Successful e-government initiatives depend on the alignment of the objectives of public services with its stakeholders. While stakeholder engagement is widely acknowledged to be essential for developing e-government services, debates persist around the definition and categorisation of stakeholders within e-government contexts. This article critically examines how stakeholder categorisations can help e-government public services to be user-centric. However, the study also highlights the complexity of involving different stakeholders in e-government, particularly those with varying levels of interest or influence, the difficulties inherent in building and maintaining relationships with stakeholders, unequal power dynamics, inadequate legislation and political structure, lack of time and willingness throughout the public service cycle. The findings from a systematic literature review are synthesised into a framework and reveal the relevance of more targeted and context-sensitive approaches to stakeholder categorisations, linking them to different engagement methods and challenges to ensure the long-term success and sustainability of user-centric public services.
Artificial Intelligence (AI) adoption in the public sector increasingly depends on the readiness of Public Data Ecosystems (PDEs) and their governance. However, limited empirical evidence exists on how governments articulate AI-enabled activities in policy documents and whether these ambitions align with the data-related conditions required for implementation. This study presents a cross-national, data-centric assessment of AI activity patterns across eight European Union Member States. Using computational text analysis and expert assessments, we identify 210 AI-enabled activity patterns across seven categories and evaluate their perceived strategic importance and documented operational salience. The results reveal gaps between strategic priorities and operational articulation, particularly in PDE-dependent areas such as data interoperability, data quality, stakeholder engagement, and explainable and human-centric AI. A two-stage clustering approach combining hierarchical clustering and K-means identifies three data-centric profiles of policy-visible AI configurations: AI-first, data-first, and balanced. These profiles indicate that commitments to AI coexist with uneven policy articulation of the governance and data conditions associated with deployment. Rather than measuring implementation outcomes, the study provides a comparative framework for analysing how AI-enabled activities are framed, prioritized, and made visible in policy documents relating to PDEs, while informing future research on data-centric AI governance in the public sector.
As Artificial Intelligence (AI) penetrates all sectors of society, academic attention to its relationship with democracy is rapidly increasing, leading to a growing body of literature. Optimists expect AI to strengthen democracy, while pessimists argue that it may undermine its essential features. Yet, the analysis of the relation between AI and democracy often remains fuzzy, fragmented, and focused on a narrow range of topics. Therefore, this study aims to provide a systematic understanding of the AI–democracy interplay. A basic framework with four dimensions of democracy – comprising four Ps: Polity, Politics, Policy, and Polis – is introduced to structure expert opinions about the potential opportunities and threats that AI poses for democracy over the next five to ten years. Thirty experts contributed to a Delphi survey, and the analysis shows that the 4Ps framework offers a valuable way to organize the AI–democracy debate, allowing for nuance and less obvious aspects. While, similar to the literature, experts generally identified more threats than opportunities, the survey identifies various opportunities to strengthen democracy that quite absent in debates. In addition, the Delphi survey brings to the fore a fifth P – Performance – which involves a general addition to the basic framework – and emphasizes the (in)ability of the democratic system to solve citizens’ problems and to generate desirable results. As such, this article provides both an empirical diagnosis of the relation between AI and democracy, and a potential agenda for further actions to preserve and strengthen democracy. This article contributes to the literature in three ways: by offering a novel framework to structure the fuzzy AI–democracy debate across its different dimensions; by surfacing less obvious and underexplored aspects of this interaction through a Delphi expert survey; and by moving beyond a predominantly risk-focused perspective in AI governance to also consider an array of opportunities AI adoption may create for democracy.
The rise of digital government has introduced the demand for new skills that have become necessary for citizens to independently and meaningfully engage with public services. However, rather than ensuring accessibility regardless of skill levels, governments are shifting the burden onto individuals to actively address their own skill deficits in order to exercise fundamental rights, such as applying for welfare benefits. Drawing on qualitative interviews with Dutch professionals in the field, this research uses abductive reasoning to build the conceptual framework of the ‘responsive citizen.’ This concept reflects the expectation that citizens respond to skill deficits and proactively meet institutional skill demands to access public services. It is developed in conversation with the notion of the ‘responsive state,’ which places responsibility for inclusive service delivery on the public administration. By contrasting these two perspectives, the article contributes to public administration scholarship by critically analyzing shifting responsibilities in citizen-state relations.
