WASHINGTON \u2014 The administration\u2019s latest crackdown on Anthropic further complicates the AI titan\u2019s already ambivalent relationship with the Pentagon, which had been simultaneously banning most Anthropic products while carving out an exemption for its latest model, Mythos, analysts told Breaking Defense.<\/p>\n
On Friday, the Commerce Department classified Anthropic\u2019s Mythos\/Fable 5 as cyber weapons subject to export controls \u2014 which makes it illegal for Anthropic to provide them to any foreign national, even its own employees. The reasoning: Amazon<\/a> reportedly<\/a> told the government it had found a way around safeguards Anthropic put in place to prevent misuse of the bug-hunting capabilities<\/a> that made Mythos so attractive to hackers, including the NSA<\/a>.<\/p>\n Anthropic, in turn, shut down public access to all users<\/a>, including US citizens, late on Friday, arguing it had no way to exclude only foreigners. The company says its safeguards still largely hold and the breach can be easily repaired. As of Monday, a delegation of senior Anthropic executives has reportedly<\/a> flown to DC to try to resolve the dispute.<\/p>\n Defense Secretary Pete Hegseth seized on the Commerce ruling as new ammunition for his highly public criticism of Anthropic as irresponsible and unreliable.<\/p>\n \u201cThree months ago, @DeptofWar kicked @AnthropicAI out of our building\u2014forever,\u201d he posted on X.com Saturday<\/a>, using the informal name for the Department. \u201cEvery passing day proves why that was the right move.\u201d<\/p>\n But, as an X.com \u201ccommunity note\u201d quickly pointed out, \u201cthis official statement is not accurate or truthful.\u201d<\/p>\n First of all, when Hegseth and President Donald Trump issued a ban on federal agencies using Anthropic<\/a> back at the end of February, it didn\u2019t take immediate effect. Instead, Trump allowed<\/a> \u201ca Six Month phase out period,\u201d tacitly acknowledging how dependent some defense agencies had become<\/a> on Anthropic\u2019s earlier AI, Claude, for rapid-fire coding and even, reportedly, planning against Venezuela<\/a> and Iran<\/a>.<\/p>\n Meanwhile, Anthropic had begun testing Mythos, a new model that is widely reported to have superhuman ability to find bugs and security vulnerabilities in code \u2014 making it immensely useful for both cyber defense and cyber attack. The secretive National Security Agency reportedly<\/a> started using Mythos in offensive operations at the same time as the administration was trying to phase out Anthropic overall.<\/p>\n Just last week, a sweeping National Security Presidential Memorandum on AI issues<\/a> formalized this ambivalent stance. Without naming Anthropic, NSPM-11<\/a> orders agencies to terminate contracts with AI companies deemed uncooperative. But the same paragraph of the memo also establishes a (strictly limited) waiver process to continue such contracts if \u201cnecessary to responsibly steward United States national security.\u201d<\/p>\n That\u2019s a delicate wire to walk amidst a political windstorm.<\/p>\n Expert Warnings<\/p>\n The \u201cpreview\u201d version of Mythos in use by NSA is not believed to be affected by the shutdown of the public version, according to both expert interviews and published reports<\/a>. Nevertheless, the latest blow-up between Anthropic and the administration could well mean blow-back for the Pentagon, including in two ongoing lawsuits<\/a> filed by Anthropic.<\/p>\n The Commerce Department declaration does not change the legal fundamentals of either case, explained Jessica Tillipman<\/a>, associate dean at the George Washington University Law School, because they challenge the specific \u201csupply chain risk\u201d designations against the company, not export restrictions. But Commerce\u2019s actions still provide legal ammunition in the lawsuits, she told Breaking Defense \u2014 potentially to either side.<\/p>\n \u201cBoth of those [lawsuits] are under different authorities than the export control directive,\u201d Tillipman said. \u201cBut I do think both sides may use this to support their cases \u2014 the government using this to suggest Anthropic is risky, while Anthropic may claim it is being singled out.\u201d<\/p>\n In the court of public opinion, meanwhile, imposing export controls on Anthropic\u2019s new AI models, on top of the previous ban and the administration\u2019s ferocious rhetoric<\/a>, could easily be seen as politically motivated persecution, warned Jack Shanahan<\/a>, the former head of the military\u2019s groundbreaking Project Maven and its Joint AI Center. That would chill relations between the administration and the wider tech sector it has been wooing, especially for military use, the retired three-star general told Breaking Defense.<\/p>\n \u201cEven if I take, at face value, the claim that Amazon found a way to jailbreak the latest Anthropic models and hence this is a legitimate national security concern, it still smacks of an ongoing vendetta against Anthropic and Dario personally,\u201d Shanahan said, referring to Anthropic CEO Dario Amodei. \u201cEven the companies who are in the administration\u2019s good graces for the time being will see the ramifications of this latest order: the ability to do major damage to their business models at the drop of the hat.\u201d<\/p>\n The shutdown of Mythos\/Fable 5 was truly sudden: \u201cI was using it at five o\u2019clock and then, at eight o\u2019clock, it was offline,\u201d said Charlie Bullock<\/a>, of the Institute for Law & AI.<\/p>\n