The EU has held discussions with the Trump administration after the United States earlier this month restricted foreign use of Anthropic PBC's most advanced AI model. European Commission Executive Vice President Henna Virkkunen said the move, which led Anthropic to completely cut off access to its Fable 5 and Mythos 5 models, was discussed during a visit to Washington this week by a senior EU policy official.
"I myself have discussed this issue with the Trump administration," Virkkunen said in an interview in Milan on Thursday, "and we have been in communication with Anthropic."
Even as the EU works to achieve digital sovereignty and reduce dependence on U.S. technology, these negotiations highlight the EU's emphasis on access to cutting-edge AI systems that are powerful tools for identifying cybersecurity vulnerabilities. The U.S. decision to prevent foreign citizens from using the models over national security concerns shows that even allies are affected by Washington's policies.
Anthropic granted the EU Cybersecurity Agency access to Mythos in early June, making it the first EU agency to be granted access to this powerful tool. Previously, the company had only allowed access to the model to U.S. government agencies and some U.S. financial firms, technology companies and international organizations.