ALGORITHMIC INEQUALITY AND SOCIAL TRUST IN AI-MEDIATED PUBLIC SERVICES

Abstract

Advancements in AI are opening up possibilities for better efficiency in public service delivery, quicker decision-making, and more service access. In developing societies, however, the swift integration of AI-driven systems brings about significant concerns about algorithmic inequity, opacity of procedures, and the loss of social trust. This research examines the factors that shape citizen trust in public institutions that use AI tools, such as their perception of fairness, transparency, accountability, and accessibility. The research design is a sequential explanatory mixed methods that includes 2,500 citizens in the survey and 14 public practitioners and marginalized citizens interviewed qualitatively. The results show that the factor most strongly associated with institutional trust is perceived transparency, followed by perceived procedural fairness; opaque automated decisions, inequal service outcomes, and lacking grievance mechanisms are associated with distrust and civic alienation. The findings also reveal that vulnerable groups are more likely to suffer a negative outcome of an algorithm in systems involving welfare, health care, and administrative eligibility. Qualitative evidence corroborates these findings, illustrating that citizens frequently misread an automated decision when there is no explanation or justification as exclusionary, unfair, and not ‘human’. The study finds that AI-powered public services can enhance governance, but they must be accompanied by clarity, openness, accountability and contestability in the design process, as well as in the use of data, and monitoring and oversight. The study's results add to the on-going body of knowledge in the fields of AI ethics, public administration, and digital governance and highlight the need for a balance between technological efficiency and fairness, explainability, and democratic accountability in the context of developing countries.