Enhancing Conflict Resolution Skills through Artificial Intelligence-Based Problem-Based Learning in Civic Education at Indonesian Primary Schools

Background/Purpose. This study investigates the effectiveness of integrating Artificial Intelligence (AI) with Problem-Based Learning (PBL) to enhance high school students’ conflict resolution skills within the context of civic education. The initiative responds to the increasing need for pedagogical models that simultaneously foster ethical reasoning, empathy, and peaceful negotiation in adolescent learners, particularly in complex socio-political environments. Materials/Methods. Using a quasi-experimental one-group pretest-posttest design, the study was conducted with 90 eleventh-grade students from three urban schools in Yogyakarta, Indonesia. The intervention utilized a custom-built AI platform featuring real-time simulations of social conflicts, an NLP-based chatbot for ethical guidance, and personalized feedback mechanisms. Conflict resolution skills were measured across three core dimensions: empathy, peaceful negotiation, and ethical reasoning. Results. Statistical analyses showed significant improvements across all assessed indicators (p < 0.001), with very large effect sizes (Cohen’s d > 1.6). Post-test scores not only increased in average value but also displayed more uniform distribution, indicating effectiveness across diverse student profiles. No significant gender-based differences in learning gains were observed, suggesting the model’s inclusivity. Regression analysis revealed that students with lower baseline scores experienced the greatest improvements, underscoring the adaptive strength of the AI-PBL approach. Conclusion. The findings support the potential of AI-enhanced PBL to develop both cognitive and affective aspects of conflict resolution skills in civic education. Theoretically, this research contributes to social constructivist and differentiated instruction paradigms. Practically, it offers a promising direction for equitable technology integration in education. However, the absence of a control group and the urban-limited sample constrain broader generalization. Future studies should explore longitudinal and cross-cultural applications to evaluate the model’s sustainability and scalability.