Anthropic claims it is Disabling the two AI models that were launched this week Claude Fable 5 and Mythos 5It is complying with the export control directive received from US Government on Friday afternoon citing concerns about national security.
The latest incident between the two sides is marked by this unprecedented event. Anthropic and the Trump administration. The company claims that the order required it to restrict access to “any foreign national, whether inside or outside the United States, including foreign national Anthropic employees,” To ensure compliance, it has blocked all access to its clients.
Anthropic, a company that Trump’s Department of Defense branded as a “supply chain riskThe Claude maker wanted to set red lines about how US military personnel could use their technology. Anthropic was effectively banned from using its technology by government agencies, contractors and other companies. Anthropic’s response was filing lawsuits Trump’s administration is a threat to democracy.
Anthropic released its public release on Tuesday. Claude Fable 5Anthropic has released a new version of its Mythos AI, which includes safeguards to prevent the model from answering any questions regarding cybersecurity, biology or chemistry. Anthropic, in conjunction with the US federal government, conducted the research prior to its public release. Mythos Preview AI model The limited release in April. This was done to provide companies with an opportunity to utilize its powerful cybersecurity capabilities to improve their defensesThe technology is being exploited to create powerful hacking tool by malicious actors.
It is not a blog post Anthropic reports that it has received an official letter from the US at 5pm Eastern Time on Friday. “The letter did not provide specific details of its national security concern,” Anthropic Writer
“Our understanding is that the government believes it has become aware of a method of bypassing, or ‘jailbreaking’ Fable 5,” The company has added. “We reviewed a demonstration of this specific technique being used to identify a small number of previously known, minor vulnerabilities. These vulnerabilities all appear relatively simple, and we have found that other publicly-available models are able to discover them as well without requiring a bypass.”
In a recent blog, the company argued it had implemented robust safeguards in order to limit the chances of Claude Fable 5 becoming misused. Anthropic said that the US government’s jailbreak of Claude Fable was narrow and wouldn’t have rendered an attacker any more dangerous than if he had used a different AI model.
“To date, the government has only given us verbal evidence of a potential narrow, non-universal jailbreak, which essentially consists of asking the model to read a specific codebase and fix any software flaws,” The company stated in a blogpost. “Our understanding is that one potential jailbreak was shared with the government.”
WIRED’s requests for comments were not responded to immediately by the White House or US Commerce Department.
Dario Amodei, CEO of Anthropic, said earlier in the week, that the company and he support an open, transparent, fair and structured government process to block unsafe AI models. Anthropic’s Friday blog post argued for a fair, structured and transparent government process that would block the release of unsafe AI models. “this action does not adhere to those principles.”

