US Export Controls on Anthropic's AI Models Tied to Safety Fix, Says David Sacks

US Govt's export controls on Anthropic tied to safety fix, says US President Advisory council member David Sacks

New Delhi, June 14

After this development, David Sacks the Co-Chair of the US President's Council of Advisors on Science and Technology in a social media post on Saturday said that the Trump administration is pushing Anthropic to remediate a reported jailbreak in its Fable AI model and hopes the export control issued over national security concerns can be lifted quickly.

Anthropic had rolled out its latest AI model, Claude Fable 5, earlier this week. The company had acknowledged in its June 9 release statement stated that "releasing a model this capable comes with risks," and said it had introduced safeguards that would block its use for some topics.

However, the US government appeared to be claiming that a "jailbreak" exists that allows users to work around those safeguards.

Anthropic, in a statement on June 12 had stated, "The US government, citing national security authorities, has issued an export control directive 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."

"The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Anthropic models will not be affected," it said.

The statement said that, "We received the directive from the government today at 5:21pm (ET). The letter did not provide specific details of its national security concern. Our understanding is that the government believes it has become aware of a method of bypassing, or "jailbreaking" Fable 5."

"Jaibreaking" refers to a workaround that could let the model be used to find software vulnerabilities, or access other illegal information such as making weapons or explosives.

"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," Anthropic said

The statement read, "We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible."

Sacks claimed a "highly credible trusted partner of both Anthropic and the USG who was testing Fable came forward with a jailbreak of those guardrails," and that the administration asked Anthropic CEO Dario Amodei to fix it or de-deploy the model. According to Sacks, "Dario refused."

Sacks' post took issue with Anthropic's public response. "In their blog post, Anthropic defended its decision by saying the jailbreak isn't serious. That is not what the trusted partner and the USG believe," Sacks said.

The Co-Chair of the US President's Council of Advisors argued this was inconsistent with Anthropic's branding as an AI safety company, noting the firm had previously promoted Mythos as a "cyberweapon" that needed regulation. "It's difficult to fathom how they could claim a jailbreak allowing operability of a cyber weapon could be defined as not 'serious,'" he wrote.

Sacks said the export control was issued "reluctantly" and expressed surprise that Anthropic had not cooperated with what he called a "reasonable safety request."

He added the administration values Anthropic's technical capabilities and believes the issue "should be easily resolved," with the ball "in Anthropic's court." The stated hope is that Anthropic remediates the safety issue, the export control is lifted, and Fable returns to general release "as soon as possible."

He also pushed back on speculation linking the action to earlier disputes and said, "Those trying to misdirect and tie this action to the prior Anthropic issues are wrong. The Administration values Anthropic's technical capabilities and feels that this issue, while serious, should be easily resolved."

Meanwhile, Pentagon's Chief Information Officer, Kirsten Davies also took to X to double down on the administration's aggressive stance stating they "fully support" the decision of President of the united states and the Secretary of War in "prioritizing national security and the security of our warfighters, DIB partners, critical infrastructure, international partners and allies."

Some things are simply more important than revenue cycles, clickbait, and pre-IPO valuation. America First. Always. "Some things are simply more important than revenue cycles, clickbait, and pre-IPO valuation. America First. Always." Davies posted.

— ANI