Working with semantic machines: The role of semantic reconciliation in public sector data governance

Abstract

Local governments are beginning to use artificial intelligence (AI) in internal administrative work, including recruitment. In a decentralized municipality, however, a standardized AI assessment must travel between a technology vendor, central HR specialists, recruiters, and hiring managers responsible for different services. This creates a tension between representational portability and contextual adequacy. Based on a qualitative case study of a large Swedish municipality, we examine an AI interviewing system that translated candidates' open-ended responses into competency indicators and narrative summaries used in early-stage recruitment. We develop a process model of semantic reconciliation comprising five recurrent forms of work: scoping what could be delegated to the system, mapping vendor categories to municipal and occupational vocabularies, refining interpretations when role context was lost, embedding shared interpretations across a distributed organization, and composing AI reports with locally accountable human judgment. The study contributes to research on local-government AI and data governance by showing how semantic machines make evaluative information portable while requiring continuous organizational work to restore context. It also specifies practical arrangements through which municipalities can use common AI tools without allowing standardized outputs to displace role-specific knowledge or human decision authority.