When Automation Risk Becomes Public Responsibility: Institutional Signals, Normative Orientations, and the Politics of Labor Protection

This project examines how automation-related labor market risks are translated into judgments of public responsibility. As artificial intelligence and related technologies increasingly reshape work tasks and employment conditions, workers face heightened uncertainty regarding job security, redeployment, and income stability. While existing research has extensively documented the economic and distributive consequences of automation, less is known about how members of the public evaluate such risks politically—specifically, under what conditions automation-related displacement is perceived as a matter of collective responsibility rather than individual adaptation. The central purpose of the project is to identify the evaluative mechanisms through which automation risk becomes politically salient. The study focuses on two factors that are theoretically central to this translation process. The first is structural risk, operationalized as exposure to automation that is plausibly outside the control of individual workers. The second is institutional signaling, captured by whether public institutions provide formal protection or adjustment assistance in response to automation-related disruption. The project further considers how these factors interact with individuals’ normative orientations toward collective responsibility, understood as evaluative lenses that shape how institutional signals are interpreted rather than as direct predictors of policy support. To address these questions, the project employs a randomized vignette experiment embedded in a nationally administered online survey in South Korea. Each respondent evaluates a single hypothetical work context in which an AI-based system is introduced. The vignettes vary experimentally along two dimensions: the degree of automation exposure and the presence or absence of institutional protection. This design allows the analysis to distinguish between the recognition of automation-related risk as a structural condition and the subsequent attribution of responsibility to public institutions. By holding the basic work context constant while varying these dimensions, the study isolates how evaluative judgments change across institutional and normative environments. The analysis proceeds by examining three analytically distinct outcomes. First, it assesses whether automation exposure is recognized as a structural condition rather than as the result of individual choice. This outcome captures the initial stage of evaluation and serves as a diagnostic measure of risk recognition. Second, the study examines attribution of public responsibility, which constitutes the primary outcome of interest. This outcome captures whether respondents believe that the government should bear responsibility for protecting workers affected by automation. Third, the analysis evaluates support for costly protection policies, such as income support and employment protection measures, to assess whether responsibility attribution extends beyond symbolic acknowledgment to endorsement of policies that entail material costs. The project is designed with clear expectations regarding these outcomes. Automation exposure is expected to increase recognition of structural risk regardless of institutional signals. In contrast, attribution of public responsibility is expected to emerge only when automation exposure is paired with institutional protection, indicating that risk is publicly recognized rather than privately borne. Moreover, this translation from risk to responsibility is expected to be stronger among respondents who endorse norms of collective responsibility. Finally, support for costly protection policies is expected to follow a similar pattern, though normative conditioning may attenuate at this stage as material costs become salient. By clarifying when automation-related risk becomes a matter of public responsibility, the project contributes to broader debates in political economy and labor politics. It advances existing research by moving beyond accounts that treat attitudes toward automation as direct reflections of economic self-interest or generalized technological optimism or pessimism. Instead, it highlights the conditional and staged nature of evaluative judgments, emphasizing the role of institutional signals and normative orientations in shaping political responses to technological change. In doing so, the project provides a framework for understanding why similar levels of automation risk can generate divergent political reactions across institutional contexts and normative environments.