European Union Science Diplomacy Alliance - Fostering EU Science Diplomacy to contribute to a better future for all. in

Image created with ChatGPT

Anna-Lena Rüland from our partner, University College London, has published a highly interesting article in Minerva on “Organized Scientific Interests and Agenda-Setting in European Science, Technology, and Innovation Policy” (link). She was so kind to briefly summarize her key findings for the EU Science Diplomacy Alliance.

Scientific advocacy and diplomacy for science

Large-scale, international science projects, such as the Large Hadron Collider or the International Thermonuclear Experimental Reactor, are often cited as typical examples of diplomacy for science.[1] Without the support and advocacy of high-level diplomats and policymakers, these “big science” projects would not come into being. Scientists and scientific communities typically play a subordinate role. Or so the argument goes in the policy-driven science diplomacy scholarship. Yet this framing largely glosses over the fact that scientists and scientific communities often put a “big science” project on the agenda of policymakers in the first place. Shedding light on the strategies and tactics that they use in the process, may help complicate our current understanding of diplomacy for science. I will use the case of the so-called “CERN for Artificial Intelligence (AI)”, a proposal to organize Europe’s AI research in a CERN-like ecosystem, to do so.

What is CERN for AI?

There are multiple, and in part clashing, visions of CERN for AI (this blog post provides a detailed overview). I will focus on the vision of the Confederation of Laboratories for Artificial Intelligence (AI) Research in Europe (CAIRNE), possibly one of the fiercest advocates for a CERN for AI. CAIRNE, and in particular its founding members, promote a CERN for AI that:

• Consists of a central hub, supported by a network of research clusters;
• Is used to conduct research on all aspects of AI;
• Has close links to industry;
• Develops AI that aligns with European norms and values; and
• Is publicly funded.

CAIRNE’s founding members started to advocate for CERN for AI in 2017, based on two central arguments. First, they consider it vital to bring Europe’s top AI researchers together in a central place to coordinate their research efforts and to agree on a few top research priorities for the field. According to CAIRNE’s founding members, this is needed to help the European AI community overcome its fragmentation. Second, CAIRNE’s founding members argue that more hardware is needed to enable cutting-edge AI research in Europe and to catch up with AI frontrunners like the US and China. When CAIRNE was founded in 2018, CERN for AI became one of the central issues that the organization advocated for. However, despite CAIRNE’s continuous advocacy, CERN for AI has not (yet) materialized. This may change soon, though, as the proposal has made it to the highest political level: Ursula von der Leyen recently proposed to set up a “European AI Research Council where we can pool all of our resources, similar to the approach taken with CERN.”

Lobbying pays off

Between 2018 and 2025, CAIRNE used different means to draw policymakers’ attention to CERN for AI. As I argue in my recently published article “´We need a CERN for AI’: Organized scientific interests and agenda-setting in European science, technology, and innovation policy,” three strategies proved particularly useful for the organization. First, CAIRNE used a parliamentary strategy to convince different members of the European Parliament (MEPs) of its proposal’s merit. This strategy has paid off, as several MEPs began to back CERN for AI during the deliberations for the AI Act. Second, CAIRNE’s founding members invested a great deal of time and effort into promoting CERN for AI by writing countless opinion pieces and giving interviews for prominent media outlets across the EU. For example, their proposal featured in “der Tagesspiegel”, “Science Business”, and “Frankfurter Allgemeine.” This media coverage drew attention to CERN for AI beyond a specialist audience. Finally, and probably most importantly, CAIRNE mobilized the AI community, think tanks and the broader public for its proposal by organizing CERN for AI-dedicated events and speaking about CERN for AI within different fora. Across all these strategies, CAIRNE credibly framed CERN for AI as an issue that needs to be addressed at the EU-level: Firstly, because the funding needed for CERN for AI would likely exceed the capacities of any one EU member state (a key rationale of diplomacy for science initiatives) and, secondly, because the entire European AI research community should benefit from and participate in CERN for AI.

