Artificial intelligence company Anthropic and the Trump administration are sparring again.
Tensions between the San Francisco company and the Trump administration escalated over the weekend after the U.S. government ordered Anthropic to cut off foreign access to its most powerful AI system.
The government’s order poses a challenge to Anthropic’s business and branding, which focuses on AI safety. The company said it has made the latest models inaccessible to all customers in order to comply with the directive. Customers will still have access to other Anthropic models.
The unusual move comes days after Anthropic released AI models known as Fable 5 and Mythos 5. Anthropic limited the release of Mythos 5, which can be used to patch software vulnerabilities, to a “small group of cyber defenders and infrastructure providers” over concerns that hackers could misuse the tool to break into computer networks.
The company says Fable 5 excels in software engineering, scientific research, and other complex tasks. Currently, Anthropic is facing security concerns regarding the release of its models.
Anthropic is competing with other big tech companies to advance AI. Anthropic has filed to go public, valuing it at nearly $1 trillion, outpacing rival OpenAI.
Here’s what you need to know about the latest high-stakes public fight.
Why did the US government order Anthropic to cut off foreign access to its latest AI models?
Anthropic said in a blog post Friday that the government had raised national security concerns but did not provide details.
“Our understanding is that the government has realized a way to circumvent, or ‘jailbreak’, Fable 5,” the company said in a post. Anthropic said safeguards are in place to prevent AI systems from being exploited for cybersecurity tasks, and no tester can broadly circumvent these guardrails.
The company added that it worked with the US government ahead of the model’s launch.
However, David Sachs, a venture capitalist who served as a crypto and AI advisor in the Trump administration, accused Anthropic of prioritizing “continuation of the consumer model over security.”
He said the Trump administration hopes to resolve the security issue and then restore access to Fable.
“Administrators want all of this to happen as soon as possible. We are frankly perplexed by Anthropic’s unwillingness to comply with the safety requirements they have previously stated are their top priority,” he wrote.
Over the weekend, Trump officials also attacked Antropic, intensifying the public feud.
“Three months ago, @DeptofWar permanently removed @AnthropicAI from our buildings,” Defense Secretary Pete Hegseth told X. “Every day that passes proves why it was the right thing to do.”
Why did the government issue this order?
The president’s order was triggered by conversations Amazon CEO Andy Jassy had with Trump administration officials, including Treasury Secretary Scott Bessent, according to people familiar with the matter. wall street journal.
According to the report, Amazon researchers were able to obtain Anthropic’s Fable 5 model and provide information that could be used to assist cyber attackers. An Amazon spokesperson told the press: “It is not uncommon for governments to seek advice from us regarding potential security risks.”
Amazon has also invested billions in Anthropic, providing chips to train its AI models.
Anthropic said it was following government orders, but disagreed that the findings “should be cause for a recall of commercial models that have been deployed to hundreds of millions of people.”
“We believe that if this standard were applied industry-wide, it would effectively halt all new model deployments for all Frontier model providers,” the company said.
Have Antropics and the Trump administration sparred before?
Yes, but it involved another issue.
Anthropic and the Department of Defense were in a dispute over the language of a $200 million government contract with the company. Anthropic wanted further assurances that its technology would not be used to surveil American citizens or for autonomous weapons. But the Pentagon characterized the pushback as Anthropic’s attempt to “seize veto power over U.S. military operational decisions.”
In February, President Trump called Anthropic “radical leftist” and “woke” and ordered federal agencies to stop using Anthropic’s technology.
Anthropic is suing the Trump administration to revoke the company’s “supply chain risk” designation, accusing the government of illegally retaliating against the company.
What’s next?
Anthropic and the White House are trying to resolve concerns raised about the AI model.
Anthropic’s senior technical staff will travel to Washington, D.C., on Monday to meet with White House officials, people told the Times.
Trump administration officials said the meeting will take place at the Commerce Department. The ministry did not immediately respond to a request for comment about the meeting.
