Trump Says He No Longer Views Anthropic As A National Security Threat

Authored by Jacob Burg via The Epoch Times,

President Donald Trump said he no longer views the artificial intelligence (AI) giant Anthropic as a national security threat.

"We have a situation with Anthropic, and we didn't like what they were doing, and so far I think they behaved very responsibly to our request," Trump told Axios's Marc Caputo in an interview that aired on June 19.

Caputo then asked Trump if he still viewed Anthropic and its CEO, Dario Amodei, as a threat to national security.

"Well, not now, but a week ago, maybe," Trump said, describing a meeting with Amodei at the G7 summit this week that influenced his change of heart.

"[Amodei] responded to us very quickly, because you know it's tremendous liability," Trump said. "You know, you can't play games with it. And he responded very responsibly."

The comments come one week after the Trump administration directed Anthropic to shut down foreign nationals' access to its new Claude Fable 5 and Mythos 5 models, which resulted in the company suspending access to all users entirely.

Issued on June 12, the directive from U.S. officials did not include specific details of potential security threats or concerns, according to Anthropic.

"Our understanding is that the government believes it has become aware of a method of bypassing, or 'jailbreaking' Fable 5," the company said in a statement at the time.

Trump told Axios that one of the company's competitors "turned Anthropic in" and raised alarms over the new models.

"They didn't like what [Anthropic was] doing. They were very concerned," he said.

Anthropic did not respond to a request for comment by publication time.

When Anthropic released Fable 5 to the general public on June 9, it said the model had exceeded the capabilities of "any model we've ever made generally available."

"It is state-of-the-art on nearly all tested benchmarks of AI capability, showing exceptional performance in software engineering, knowledge work, vision, scientific research, and many other areas," Anthropic stated.

As a "Mythos-class" model, Fable 5 is essentially as strong as Mythos, but with key safety features to make it safe for public use.

In the same announcement, Anthropic made Mythos 5 available to a small group of cyberdefenders and infrastructure providers. But after the company's decision to suspend access to all users on June 12, both models are currently unavailable.

Two days before Anthropic pulled access, Amodei wrote on X that he believes frontier models such as Mythos 5 "should face mandatory third-party testing for cyber, bio, and autonomy risks - with the power to block or revoke deployment of models that pose catastrophic risk."

Trump signed an executive order early this month asking AI firms to voluntarily submit their frontier models for government review 30 days before they're available to the general public.

At the time, Sen. Bernie Sanders (I-Vt.) criticized the order for only being "voluntary," saying that mandatory testing and review of frontier models is needed to "protect Americans."

Jacki Thrapp contributed to this report.