European Commission officials are planning to meet with Anthropic PBC to seek more information on the company’s Mythos model and request making the groundbreaking artificial intelligence tool available to the bloc, according to people familiar with the matter.
Officials from the commission were set to visit San Francisco on Thursday for meetings with representatives from the AI startup to learn more about Mythos’s capabilities, according to one of the people, all of whom asked not to be identified because the plans are still taking shape.
Any final decision on Mythos access for the EU would likely require approval from the US government, according to one of the people. For now, there’s no confirmed date for a meeting between representatives from the bloc and Trump administration officials, the person said.
The European Union has been pushing for access to Mythos since Anthropic disclosed last month that the new model was extraordinarily adept at finding network vulnerabilities and could pose a major cybersecurity risk. Anthropic has limited its availability to a handful of large tech and Wall Street companies through an initiative known as Project Glasswing, amid broader global alarm about the new threats Mythos could pose to critical systems.
The meetings in San Francisco would mark the latest step in weeks of negotiations between Anthropic and the bloc over Mythos that so far have failed to yield a breakthrough. In those discussions, the EU has received "extensive information” from the company on the risks that the model could pose, commission spokesperson Thomas Regnier said Thursday at a press briefing in Brussels.
A spokesperson for the commission didn’t respond to questions on the meeting plans but said that the EU generally is seeking to ramp up discussions with the US on advanced AI models. An Anthropic spokesperson said the visit is part of the company’s ongoing engagement with the EU, allied democracies and international institutions on safe AI deployment and Project Glasswing.
In the US, Trump administration officials have pressed to allow federal agencies to use Mythos to test their networks for security flaws, and the National Security Agency has already been using it. But White House officials recently rejected Anthropic’s plans to distribute Mythos to several dozen additional companies and organisations, citing security concerns.
Anthropic is urgently working to expand Mythos access in Europe, though the outcome of that effort remains unclear, another person familiar with the matter said.
Finance ministers across Europe have been demanding access for European banks and companies, so they can check their systems for vulnerabilities. Following a May 4 meeting of Eurogroup finance ministers, Eurogroup President Kyriakos Pierrakakis acknowledged the importance of working with the US in addressing the new challenges posed by technologies like Mythos.
"I don’t think we have the luxury of not trying to establish channels of communication with the US,” Pierrakakis said. "The challenge here is that technologies like AI necessitate international governance frameworks at a moment when multilateralism is challenged.” – Bloomberg
