Anthropic officials and industry experts say Trump’s model intervention is targeted and unreasonable

The battle between the Trump administration and anthropic artificial intelligence has been raging for months, and its mercurial nature has made it one of the term’s most difficult debates to follow, even for those of us in the tech news world. But after last Friday’s surprise de facto ban on the company’s latest models, the Fable 5 and Mythos 5, many are wondering if the government’s intervention is crossing the realm of hypervigilance into more nefarious territory. In fact, according to recent reports in New York Timessome Anthropic employees are downright convinced that this administration is actively trying to hurt their business.

The company, led by Dario Amodei, was initially brought into Trump’s orbit late last year with a Defense Department contract offer as part of the War Department’s rollout of an « AI-first » military strategy. The apparent start of things going off the rails came when Anthropic asked for minimal regulation and some guarantees about how its technology would be implemented — mass home surveillance and autonomous killing were excluded. The obstacles to negotiation presented by these crumbs of ethics were clearly an interruption of more let it happen deals made by their industry peers and ranked Pete Hegseth.

Soon after, in an unprecedented and seemingly punitive measure, the Pentagon designated Anthropic as a « supply chain risk. » That designation prohibits any « contractor, supplier or partner doing business with the United States military » from doing business with Anthropic, according to Hegseth’s subsequent post on X. Despite this alleged security risk, Hegseth also made it clear that Anthropic will still have to spend the remaining months of its contract providing services to the Pentagon until a « better and more patriotic » replacement is found.

As lawyers for both teams challenged the legality of the blacklist in court, Anthropic continued to develop its models and gradually its relationship with the administration became less adversarial. Or so it seemed until the June release of Fable 5, Anthropic’s publicly available version of the Mythos 5 model that powers Palantir’s Pentagon effort, albeit with cybersecurity railings. But just days after that launch, the company was forced to disable its Fable 5 and Mythos 5 models after another White House order.

Citing vague concerns about jailbroken versions of the models falling into the hands of « foreign nationals, » posing a threat to national security, the federal government issued a so-called « export control directive » to Anthropic at 5:21 p.m. (ET) on Friday, requiring them to prevent any foreign access to the models in question. But with paying customers around the world and so many Anthropic employees who are themselves foreign nationals, the only way for the company to ensure full compliance was to agree to the requirements and remove everyone’s access, rather than sifting through its users one by one.

Anthropic says they were just as shocked as their users about the government’s directive, and argue that the ban was prompted by overstating narrow, known vulnerabilities — none that come close to justifying such an intervention.

Before reissues appear, the administration wants to see versions of models where the guardrails clearly cannot be bypassed. According to reports in WIREDa number of experts doubt that this hypothetical impregnable model is even possible, and believe that all cybersecurity barriers are merely temporary loopholes waiting to be improved once their inherent weaknesses are inevitably exposed.

But as the opposing powers continue to meet for talks that the president says are « going well, » an ugly precedent is being set. Administrations have always picked horses, but harassing Anthropic for daring to ask for some of the regulation the AI ​​industry desperately needs is something else entirely. If defending against foreign threats is the real goal here, why did the administration’s recent crackdown on Anthropic force officials from their rival, Open AI, to come in to defend them? More importantly, why do these decisions, ostensibly for the good of the nation, always seem to leave America in a more isolated, insecure state than before?