Public relations as legitimacy infrastructure in the algorithmic state

Public governance is increasingly mediated by algorithmic decision systems and artificial intelligence (AI). While public governance refers to the broader institutional practices through which collective authority is exercised, AI governance designates the more specific arrangements—ethical, technical, and regulatory—through which algorithmic systems themselves are managed; the two are connected because AI governance now reshapes the communicative conditions under which public governance is legitimated. Although prior research has examined AI governance from ethical, technical, and regulatory perspectives, limited attention has been given to how algorithmic authority is communicatively constructed and evaluated. This article reconceptualizes public relations (PR) in algorithmically mediated public-sector environments as legitimacy infrastructure rather than symbolic communication. Drawing on legitimacy theory and core PR traditions—including dialogic communication, organizational listening, transparency, and relationship management—the article argues that in the algorithmic state, PR evolves from strategic messaging to legitimacy infrastructure, becoming embedded in the conditions under which algorithmic authority is experienced and evaluated. It introduces the Legitimacy Infrastructure Model (LIM), which identifies four relational domains—value framing, justificatory signaling, narrative positioning, and trust calibration—through which legitimacy is formed when authority is embedded in socio-technical systems and encountered through system outputs prior to dialogue. The model extends PR theory by reframing legitimacy as a continuous relational accomplishment under infrastructural authority, articulating testable theoretical propositions, and clarifying the advisory and design-sensitive role of communication leadership in AI-enabled public institutions. In doing so, the article positions PR as a constitutive capability shaping how institutional authority is interpreted and evaluated in datafied governance contexts.