Anthropic suspends AI models after US government directive
Auf einen Blick
- Anthropic has suspended access to its Fable 5 and Mythos 5 AI models following a US government directive citing national security concerns.
- The directive, received Friday, instructs the company to block all foreign nationals from accessing the models, including its own employees.
- Anthropic stated that the government is concerned about a potential "jailbreak" method that could bypass the models' safeguards, though Anthropic disagrees with the severity of the threat.
KI-generierte Zusammenfassung
Anthropic said it suspended access to its Fable 5 and Mythos 5 AI models after receiving a US government export control directive citing national security concerns.
In a statement posted Friday, Anthropic said it received the directive at 5:21 pm ET, instructing it to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
Anthropic abruptly disabled the models for all users in order to ensure compliance. It said all other Anthropic models, such as Opus 4.8, are not affected.
"We are complying with the government’s legal directive and are removing access to Fable 5 and Mythos 5 for all users," the firm said.
The directive comes just days after Anthropic released Fable 5 and Mythos 5, two powerful AI models built on top of Mythos Preview, a general-purpose language model that the company previously said had found thousands of vulnerabilities in critical software.
Anthropic said the government did not provide specific details about the alleged threat, but said it believes authorities are concerned about a possible “jailbreak” method capable of bypassing Fable 5's safeguards.
Related: AI researcher claims he's already bypassed Anthropic's Fable 5 guardrails
“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,” said Anthropic.
A non-universal jailbreak is a far lower threat than a "universal jailbreak," a method to broadly bypass a model's safeguards, it explained.
"We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people. If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers," it added.
Anthropic said it believes the government order is a result of a misunderstanding and is working to restore access for users as soon as possible.
