Intervening to Increase Community Trust for Fair Network Outcomes

Refugees or immigrants who arrive in new countries often feel isolated. In this work, we examine how a resource-bounded public entity can make recommendations to increase integration of these new arrivals into a community. The community is made up of agents who engage in a strategic network formation process; agents join periodically - new arrivals are the refugees. The public entity meanwhile makes a limited number of edge-formation recommendations (according to its resource constraint) per iteration in order to increase integration of refugees. This work investigates the relationship between community trust and network fairness. First, we show that increasing the public entity's resource allocation will not compensate for low trust in the community. Then, we introduce two trust-increasing interventions by the public entity: a targeted advertising campaign, and an announcement to increase transparency. We find that diverting a fraction (20%) of the public entity's resources to a targeted advertising campaign can increase trust and fairness in the community, especially in low trust scenarios. We find that personalized, local announcements are more effective at increasing fairness than global announcements in low trust scenarios; they almost double our fairness metric in some cases. Importantly, the transparent announcement requires no extra resource expenditure on the part of the public entity. Our work underscores the importance of community trust - low trust cannot be compensated for with resources. This work provides theoretical support for these trust-increasing interventions, which we show can lead to increased integration of refugees in communities.