US lifts restrictions on Anthropic’s most powerful AI models

SHEERA FRENKEL AND ANA SWANSON

SAN FRANCISCO: The Commerce Department lifted restrictions on all of Anthropic’s artificial intelligence models Tuesday, allowing the company to bring its most powerful AI technologies back online, according to a letter from the government to the company that was viewed by The New York Times.

In the letter, Commerce Secretary Howard Lutnick told Anthropic that it would no longer need a license for exports or in-country transfers of its Claude Mythos and Claude Fable AI models, reversing an order issued June 12.

“Anthropic has taken steps in close coordination with the US government to address the risks associated with Claude Mythos 5 and Claude Fable 5,” Lutnick wrote.

The move was the latest chapter in Anthropic’s often prickly relationship with the federal government and is set to de-escalate their feud, paving the way for the company to largely return to business as usual. Citing national security, the Commerce Department had initially ordered Anthropic to suspend access to its latest AI models for all foreign nationals.

Those controls raised questions about how hands-on the Trump administration might be with the fast-growing artificial intelligence industry. They also could have caused significant problems for technology development at Anthropic, which along with OpenAI is among the world’s most influential AI companies. Mythos and Fable, which are powerful at identifying security flaws in software, are considered a significant improvement on earlier AI models.

The restrictions were the second time the Trump administration had targeted Anthropic, which is based in San Francisco. In March, after tense talks over how Anthropic’s AI could be used in warfare, the Pentagon labeled Anthropic an unacceptable supply chain risk, potentially limiting its use by federal agencies. Anthropic has sued the federal government over that designation. Lutnick’s lifted the export controls after weeks of deliberations between the Commerce Department and the company over safeguards built into its AI models. Last week, he announced a partial deal with Anthropic to bring back access to its Mythos model for some clients.

In Tuesday’s letter, Lutnick said Anthropic had agreed to take steps including proactively detecting and addressing security risks for the AI models, working with the US government on protocols and standards for these and future models, and informing the government of any malicious activity.

Lutnick added that the Commerce Department “reserves the right to reevaluate the decisions made in this letter” if Anthropic fails to adhere to its commitments or circumstances change.

In a statement posted on social media, Anthropic said it would begin restoring access to its AI models Wednesday.

“We’re grateful to our users for their patience, and to everyone who worked with us on redeploying the models,” the company said.

In a social media post Tuesday, Susie Wiles, the White House chief of staff, expressed her “gratitude to companies across industries who continue to work closely with the White House” to carry out Trump’s executive order on AI.

“Our shared priority remains: get the best tech deployed as quickly and safely as possible,” she said.