ALGORITHMIC GOVERNANCE IN THE UNITED STATES PUBLIC SECTOR: DESIGNING TRANSPARENT AI SYSTEMS FOR POLICY DECISION-MAKING

The rapid and intensively high tempo of embedding artificial intelligence (AI) in the public domain has changed policy-making processes, especially in the US. Algorithmic rule-making, while presenting challenges of efficiency, predictive performance, and data-driven policy-making, also leads to basic questions of transparency, accountability, and regulation. This paper uses a systematic literature review (SLR) approach in analysing the influence of transparent AI systems in facilitating democratic governance and legitimacy within the U.S. public sector. The review used the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) to provide rigor and replicability. A broad search on Scopus, Web of Science, JSTOR, IEEE Xplore, and Google Scholar, supported by policy briefs of US agencies like the Office of Science and Technology Policy (OSTP), National Institute of Standards and Technology (NIST), and Government Accountability Office (GAO) resulted in 512 records. During the application of inclusion and exclusion criteria, 42 quality studies to be synthesized were identified. There are three clusters of themes overall: (i) the requirement for interpretable and explainable AI to foster citizen trust and policy legitimacy; (ii) systems of institutional accountability that strike a balance between risk and innovation; and (iii) mechanisms of oversight such as auditing, ethical rules, and governance structures that ensure responsible adoption of AI. The review picks up major gaps in current literature, i.e., empirical testing of algorithmic systems in actual policy environments and scarce citizen participation in AI governance mechanisms. While U.S. federal initiatives are being undertaken, e.g., the Blueprint for an AI Bill of Rights and NIST's AI Risk Management Framework, scaling transparency remains a concern. This research adds to the expanding literature on algorithmic governance through a systematic survey of extant knowledge, recognition of enduring gaps, and articulation of policy guidelines for transparent, accountable, and trustworthy AI systems in the U.S. public sector. Ultimately, the research asserts that algorithmic governance needs to draw upon technological advancements while protecting democratic values in order to protect fairness, legitimacy, and public trust in policymaking by AI.