Artificial intelligence (AI) is increasingly reshaping public administration, with implications for administrative efficiency, public service innovation, data-driven governance, and responsible digital transformation. This article examines AI readiness in the Romanian public sector by focusing on the balance between individual-level readiness and institutional support for AI adoption. Drawing on the Romanian dataset of the Global Artificial Intelligence Adoption Survey, the study analyzes responses from 353 public sector employees, with a main analytical focus on 290 respondents who reported using at least one AI tool at work. The analysis examines AI use intention, perceived AI work impact, individual AI competence, workplace support, organizational AI readiness, trust/ethical confidence, and risk perception. The study does not measure public service innovation directly; instead, it examines individual-level AI adoption outcomes and discusses their implications for the capacity of public institutions to support responsible public service innovation. The results show that, among Romanian public sector respondents who use AI, AI adoption outcomes are strongly associated with individual-level factors, especially perceived competence, frequency of use, positive experience, and trust in AI. By contrast, organizational AI readiness remains comparatively weaker and does not show a significant unique association with AI use intention or perceived work impact once individual competence, workplace support, trust, and practical AI exposure are included in the regression models. Within the available national sample, differences between local, regional, and national administrative levels are limited, suggesting that the gap between individual readiness and institutional support is not confined to a specific level of government. The findings suggest that individually driven AI adoption must be translated into institutional capability if AI is to support responsible public service innovation and sustainable digital transformation in the public sector.
Bunea et al. (Tue,) studied this question.