Framing is key

A CAIRNE strategy that created controversy, especially within the European AI community (but also among some sociologists of science), was the framing of the proposed AI research infrastructure as a CERN for AI. Several people that I interviewed for my article argued that, in principle, it is beneficial to use the renowned CERN “brand” to promote a large-scale, European (AI) research infrastructure. Yet in the case of the CERN for AI initiative this strategy partly backfired. There are two reasons for this. First, at about the same time as CAIRNE started to promote CERN for AI, a group of prominent AI researchers that is now known as “ELLIS” advocated for a networked AI research infrastructure modeled on the European Molecular Biology Laboratory. Given the EU’s limited funding for AI and the enormous price tag of any AI research infrastructure, ELLIS had no interest in supporting CAIRNE’s rival proposal. Second, critics of CERN for AI, among them ELLIS members, argued that the CERN label did not make sense in the context of AI because, unlike to the massive colliders at CERN, the computing power necessary for AI research does not need to be centralized. (Somewhat ironically, while criticizing CAIRNE’s framing, ELLIS failed to acknowledge that back in the day the European Molecular Biology Laboratory was promoted as a Conseil Européen de la Recherche Biologique [see Cassata, 2024]). Critics of CERN for AI further argued that using the CERN label for a large-scale AI research infrastructure was “misleading” because the current political circumstances are not comparable to those that eventually facilitated the establishment of CERN.

Tailor your message to your audience

In the future, scientists and scientific communities may therefore want to frame their proposals for large-scale, science infrastructures more strategically. Historical research has shown that promoting a “big science” project as a CERN of [insert relevant scientific discipline] does not necessarily have to backfire. What seems to play an important role is that those advocating for a large-scale science project in a research field that does not require big instruments chose a framing that manages to balance the best of both worlds: the unifying moment of centralization and the efficiency of decentralization (see Cassata, 2024). Such a framing is also likely to more strongly appeal to policymakers, as a central research facility supported by a network of research clusters would maximize the number of constituencies that see a return on investment from diplomacy in science.

References

Armbruster, Alexander. 2023. “Wir brauchen ein CERN für KI”. Frankfurter Allgemeine Zeitung, 16 October 2023.

Cassata, Francesco. 2024. A ‘Heavy Hammer to Crack a Small Nut'? The Creation of the European Molecular Biology Conference (EMBC), 1963–1970. Annals of Science: 1-48. https://doi.org/10.1080/00033790.2024.2351511.

Hoos, Holger. 2023. ‘AI made in Europe’ am Wendepunkt. Der Tagesspiegel, 7 July 2023.

Kelly, Éanna. 2021. Call for a ‘CERN for AI’ as Parliament Hears Warnings on Risk of Killing the Sector with Over-Regulation. Science Business, 25 March 2021.

Kohler, Kevin. 2024. CERN for AI: An Overview. https://machinocene.substack.com/p/cern-for-ai-an-overview. Accessed 13 February 2025.

Matthews, David. 2024. Call for the EU to Build Publicly Funded Cutting-Edge Artificial Intelligence. Science Business, 4 January 2024.

Royal Society, and AAAS. 2010. New Frontiers in Science Diplomacy. Navigating the Changing Balance of Power. The Royal Society (London). https://royalsociety.org/~/media/Royal_Society_Content/policy/publications/2010/4294969468.pdf.

Rüland, Anna-Lena. 2025. “We Need a CERN for AI”: Organized Scientific Interests and Agenda-Setting in European Science, Technology, and Innovation Policy. Minerva Online first. https://doi.org/https://doi.org/10.1007/s11024-024-09568-6.

Wulff Wold, Jacob. 2024. Von der Leyen Gives Nod to €100 Billion CERN for AI Proposal. Euroactiv, 25 July 2024.

[1] Science diplomacy, broadly understood as all activities at the intersection of foreign affairs and science, is commonly seen to be made up of three dimensions:

1. Science in diplomacy, i.e. using scientific evidence to inform foreign policy decisions,
2. Diplomacy for science, i.e. diplomacy facilitating international scientific cooperation, and
3. Science for diplomacy, i.e. using international science cooperation to improve international relations (Royal Society and AAAS, 2010).