Europe’s current role in the global governance of AI
Some thoughts and open questions
What’s on Chiara’s mente recently? Many questions about Europe’s role in the global governance of advanced, beneficial AI.
As the title suggests, this is not another post on reflection or event ideas to facilitate reflection. I do think about other things sometimes.
This is a summary of my current thinking on Europe & AI, partly in the context of my work at Talos, adapted from a short memo I submitted to a conference. This post represents my views only.
Tl;dr
A stronger European AI policy community is essential to complement US and UK efforts in ensuring the safe development of advanced AI. The EU's unique position - with its large highly educated population, market power, and regulatory expertise - enables it to potentially act both as a geopolitical stabilizer and a pioneer in effective regulation through the Brussels effect. However, for Europe to fulfil either of these roles, it must build sovereign AI capacity.
Building this sovereign AI capacity isn't limited to building AI infrastructure. It requires developing both exceptional talent and creating strong networks that bridge technical and policy domains. A thriving European AI governance talent pipeline is therefore critical not just for fostering expertise, but for enabling Europe to meaningfully shape the development of advanced AI in alignment with societal wellbeing.
Europe must build sovereign AI capacity to play a meaningful role in global AI governance
Sovereign AI capacity means having sufficient local capabilities (e.g. regulatory expertise, AI infrastructure, venture capital) to understand, evaluate, regulate, and develop AI systems without being wholly reliant on foreign capabilities.
Currently, Europe remains highly dependent on digital technologies from outside the continent, with >80% of its digital products and services being imported. The barriers to European sovereign capacity are structural: fragmented digital markets; a lack of concentrated technical talent hubs; a risk-averse investment culture; weaker academia-industry collaboration compared to the US; brain drain; and uncoordinated policies on education, immigration, IP, taxes, and antitrust.
Building sovereign AI capacity would enable Europe to play two vital roles in global AI governance:
Act as a global stabilizing force. Europe could develop its distinct technological model, serving as a stabilizing third pole in global AI governance while setting regulatory standards through the Brussels Effect. This independent position would provide crucial checks and balances on both US tech companies and other global actors.
Strengthen the democratic alliance. Europe could become a more powerful partner in promoting democratic AI governance. That is, assuming the US doesn’t have the resources to police global AI development on its own, and Europe maintains the stance that "The US are Europe's most important partner and the EU should work closely with the US on AI as well as on other topics". Sovereign capacity doesn’t necessarily mean matching US capabilities entirely. At the very least, Europe should probably shoot for having the capacity to earn access to US AI innovations – it can’t be taken for granted that US-aligned allies get unfettered access.
Idk where else to put this but good to mention: the current ramping up of investment into European defense seems like a good opportunity to build sovereign AI capacity.
The European AI governance talent pool is essential for finding and developing the exceptional capabilities needed to address global challenges in the governance & development of advanced AI.
In the face of rapid tech development, you need top talent to understand what’s happening to society and how to handle those changes. The recent Draghi report similarly identifies the need for Europe to build cross-sectoral STEM talent who can bridge technical and strategic domains.
While the US leads in frontier AI development, addressing global AI challenges requires drawing on talent from across the world. Europe's large, well-educated population represents a vital pool of potential leaders and entrepreneurs in AI governance. By casting a wide net and nurturing diverse perspectives, we can better develop the broad range of capabilities needed to address complex AI challenges. This is where my work comes in. Our fellowship program leverages Europe's extensive talent base to identify and develop promising leaders in AI policy, graduating >40 fellows/year across two cohorts.
While each fellow brings unique strengths and perspectives, success is amplified through collaboration and mutual support. This is why Talos is also developing a broader alumni/network strategy to nurture a thriving community of EU AI policy professionals who become potential nodes in a powerful network of support and collaboration.
Some open questions
On Europe as a stabilizing “third pole”
The EU exercises strong central control over economic policy, but lacks similar authority over security and defense - areas where individual Member States maintain control. If it is true that AI becomes increasingly securitized, and European member states may want to develop frontier AI policy in-house instead, will the EU be able to translate its economic power into meaningful security influence?
On the Brussels Effect
Is the Brussels Effect an outdated model? Or, at least in AI/tech, is Europe too small to make everyone follow their lead? The EU has had historical success in exporting its standards, e.g. GDPR. This worked through market power: companies adopted EU standards for market access. But this was at a time when AI was not yet on the geopolitical radar. Europe can't rely solely on regulatory pressure to influence the governance of the world’s most transformative technology.
If we assume a "de jure" model were to hold, what regulation would we want other jurisdictions to adopt from the EU?
On Talos strategy & programs
Talos focuses on frontier AI governance talent development. Given limited resources, should Talos:
Perfect our existing programs (fellowship and alumni network)?
Launch a new fellowship stream (e.g. focused on technical people pivoting into AI governance)?
Some other activity that we aren't currently doing?
Thanks for writing this!
Re what Talos should do (though higher level than the questions you actually asked), my hunch at the moment is something like: I'd want Talos to be referred to as 'this is the place we want our next intern/hire from'. For this to happen, I'd imagine that the org would need to:
- build legitimacy toward the traditional EU policy ecosystem, get known and be taken seriously
- have a strong brand, something people easily understand, something that sticks
Though I can't tell what is Talos' current situation, nor how this is supposed to happen in your specific context.
Besides, maybe the winning strategy is something much more underground, actually. Or maybe it's both at the same time? Though public scrutiny usually undermines one's capability to act behind the scenes.
One takeaway from your post for me is that diverse strategies can be really important, and that we shouldn't but all our eggs in the same basket (eg focus on US AI governance). I don't know yet at which point of resource scarcity I'd change my mind though.