ARTIFICIAL INTELLIGENCE AS A PILLAR OF ACCOUNTABILITY IN THE PUBLIC SECTOR: AN ANALYSIS OF STRATEGIC CHALLENGES IN THE DIGITAL ERA
TL;DR
The results indicate that the automation of audit processes can lead to significant reductions in processing time (up to 96% in pilot workflows), while simultaneously highlighting a structural skills deficit (“humanware”) that may affect the quality, interpretation, and legitimacy of algorithm-assisted decisions.
Abstract
: This paper examines the transformative impact of Artificial Intelligence (AI) on accountability mechanisms in the public sector, with a focus on the transition from sample-based auditing to emerging forms of extended algorithmic monitoring. The research adopts a mixed-methods design, combining documentary analysis of European Court of Auditors (ECA) reports with a contextual assessment of institutional barriers within the Romanian public administration. The results indicate that the automation of audit processes can lead to significant reductions in processing time (up to 96% in pilot workflows), while simultaneously highlighting a structural skills deficit (“humanware”) that may affect the quality, interpretation, and legitimacy of algorithm-assisted decisions. In this context, the paper proposes a multidimensional framework for algorithmic oversight, supplemented by a SWOT analysis of AI implementation in Romania and recommendations aimed at reducing fragmentation in administrative data. The main contribution of the study lies in conceptualizing the relationship between technological efficiency and institutional accountability, emphasizing that AI performance in the public sector is contingent on organizational maturity and the quality of data infrastructure. The findings offer relevant implications for public policy design and the responsible implementation of AI, in alignment with European regulatory requirements, including the EU AI Act.