ABSTRACT
Digital public administration scholarship has expanded on broad trends in government digitalization, yet its generic categories, such as ‘citizen’, ‘user’ and ‘service’, systematically overlook what is distinctive about welfare administration. From automated debt recovery and predictive risk modelling to algorithmic child protection and emerging generative AI applications, the digital transformation of social policy creates profound dilemmas that generic frameworks cannot address. This paper proposes digital social administration (DSA) as a new conceptual paradigm—the ensemble of digitized and automated systems, data infrastructures, algorithmic tools and organizational practices through which welfare states assess eligibility, allocate benefits, govern risk and manage disadvantaged populations. We develop the TOV Three-Dimensional Analytical Framework to operationalize this paradigm, focusing on the Technical, Organizational and Valuational dimensions of DSA and their relationships. Applying the framework to four signature dilemmas, and considering the emerging challenges posed by advanced AI and AGI demonstrates its ability to reveal dynamics that single-dimensional approaches overlook. DSA constitutes a qualitatively new administrative form, comparable to the bureau-professional state or new public management, and demands concepts adequate to its specificity. We conclude by outlining a research agenda organized around the TOV dimensions and identifying implications for social policy, practice, advocacy and legal reform.