Scaffolding mediation for group discussion with enhancement participation empowered by the AI mediation

Addressing the persistent challenge of disorganized group discussions in social interactions requires effective strategies. This study leverages recent advancements in group dynamics to develop a structural diagram designed to enhance members’ engagement in group discussions. By implementing this improved participation strategy, we aimed to facilitate smoother group discussions within a structured framework involving participants from diverse professional backgrounds (N=264 participants across 44 groups), conducted in real-world settings such as language laboratories. Utilizing permutation techniques, we identified a phenomenon termed "scaffolding mediation", which optimizes, mediates, and corrects permutations among group members. This approach significantly improved the efficacy of group discussions, achieving an average mediation rate of 81.3% compared to 34.2% in non-control and self-mediated structural diagrams (N=132 participants across 22 groups). Unlike previous strategies that isolate mediators from the chaos of group discussions (N=66 participants across 11 groups), the scaffolding mediation approach remains neutral. It enhances group dynamics and boosts discussion effectiveness by fostering greater participation.