or group writing performance. However, it was found to significantly enhance students' cognitive engagement, particularly by supporting students in sharing opinions, explaining concepts, and engaging in analysis. This research offers both theoretical and practical implications for better utilizing GAI technology to facilitate OCW activities.

中文翻译：

---

通过基于 AI 的生成式对话代理提高学生对在线协作写作的参与度


促进学生参与在线协作写作 （OCW） 活动一直是教育工作者关注的关键问题。以前的研究试图设计对话代理 （CA），利用基于检索的模型来吸引学生参与协作学习。然而，很少有研究探索基于生成式 AI （GAI） 模型的 OCW 的 CA 设计。研究人员呼吁研究如何更好地利用 GAI 技术来支持学习。为了解决这一差距，本研究整合了先进的人工智能技术（即基于检索的模型、GAI 模型和检索增强生成）来开发一个旨在提高学生对 OCW 活动的参与度的 CA。此外，进行了一项涉及 78 名本科生的准实验，以探索该 CA 对学生参与度（包括行为、认知和情感参与度）和小组写作表现的影响。结果表明，CA 没有显着影响行为参与、情感参与或小组写作表现。然而，研究发现它显着提高了学生的认知参与度，特别是通过支持学生分享观点、解释概念和参与分析。本研究为更好地利用 GAI 技术促进 OCW 活动提供了理论和实践意义。