The Anthropic Year continues.
The artificial intelligence firm announced late Friday that it was disabling two of its newest language models just three days after launching them, subject to new export controls that had only been imposed by the US government that prevent any foreign nationals from using the models.
The Anthropic Year continues.
The artificial intelligence firm announced late Friday that it was disabling two of its newest language models just three days after launching them, subject to new export controls that had only been imposed by the US government that prevent any foreign nationals from using the models.
The sudden decision on export controls has exacerbated ongoing tensions between the Silicon Valley company and the Trump administration over national security concerns over the use of Anthropic AI models, and sets off the second high-profile dispute between them in less than a year.
Anthropic Executives is reported they held talks with administration officials in Washington on Monday to try to find a solution, but no agreement has yet been reached. Neither side responded to requests for comment on what was discussed.
Here’s what you need to know.
What happened?
Anthropic received an export control order from the US government at 5:21 pm EDT on Friday that cited “national security authorities” to prevent any foreign nationals (including non-Anthropic US employees) from using the company’s Claude Fable 5 and Claude Mythos 5 models, the company said in a statement. statement. It added that the directive “did not specifically address its national security concerns.”
Anthropic said the government had voiced concerns about vulnerabilities in models that could allow users to “jailbreak” or bypass their security measures. However, the company claimed that the vulnerability “look(ed) simple” and could also be exposed by other publicly available frameworks, such as OpenAI’s ChatGPT.
The White House did not immediately respond Foreign Policyrequest for comments on its specific issues, which were is reported flagged for administration by Amazon. Amazon shared its problems with Anthropic before reporting them to the US government, according to a person familiar with the matter.
The tech giant is a major investor and partner in Anthropic as well as a leading provider of cloud computing services to the US government. “As a leading cloud provider serving a large number of private and public sector customers, it is not uncommon for governments to seek our advice on potential security risks,” an Amazon spokesperson said in a statement.. “As they occur, we do not share details of these discussions.”
Instead of trying to restrict access to only US users, Anthropic chose to disable the models entirely. “The main impact of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance,” the company said, noting that access to the company’s other models will remain unrestricted.
Wait, why does “Mythos” seem familiar?
Anthropic he raised the alarm back in April when it launched an early version of one of its new models, known as the Claude Mythos Preview. At the time, the company warned that the model was too powerful for full public release because it had demonstrated a pre-existing ability to detect cybersecurity flaws in computer systems—thus theoretically making it easier for hackers to infiltrate. Instead, Anthropic released a few dozen (no growing up) companies and government agencies to access Mythos so they can use it in a closed environment and integrate their systems before it is released more widely.
That wide release came last Tuesday, when Anthropic he announced Claude Fable 5-model that it said was built from Mythos but with additional protections and restrictions for certain questions that made it “safe for general use.” It also announced Claude Mythos 5, a slightly modified version of the model without some of those protections that can only be used by “a small group of network advocates and infrastructure providers.”
That massive release was short-lived, however, with the Trump administration putting the kibosh on both brands just three days later.
This isn’t the first time the Trump administration has had trouble with Anthropic, is it?
Nope. Back in February, Anthropic entered a great struggle and the US Department of Defense on the deployment of the company’s models to the US military. Anthropic demanded that the Pentagon commit to preventing the designs from being used for autonomous weapons or internal surveillance.
US Defense Secretary Pete Hegseth responded by giving Anthropic a decision: Give the military unfettered access to Anthropic’s technology or be designated a “supply chain risk” — a label that would prevent Anthropic from serving any customers working with the Pentagon.
Hegseth later he made good on that threatand cause the company sue the US government in response. The case is still ongoing, even as Anthropic and the Trump administration are waging a new battle this week.
“Three months ago, (the Department of Defense) kicked @AnthropicAI out of our building—forever,” Hegseth. he wrote in a post on X on Saturday. “Every day that passes proves why that was the right move.”
What does this mean for the rest of the world?
The Anthropic-Trump showdown has further strained the already rocky transatlantic relationship, as European countries scramble to carve out which would allow them to continue using America’s most advanced AI designs.
European diplomats reportedly held talks with US Commerce Secretary Howard Lutnick on the sidelines of the G-7 summit in Évian-les-Bains, France, this week, seeking to discuss the availability of a “trusted partner”. Financial Times information. Britain has also requested a special allocation, although a Trump administration official he told it of New York Post that exempting any country from export controls, including allies, would be “absolutely illogical.”
Canadian Prime Minister Mark Carney warned that US export controls show the dangers of global dependence on US technologies. “The situation we have right now with Mythos and Fable is something that can happen with over-reliance on certain models,” Carney. he said during a visit to Ireland before the G-7 meeting. “Nobody has done anything wrong in that situation,” he added. “But we will be doing something wrong if we just accept this, don’t take the lesson, don’t build and differentiate.”
