Anthropic mentioned on Friday it would “abruptly disable” its most superior AI fashions for all customers after the US authorities ordered it to droop entry to the fashions for overseas nationals, citing nationwide safety considerations.
The corporate acquired the export management directive to droop entry to Fable 5 and Mythos 5 for all overseas nationals, with out being given particular particulars of its nationwide safety concern, Anthropic mentioned in a press release.
It’s Anthropic’s understanding that the federal government believes there’s a technique of bypassing, or “jailbreaking,” a safeguard that might forestall Fable 5 from being utilized in figuring out software program vulnerabilities, the corporate mentioned.
The order comes simply as a earlier dispute between Trump administration officers and IPO-bound Anthropic confirmed indicators of easing throughout components of the US authorities.
Anthropic’s relationship with the federal government ruptured this yr after it refused to permit the US army to make use of its AI fashions for home surveillance and absolutely autonomous weapons programs. The federal government responded by placing Anthropic on a provide chain blacklist, set to take impact later within the yr.
The motion additionally marks a significant escalation of US efforts to halt overseas adversaries’ AI capabilities. For years, US export controls have targeted on the chips and instruments that energy AI slightly than on proscribing overseas entry to AI itself.
Anthropic mentioned the federal government has given it solely “verbal proof of a possible slender, non-universal jailbreak”.
“We disagree that the discovering of a slender potential jailbreak ought to be trigger for recalling a industrial mannequin deployed to tons of of thousands and thousands of individuals,” the corporate mentioned.
The federal government directive and Anthropic’s response spotlight rising stress between AI builders and regulators over methods to assess dangers from so-called “jailbreaks,” or strategies used to bypass mannequin safeguards.
As just lately as Wednesday, Anthropic had referred to as for higher US oversight of AI, together with the power to dam fashions with unacceptable dangers. It mentioned, nevertheless, the federal government motion on Friday didn’t observe rules of truthful and fact-based regulation.
The Pentagon’s chief data officer, Kirsten Davies, mentioned in a publish on X that the Protection Division supported prioritizing nationwide safety.
“Some issues are merely extra essential than income cycles, clickbait, and pre-IPO valuation. America First. All the time,” Davies mentioned.
Anthropic confidentially filed for a US IPO final month, edging forward of rival OpenAI within the race to succeed in public markets.
SOPHISTICATED CYBERATTACKS
Earlier this week, Anthropic rolled out an AI mannequin named Claude Fable 5, representing a brand new tier of functionality it calls “Mythos-class.” The mannequin is accompanied by guardrails barring its use in dangerous areas corresponding to cybersecurity, which some customers have complained are “overly broad,” Anthropic mentioned.
Consultants have mentioned that Mythos fashions, within the flawed palms, might dramatically speed up subtle cyberattacks, significantly in sectors corresponding to banking that depend on advanced, interconnected, and infrequently decades-old expertise programs.
Anthropic mentioned it had labored with the US authorities, amongst others, on security forward of the Fable launch and that fashions from rival AI suppliers confirmed an identical capability to unearth minor bugs in code.
“The web impact of this order is that we should abruptly disable Fable 5 and Mythos 5 for all our clients to make sure compliance. Entry to all different Anthropic fashions is not going to be affected,” Anthropic mentioned.
Anthropic mentioned that it believed there was a “misunderstanding” and that it’s working to revive entry to the fashions as quickly as doable.
“If this customary was utilized throughout the business, we consider it will basically halt all new mannequin deployments for all frontier mannequin suppliers,” the corporate mentioned.
Amazon’s cloud unit AWS mentioned late on Friday that Anthropic has requested it to revoke entry to the fashions for “all customers in all areas.”
A US official confirmed that the Commerce Division had issued an export management directive to droop all entry to Fable 5 and Mythos 5 by overseas nationals.
Dean Ball, a former White Home official who contributed to the AI Motion Plan the administration issued in the summertime of 2025, mentioned in a publish on X that the order suggests all “non-Individuals” can be restricted from utilizing Anthropic’s newest fashions, together with these based mostly within the US.
“This implies you must anticipate to should show your citizenship to make use of Anthropic fashions,” Ball mentioned.
A number of key Anthropic personnel, together with co-founder Chris Olah, AI researcher Andrej Karpathy and thinker Amanda Askell, had been born exterior the US. Reuters was unable to find out their citizenship standing, and an Anthropic spokesperson declined to touch upon whether or not such employees would lose AI mannequin entry.
– Ends