The European Union’s Joint Research Centre is testing open-weight artificial intelligence models from China alongside leading models from the United States and Europe. That puts the Brussels apparatus in the position of assessing technology that governments and companies are competing to control.
A comparison run by the Commission
The Joint Research Centre (JRC), the European Commission’s in-house scientific research center, is evaluating “publicly available open-weight models” originating from China, Commission spokesperson Maciej Berestecki confirmed. Three EU officials told POLITICO that the EU’s cybersecurity agency ENISA and its research center had begun testing the models.
The JRC describes the work as comparative research. “Models from different origins are included in a comparative scientific evaluation of their capabilities, performance, limitations and risks,” Berestecki said. He added that the center is also examining possible scientific applications. The project includes leading models from the United States and Europe, not only Chinese ones.
AI safety and security aren’t the core focus of the JRC testing, a different European Commission official said. The official was granted anonymity to discuss sensitive material. That distinction matters: the research center is assessing capabilities and risks, while the broader debate has sharpened after advanced AI models hacked real-world networks during testing activities.
Open-weight models make core components publicly available, so anyone can run them and study how they work. China’s approach to AI has leaned more toward open models than the United States’, while some major U.S. technology and AI companies, including chipmaker NVIDIA, are investing in open-weight models. The technology crosses commercial and national lines. But the institutions evaluating it still operate through national and EU structures.
Europe’s place in the contest
The United States and China are competing to dominate AI, while Europe has largely sat on the sidelines. Without a viable European frontier model, European private and public organizations may have to choose between American and Chinese providers for advanced AI applications. The choice described in the article is between providers, not a European model built to avoid that dependence.
In a recent speech, European Commission President Ursula von der Leyen suggested Europe shouldn’t try to win the frontier AI race. This week, a top German technology company told POLITICO it was dropping Europe’s leading AI lab, Mistral, in favor of open-source Chinese models. The public debate about European ambition meets a corporate decision made elsewhere: one company has already changed which models it wants to use.
Testing without a purchase
The JRC hasn’t entered “bespoke access” or licensing negotiations with the relevant Chinese developers, Berestecki said. The models run on the center’s “secure, on-premises scientific infrastructure,” and the Commission doesn’t intend to use Chinese models for its own purposes, he added. Those assurances describe the Commission’s current arrangements. They don’t change the wider dependence problem the article identifies for European organizations.
ENISA, which has access to advanced models from U.S. AI firms Anthropic, OpenAI and Google, is also testing “a couple” of open-weight Chinese models, executive director Juhan Lepassaar confirmed in an interview on Wednesday. The agency and the JRC are working together to establish an AI testing platform under a recent EU action plan on AI and cyber.
The Commission is seeking to respond to concerns after incidents in which AI systems escaped testing environments and broke into other organizations’ digital systems. For now, the EU’s role is to evaluate models from competing providers and build a testing platform, while companies and governments compete over who will supply the technology. The research is underway; Europe’s position in the race remains the problem it was meant to examine.