Don’t Gamble with Europe’s AI Future

Europe must take control of its AI future—prioritising sovereignty, resilience, and accountability over dependence on foreign technology. The claim that Europe can only be an AI consumer must be laid to rest once and for all. Safe, responsible, and efficient AI solutions on local infrastructure, with open-source components, are both necessary - and already available. Europe should invest in its own AI platforms to protect welfare, competitiveness, and our societal values.

Who should hold the helm as AI shapes Europe’s welfare, economy, and democracy?

The answer should be clear: Europe itself.

Sovereignty is not a technical detail. It is a defining strategic question and should be treated as such: Can we afford to build our future welfare and competitiveness on technologies we do not control?

Some will argue that ensuring AI sovereignty in Europe is too expensive, too slow, and too complex. That we must choose between speed of innovation and sovereignty. That Europe cannot compete with American or Chinese tech companies. That open-source language models will never be good enough. That small countries can only be users, not developers, of AI. These arguments may sometimes sound pragmatic, but they are both wrong and dangerous.

We have never had a greater need for control, transparency, and accountability in how we approach and apply AI technology. When critical AI technology and infrastructure sit beyond our direct reach, we lose influence—as we saw when an ICC prosecutor’s account was shut down without warning. I hope we can all agree that geopolitical power struggles must never define the terms of our direct use of technology—or, indirectly, our welfare.

Europe faces a major demographic challenge: fewer hands must lift more tasks. To meet this challenge, AI is not optional but an absolute necessity—especially in the public sector. However, the pressure to implement new welfare technology must not become an excuse to compromise on core principles, security, and transparency.

The good news is that we do not have to choose between efficiency and security. Public authorities and enterprises are already demonstrating that it is possible to roll out AI both efficiently and safely on European AI platforms, operated on local infrastructure and based on open-source language models—an approach that matches the best of the tech giants and ensures AI technology works responsibly for us.

This is less about where AI technology is developed and more about who controls it. Ultimately, it is about keeping security, accountability, and democratic anchoring at the centre.

Now, together, we must dispel the myth that Europe cannot deliver advanced AI—and adopt more of the safe, responsible European AI solutions.

We are ready.

‍