Algorithmic Governance and Social Equity: Evaluating Bias, Transparency, and Public Trust in AI-Driven Welfare Systems (Reimagining Digital Public Policy Through Fairness, Accountability, and Inclusive Algorithmic Design)

Abstract

The growing use of artificial intelligence (AI) in public welfare systems has led to what scholars have labelled an emerging "algorithmic state" (Busuaioc, 2021; Yeung, 2018) in the conduct of contemporary governance. While AI decision systems are a promising source of administrative efficiency and data-driven objectivity, there is substantial evidence that these decision systems may replicate structural inequalities that are embedded in historical datasets (Barocas & Selbst, 2016; O'Neil, 2016). The issues of algorithmic bias, opacity, and accountability have compelled many questions about social equity, democratic legitimacy, and public trust (Mittelstadt et al., 2016; Pasquale, 2015). This research examines the role of perceived algorithmic bias and transparency in people's trust towards AI-based welfare governance. Drawing from the theoretical foundations of algorithmic regulation (Yeung, 2018), accountability in automated decision-making (Diakopoulos, 2016), as well as AI adoption in the public sector (Wirtz et al., 2019), the research builds an integrated governance-trust model. The study also uses a mixed-method approach with survey data and qualitative policy interviews in order to investigate whether transparency mechanisms mediate the link between perceived fairness and institutional trust. The findings are expected to show that algorithmic explainability and accountability structures are key to successfully increase legitimacy in digital welfare systems. Through bridging social justice theory with algorithmic governance scholarship (Lodge & Mennicken, 2021; Veale, & Brass, 2019), this research adds to the current growing conversation on ethics in AI and makes policy recommendations for inclusive and citizen-centered digital public administration.