EU cyber agency ENISA and research centre JRC test Chinese open AI models

The EU's cyber agency ENISA and the Commission's Joint Research Centre have begun testing open-weight AI models from China, Politico reported. The models run on the bodies' own secure computers, no special access was requested from Chinese developers, and the Commission does not plan to use them in daily work. The two bodies are also building a joint AI testing platform due by the end of the year.

• ENISA is testing a couple of Chinese open-weight models, JRC as part of a wider study
• JRC compares Chinese, American and European systems on its own secure computers
• The Commission does not plan to use the tested models in its daily work
• A joint AI testing platform for energy, finance and health is due by year-end