Introduction
In the era of permacrisis, complexity and urgency have become common features of policy making, challenging both the design of policies and their accountability. The 2024 EPS annual debate, which took place in Florence on 25th of November of 2024, in collaboration with the European University Institute (EUI) School of Transnational Governance, discussed the ever-increasing demands for policy makers, covering everything from how expertise can be best incorporated into the policy cycle to develop efficient and decisive policy responses to the increasingly complex policy problems that often surpass the national level. Issues such as the role of experts in policy making, evidence-informed policy making, public policy co-creation, and democratic accountability were linked and discussed.
The panellists were Heather Grabbe, Senior Fellow at Bruegel, Claudio Radaelli and Diane Stone, both Professors at the School of Transnational Governance, EUI and Gaby Umbach, Professor at the Robert Schuman Centre for Advanced Studies, EUI. The discussion was chaired by Stella Ladi, co-editor of EPS. All panellists have long academic and policy experience on the topic and provoked a lively debate with the audience in a full auditorium. The discussion drew special attention to the role of political scientists and other experts,when acting as government advisors during a crisis. The panellists explored how expert advice and democratic accountability can co-exist when multiple levels of governance should work together to respond to global challenges such as climate change or fast-paced technological developments.
The debate that we are pleased to publish in the EPS is a reflection on the discussion that took place in Florence and consists of an introduction co-authored by the three EPS co-editors, an article on ‘Evidence Use in Policy Making’ by Marylou Hamm and Imogen Bayley (2026, issue 25.2), an article on ‘Knowledge and Democratic Accountability in Policy Making’ by Imogen Bayley (2026, issue 25.2), and a final article on the ‘Epistemic Foundations of Transnational Governance’ by Diane Stone (2026, issue 25.2). All authors actively participated in the Florence debate in different capacities. What follows is a summary of the key arguments of the three papers and how they link to some of the challenges that have further consolidated in the last year in the relationship between expertise and policy making.
Knowledge and policy making in a changing world
The three articles of this debate aim to shift our attention to aspects of the knowledge – policy making nexus that have not been discussed enough in the literature by now.
The article by Hamm and Bayley (2026, issue 25.2) claims that we need to understand the relationship between expertise and policy making as political and not as purely technocratic to capture why expertise is sometimes used in policy making, but not always. They identify three core challenges. First, the literacy challenge, which refers to the uneven levels of evidence literacy among policy makers, the public, and experts themselves. Hard data, which is simpler to illustrate, seem to be privileged over more qualitative and discursive evidence. Policy-oriented research is preferred to valuable basic research because of the difficulty of immediately translating its findings into tangible actions. Second, the capacity challenge, which is often perceived as the most pressing issue, has to do with the institutional capacity of public bodies to process evidence. Finally, the political challenge denotes the political will to use, ignore, or even manipulate evidence and is on the rise when populist governments are in control. Only when these three challenges are seen as interconnected can we shed light on why and how expertise is used on different occasions.
The second article by Bayley (2026, issue 25.2) explores policy learning, which is a key policy-making mechanism and questions the learning assumption in policy making when seen as a normative ideal. By re-thinking the findings of three seminal works on the role of expertise in policy making, she argues that learning is, in practice, constrained by institutional dynamics and by the political timing, and it should not be seen as an act of idealism.
Stone (2026, issue 25.2), in the third contribution to this debate, turns her attention to the role of expertise in the evolving space of transnational governance, which makes the democratic accountability of experts even more challenging than at the national level. She points out that the complexity of transnational problems makes it necessary to involve experts from different disciplines, but also from nontraditional sources of knowledge, such as indigenous knowledge or knowledge that derives from professional experience. She emphasises the role of social scientists, and political scientists in particular, in constructing frameworks and concepts that feed into the policy debates and the challenge of epistocratic tendencies, which is more acute at the transnational level.
Further developments
It is only over a year since the debate was conducted at EUI, and some of the issues that were discussed have already exacerbated. We briefly discuss two significant developments: the changing role for experts in the second Trump administration and the popularisation of the use of Artificial Intelligence (AI) platforms.
The first development is the deterioration of the relationship between experts and policy makers in the US since the second Trump administration took office. In line with Hamm and Bayley (2026), the political challenge can be observed. The most immediate concern is the institutional and information gap that the shutting down or shrinkage of the activities of organisations such as the Woodrow Wilson Centre, the United States Institute of Peace (USIP), the United States Agency for International Development (USAID), and Voice of America creates. While these organisations were not purely research-oriented, they were the home of numerous experts and produced or held valuable primary data. They also engaged in informed analysis, debates, and the training of scholars. Their destruction is felt in the US but also beyond, since these resources were used by researchers across the world. It remains unknown who is going to feel this gap, but since demand remains high, at least in the private sector, private consultancies and risk analysts seem to be strengthened in the first instance. However, databases and basic research cannot be replaced by the private sector, but only by truly independent and resourceful research institutions. Unfortunately, US universities and their research centres are also in a dire situation, following the restructuring of higher education and its funding, which is affecting several disciplines from gender studies to climate research and which are no longer considered to be priorities to say the least.
What is even more striking is that the old guard of experts and think-tanks does not seem to be replaced by new organisations or expert advisors. The second Trump administration does not seem to value expert analysis and scientific data and prefers a more personal and political decision-making style. This is a particularly significant shift since the US was one of the pillars of evidence-informed policy making, often influencing, via soft diplomacy tools, the rest of the world. It remains to be seen whether this shift is here to stay and whether other countries and regions, including Europe, will follow a similar path. Interestingly, while experts and think-tankers are losing their ground, technological companies and AI gurus have gained more access to the White House.
This brings us to the second big development, which is the popularisation of AI. It is only in the last year that the access and use of AI have exponentially increased, at least in the West, in both the public and private sectors, but also at the citizen’s level. The way AI will interact with expert advice and evidence-informed policy making will be determined in the coming years. There are two possible directions: first, and preferably, for AI to be used and developed as a tool that can assist evidence-informed policy making by producing and analysing data that can assist in policies such as the prevention of climate disasters. In that scenario, experts will be able to provide more accurate and fast solutions to policy makers. The second direction could be an attempt to replace experts with AI agents that would be expected to assist policy makers in their work instead of experts. The outcome of such a choice is less straightforward.
Conclusions
The debate that we held in 2024 is only a first step for a continuous discussion on knowledge and democratic accountability in policy making. Given the latest developments, political scientists should be at the forefront of these discussions. As is often the case with complex problems and challenging times, not all answers to dilemmas can be immediately provided. What could be done, though, is to continue the discussion in an open and inclusive manner by holding dear the values of academic freedom and independence. Inclusivity should be interpreted broadly to also embrace more interaction with citizens for both the production of knowledge and the co-production of evidence-informed policy-making solutions. The European political science community has a lot to offer at this time in history, and EPS is a welcome forum for scientifically rigorous and analytically sharp work on these issues.
Data availability statement
Data availability is not applicable to this article as no new data were created or analysed in this study.
Financial support
Aside from the salaried position of the authors, there was no other funder in the drafting of this paper.
Competing interests
There are no conflicts of interest or competing interests in the development of the article.
Open access
