How AI is changing the future of public administration

Artificial intelligence and automation promise to make public services more efficient, reduce administrative work, and free up time for tasks requiring human judgment. But what happens when these technologies also make public institutions more complex, less transparent, and harder to govern? 

These questions were at the center of Where is the public sector heading?, a seminar organized by House of Governance and Public Policy (GaPP) & House of Innovation (HOI) together with the Swedish Agency for Digital Government (Digg) at the Stockholm School of Economics on November 2, 2026. 

The event brought together Nicklas Berild Lundblad, independent researcher, author, and policy expert; Martin Berg, Professor of Media Technology at Malmö University; and Anne Kaun, Professor of Media and Communication Studies at Södertörn University. 

The discussion was moderated by Ida Lindgren, Professor of Information Systems at Linköping University, and introduced by Magnus Mähring, Erling Persson Professor of Entrepreneurship and Digital Innovation and Head of the House of Innovation at SSE. 

Together, the participants discussed how philosophy and social theory can help us understand the changes taking place in public administration.

When automation promises more than efficiency 

One of the central themes was the gap between the promises of automation and what happens when new technologies are introduced into public organizations. 

Automation is often presented as a way to eliminate repetitive tasks, save working hours, and allow employees to focus on more meaningful work. Yet the discussion highlighted how these expectations can shape organizations long before the technology itself delivers measurable results. 

Public agencies may spend considerable time preparing for automation by mapping processes, reorganizing workflows, and adapting their operations to what future systems might require. 

This raises an important question: Are organizations changing because technology has already improved their work, or because they expect it to do so in the future? 

The discussion also examined how commercial interests and technology providers help shape expectations about what AI and automation can achieve. 

Can AI make public decisions harder to understand? 

Another important theme concerned transparency and accountability. 

Public administration has traditionally been built around the principle that decisions should be rational, consistent, and explainable. But increasingly sophisticated AI systems challenge this idea. 

As automated systems become more complex, understanding how they reach particular conclusions can become difficult, even when their outputs appear accurate or useful. 

The panel explored a particularly challenging possibility: What happens if AI systems become very effective at delivering public services, but neither citizens nor public officials fully understand how they work? 

The concern is not only that technology might produce incorrect decisions. It is also that public institutions could become dependent on systems whose internal processes are difficult to examine or challenge. 

This raises broader questions about accountability, public trust, and the role of human judgment in government decision-making. 

When technology creates new complexity 

The conversation also explored how digitalization can create new challenges for the institutions it is intended to improve. 

As societies become more technologically interconnected, public organizations must respond to increasing volumes of information, new forms of interaction, and changing expectations from citizens. 

One example discussed was the possibility of AI agents acting on behalf of individuals. Such systems could make it easier for citizens to contact authorities, submit requests, or appeal decisions. At the same time, they could generate volumes of administrative work that existing institutions are not equipped to handle. 

This illustrates a wider dilemma: Technology can help public organizations manage complexity, but it can also introduce new forms of complexity that require further technological and organizational responses. 

The discussion emphasized that these challenges cannot be understood through technology alone. They also depend on how public organizations are structured, how decisions are made, and how employees adapt new systems to everyday work. 

Who should shape the digital public sector? 

A recurring question throughout the seminar was who gets to influence the direction of technological change. 

Decisions about AI and automation are made at many levels, from political leadership and public procurement to IT departments and individual workplaces. Different groups may also have very different understandings of what automation means and what it should achieve. 

The discussion therefore turned to the importance of involving citizens in decisions about the technologies used in public services. 

Participatory approaches, such as citizens' assemblies, were discussed as possible ways to bring different perspectives into the development of digital public services. Examples of cities exercising greater influence over technology and its use were also raised. 

At the same time, the panel acknowledged that meaningful participation is not straightforward. Citizens may find it difficult to identify what they want from technologies they have not yet experienced or fully understood. 

The discussion also highlighted the role of universities in developing the knowledge, concepts, and critical perspectives needed to support informed decisions about digitalization. 

Beyond the question of efficiency 

The seminar returned repeatedly to a broader question: What kind of public sector do we want to create? 

While efficiency and productivity are important considerations, the discussion showed why they cannot be the only measures of successful digital transformation. 

Transparency, democratic participation, professional judgment, and public accountability also need to be part of the conversation. 

Rather than treating technological development as inevitable, the seminar encouraged a closer examination of the assumptions, choices, and values shaping how AI and automation are introduced into public institutions. 

The seminar was organized by the Swedish Agency for Digital Government (Digg) and the Stockholm School of Economics. It was based on the fourth edition of Digg reflekterar, a publication series exploring contemporary issues in digitalization through research and critical reflection.