As neural language models become increasingly effective, their integration into real-world applications is expanding. Despite this, many of these applications are often limited in the types of interactions they facilitate, with text-based chats being one of the most common modes of engagement. One area that remains largely unexplored is the use of AI in the scientific peer review process, particularly in the crucial editing phase. There have been many different tools that support the writing process, but recent advances in AI have created the potential for novel features. This thesis introduces ReViewer, an AI-assisted interactive editing tool designed to enhance the editing stage of the peer review process. ReViewer leverages the capabilities of large language models (LLMs) to assist editors by automating content analysis and extracting actionable suggestions from reviewer feedback. The application offers several innovative features, including statement-based highlighting, severity classification, contextual improvement suggestions, and visualizations for review overlap analysis. These features empower editors to make informed, objective decisions while maintaining control through a human-in-the-loop system that supports but does not replace their expertise. To assess the effectiveness and usability of ReViewer, we conducted a preliminary study involving both novice and expert participants in the context of scientific peer review. This study offered valuable insights into users' attitudes towards integrating AI-assisted features into their scientific editing workflows, as well as their experiences with the specific features we implemented. The findings not only highlight the benefits and challenges of using AI tools in the review process but also provide a deeper understanding of how these tools can complement human expertise. Finally, we outline potential directions for future research and development, focusing on refining AI-driven features further to improve their usability and trust for AI-assisted editing.