Abstract
This paper reports on the redesign of a supportive peer-review system developed through the practical use of generative AI within the Association for Dialogue and Support of Japan (ADSJ). The original objective was to develop operational standards for AI-assisted review. However, practical application revealed that AI review was effective for supporting research development but insufficient for evaluating overall manuscript completeness. As a result, AI review and manuscript completeness assessment emerged as distinct functions, and the effectiveness of AI-assisted pre-submission support was also identified. Based on these findings, ADSJ established a new review system integrating AI review, manuscript assessment, reviewer-based evaluation, and editorial decision-making. Furthermore, this practice demonstrated that peer review can function not only as a post-submission evaluation process but also as a pre-submission support mechanism for academic writing.
