May 7

Governments across Europe are moving beyond digitising services toward embedding Artificial Intelligence (AI) into the core of public administration. This shift is often described as a transition from e-government to “AI government.” But what does this transformation actually involve, and what does it demand from public institutions beyond deploying new technologies?

In this talk, the speaker presents insights from a reflexive, empirically grounded framework for AI adoption in the public sector, developed through an in-depth case study of Estonia’s Bürokratt initiative—a national AI-enabled virtual assistant that serves as a single gateway to public services. Rather than treating AI as a standalone tool, the framework examines how AI becomes institutionalised across legal, organisational, operational, and socio-technical dimensions.

The session traces Estonia’s trajectory from e-government toward AI government and, conceptually, toward what can be described as an agentic state—a governance model in which AI enables adaptive, data-informed action while remaining embedded in human oversight, legal mandates, and democratic accountability. The focus is not on technological sophistication, but on institutional readiness, inter-agency coordination, human–AI delegation, and the governance conditions under which AI can be scaled responsibly.

Drawing on mixed empirical evidence, the talk highlights why many AI initiatives struggle to move beyond pilot phases: regulatory ambiguity, fragmented ownership, workforce capacity gaps, and misalignment between technical ambition and institutional reality. It reframes AI adoption as a multi-phase governance challenge, requiring iterative learning rather than one-off implementation.

For policymakers and digital leaders, the session offers a practical lens to assess where their administrations stand—not on an AI maturity curve alone, but on a broader trajectory of governance transformation. The key message is clear: AI does not simply automate government; it reshapes agency, responsibility, and the state’s capacity to act—and must therefore be governed accordingly.