Social credit scores reduce interpersonal cooperation and trust

TL;DR

Social credit scores, intended to promote trust, actually reduce cooperation and trust between people, worsening social inequalities.

Contribution

This study reveals that social credit systems can decrease trust and cooperation, contrary to their intended purpose.

Findings

Access to social credit scores leads to reduced trust and cooperation between individuals.

Social credit scores create persistent biases in partner perception, resistant to correction by behavioral evidence.

Lower scores disproportionately disadvantage individuals, increasing social and economic inequalities.

Abstract

Social credit systems (SCS) are increasingly used by government agencies and private firms to assign scores to individuals based on social status and behavior. These scores subsequently impact access to social and economic opportunities, resources, and interactions. The ethical and privacy concerns of SCS are frequently overlooked due to their purported, yet unverified, social and economic benefits. In this paper, we examine the impact of social credit scores on cooperation, trust, and partner selection in economic decision-making. Contrary to the intuitive notion that social credit scores facilitate interactions by increasing transparency, we find that the availability of SCS information leads to lower trust and reduced cooperation between individuals. Additionally, we find that social credit scores create persistent biases in the perception of interaction partners, which remain resistant to change even when directly contradicted by relevant behavioral evidence. These effects disproportionately disadvantage individuals with lower scores, exacerbating existing social and economic inequalities, and demonstrate that reputational systems can amplify polarization rather than promote fairness and cooperation. These findings provide important insights for policymakers, regulators, and businesses considering the adoption of social credit systems by highlighting the potential for unintended negative consequences in both public and commercial domains.

Figures