Assessing Swedish Institutional Readiness for AI-Driven Digital Sovereignty: A Case Study

Sammanfattning

Artificial intelligence is increasingly becoming part of public administration, creating opportunities for efficiency, automation, improved decision-making, and better public services. At the same time, the growing use of AI raises questions about digital sovereignty, since public sector organisations often depend on external technology providers, cloud services, digital infrastructure, and data ecosystems that they do not fully control. This creates challenges for public agencies that must comply with legal, ethical, and security requirements while also developing the organisational capacity and capability needed to use AI effectively. Although the European Union has introduced regulatory frameworks such as the AI Act, NIS2, and the GDPR, there is limited empirical research on how Swedish public sector organisations are prepared to implement and govern AI in ways that support digital sovereignty. Therefore, this study examines the research questions “How prepared are public organisations to implement and govern AI in ways that support digital sovereignty?” and “What factors enable or hinder this readiness?” The study was conducted as a qualitative single-case study with embedded units of analysis, in which the Swedish public sector was treated as the overall case and participating agencies as embedded units. Qualitative data were collected through semi-structured interviews with 12 respondents from Swedish public agencies and through document analysis of agency documents, national policy materials, and EU regulatory frameworks. The material was then analysed using thematic analysis. The findings yielded four main themes: AI governance and organisational structure; AI implementation maturity and organisational readiness; digital sovereignty and infrastructure strategy; and organisational approaches to AI implementation. The findings show that Swedish public sector agencies are partially prepared to implement and govern AI in ways that support digital sovereignty. Readiness is strongest in governance, compliance, and organisational control, while broader implementation is constrained by competence gaps, limited resources, legal uncertainty, legacy systems, procurement challenges, and dependence on external technology providers. The study concludes that AI-driven digital sovereignty in the Swedish public sector requires not only regulatory compliance but also stronger organisational, technical, and strategic capacity to reduce dependency and support responsible AI implementation.