Continuation or reconfiguration? A Multi-Level Framework of AI implications for local government

AI technologies are increasingly adopted in local government across Europe, yet we lack a conceptual understanding how they interact with national and task-/policy-specific contexts, different administrative cultures and value logics. This paper develops a multi-level framework to explore how AI may either reinforce existing governance arrangements or reconfigure the institutional role of local government and the nature of the citizen-state encounter. Thereby we draw on literature concerning comparative public administrative, administrative culture and tradition, public ethics and values, as well as the public encounter and street-level bureaucracy as well as our own research on the digital transformation in European municipalities to generate propositions about the conditions under which the use of AI leads to continuity or change in a comparative perspective. European local governments are embedded in different administrative cultures and traditions and have a wide range of functional responsibilities: From a formalized regulatory culture emphasizing the rule-of law (typical for Continental European Napoleonic/Federal and Eastern European countries), to a cooperative contact culture (Nordics), and a public interest culture stressing pragmatism, flexibility, and negotiation (Anglo-Saxon countries). Local governments’ task profile can follow a multi- or single-purpose model with either a broad or a narrow spectrum of responsibilities. Tasks range from human services, to administrative services, to regulatory enforcement tasks; with task-/policy-specific dominant values, e.g. following the principles of universality vs. particularity or public ethics of justice vs. public ethics of care. Given these differences, different future scenarios concerning AI implications can be imagined, which we summarize under continuation vs. reconfiguration. At the macro-level, we propose that AI may either be added as just another layer to existing administrative practices without altering the role and position of local government in the multi-level system, or that AI deployment may reconfigure the local level of government by enabling new forms of (control) re-centralization. As a consequence, AI could redefine the role of municipalities as intermediaries who must interpret, translate, and support citizens navigating AI-driven services. AI may even become functionally equivalent to local service provision, fundamentally altering citizen-state interactions. At the meso-level, we propose that local government, on the one hand, continue traditional practices by, for instance, intensifying legality, standardization, and rule-bound decision-making, especially in legalistic settings, thereby strengthening hierarchical bureaucratic control (algorithmic cage, curtailment). In settings with a cooperative contact culture, AI may be another form of knowledge besides professionals (algorithmic colleague, enablement), while for contexts with a public interest culture AI deployment could be more pragmatic and flexible in order to solve problems and pursue strategic aims. On the other hand, dominant public values guiding administrative actions can be altered. For instance, AI may strengthen rule-orientation and efficiency, with fairness, transparency, and responsiveness at risk. These developments may be different across administrative cultures and organizational logics. At the micro-level, AI leads to a dual development: It can reinforce street-level bureaucrats’ role as state agents by reducing frontline discretion, and simultaneously, strengthen the role as citizen agents, when street-level bureaucrats correct or resist AI outputs, acting as guardians of the individual citizen.