Sustainable Digital Governance of AI-Based Decision Support Systems in Public Administration: A Socio-Technical Capacity Framework

Artificial intelligence (AI)-based decision support systems (DSSs) increasingly shape how public organizations classify cases, rank risks, allocate attention, and interpret administrative information. Although these systems may improve administrative performance, their contribution to sustainable digital governance depends on institutional arrangements that preserve accountability, adaptability, inclusiveness, and public justification. This conceptual article develops a lifecycle-oriented socio-technical governance capacity framework through a structured synthesis of public administration, digital government, decision support systems, responsible AI, socio-technical systems, sustainability, and risk governance scholarship. The framework distinguishes five interacting layers—technical, organizational, legal–ethical, societal, and adaptive—and eight cross-layer capacities: data governance, algorithmic accountability, human oversight, legal and ethical assurance, organizational learning, inter-organizational coordination, public justification and contestability, and adaptive monitoring and response. It further identifies four system-level relationships concerning capacity alignment, lifecycle variation, distributed responsibility, and adaptive feedback. Isolated safeguards provide limited assurance when they are institutionally disconnected or unsupported by the authority to learn and intervene. By conceptualizing responsible AI-based decision support as a configuration of interdependent capacities, the framework connects AI governance with the institutional resilience, accountability, and adaptability required for sustainable public administration. It provides a diagnostic basis for comparative research and organizational assessment throughout the lifecycle of AI-based DSSs.