Purpose: This study aims to examine the effect of 'debate education using generative artificial intelligence (AI)' on 'debate efficacy' targeting elementary school students in the 5th and 6th grades. Through this, we aim to provide valuable information on debate classes using generative AI to field teachers and researchers in debate-related studies. This study aimed to provide students with a positive communication experience by allowing them to articulate their arguments, engage with peers, and persuade others. Additionally, it sought to serve as a foundation for fostering students' collaborative communication skills and digital language literacy. Furthermore, in alignment with the introduction of the “Media” domain in the 2022 Revised Korean Language Education Curriculum, this study aimed to offer pedagogical implications for teachers regarding debate education using generative AI. Lastly, it sought to expand the scope of Korean language education by preparing students to actively adapt to the rapidly evolving communication environment of the future society. Methods: To achieve this, a triangulation study combining quantitative and qualitative research methods was conducted. The quantitative research compared and analyzed the pre-post debate self-efficacy of the participants, while the qualitative research explored the effects of debate classes using generative AI by analyzing portfolios generated by the participants. Descriptive statistics and the Wilcoxon signed-rank test were used to analyze quantitative data. As some of the response data from the participants did not satisfy the assumption of normality, the pre-and post-test changes in debate efficacy among the participants were analyzed using a nonparametric Wilcoxon signed- rank test (p<.05). The reliability of this study was verified through Cronbach’s ⍺ value. Portfolio analysis was employed for the qualitative data analysis. Results: The “Debate Self-Efficacy” of the participants was measured pre- and post-intervention using the Wilcoxon Signed-Rank Test. The results showed a significant improvement in all subcomponents of debate self-efficacy, including Emotional Self-Efficacy, Cognitive Self-Efficacy, and Social Self-Efficacy. In particular, significant improvements were observed in the components of Emotional Self-Efficacy, namely ‘expectation,’ ‘persistence,’ and ‘emotion regulation.’ Furthermore, the analysis of portfolios composed of activity sheets developed for this study revealed that engaging in debate activities using Generative AI positively enhanced the debate self-efficacy of the participants. Conclusion: This study demonstrates through action research and empirical analysis that Korean language debate classes utilizing generative AI are effective in enhancing students' debate efficacy. This study hopes to serve as a stepping stone for fostering Inclusivity and enhancing Communication Competence among learners, thereby contributing to the strengthening of their Debate Self-Efficacy. Furthermore, it is expected that this research will contribute to the activation of Debate Education Using Generative AI in school.