The US government’s enforcement letter to Anthropic, which effectively forced the company to take its latest AI models offline just before the weekend, should be a wake-up call for any US tech company — AI lab or otherwise.
To bring you up to speed on the news blitz: On Friday afternoon, the US Department of Commerce sent Anthropic a letter citing an obscure export control directive that barred non-Americans, including Anthropic employees, from accessing Fable 5 and Mythos 5, citing an unspecified national security concern. Anthropic said it believes the letter is related to a bypass of the model’s guardrails, but isn’t sure why the letter doesn’t provide specific details. The letter has not been made public.
In response, Anthropic shut down and its two flagship models to all customers to ensure it complies with the directive. The result was that the US government successfully forced a tech company to take its models offline in a swift and unilateral action that did not appear to require court approval.
Friday’s intervention by the Trump administration shows that the artificial intelligence industry is not immune to government interference. It’s also a warning to the wider tech industry: comply, or we can shut you down and your products.
Citing sources, Worthy described a tense situation over the weekend between the two major players, saying that “personality differences” between Anthropic and the Trump administration led to the export directive, rather than a technical issue with the AI products.
New details of the matter that emerged over the weekend now cast further doubt on the government’s already shaky reasoning.
Kaiti Mousouri, a veteran cybersecurity researcher and founder of Luta Security, said in a blog post that Anthropic recently shared with it a private copy of a paper written by security researchers that describes an alleged bypass of a guardrail in Fable 5. (The Wall Street Journal reports that the paper’s authors are security researchers at Amazon.) Moussouris said Anthropic reached out to ask for her opinion on the paper.
Moussouri’s blog post described how the researchers triggered the guardrail bypass, but said the bypass itself “should never have triggered an export control.” The difference is largely between asking an AI model to “check the code for security issues” and asking it to “fix that code”. The end result is largely the same, even if the questions are asked slightly differently.
“The behavior described in the document cannot be substantially corrected and any attempt would weaken the model for defence,” said Moussouris, who criticized the export control directive as hasty, heavy-handed and misguided.
Moussouri and dozens of other top researchers and security experts have since called on the Trump administration to withdraw the export control order, calling the move to withdraw advanced cybersecurity capabilities from US network defenders “dangerous”.
Previous administrations have made sweeping decisions about knowledge gaps. For example, the language used by the US government during the 2010s to define export law covering cybersecurity tools that could also be used for cyberattacks was so broad that unintentionally, almost outlawed legitimate security and vulnerability research.
However, the Trump administration’s directive appears to be retaliatory.
Justin Hendrix, o editor of Tech Policy Presssaid the Trump administration’s move is “likely to raise alarms in foreign capitals about the reliability of American artificial intelligence for critical applications.” The message is that AI companies in the United States cannot be trusted to operate without interference from the US government.
The Trump administration has not confirmed why it invoked the export control directive. Did the employees misread the report and freak out? Did Amazon CEO Andy Jassy say something to senior government officials that prompted the backlash, out of wariness or hatred? Did something get lost in translation, or was this a way to pressure Anthropic, with whom management already has an uneasy relationship? It’s possible the White House was unaware of the far-reaching implications of the letter’s demand, and officials are scrambling to undo the damage they caused themselves.
To quote Hendricks, “the climate is a cloud of suspicion that senior officials are picking favorites based on personal and political factors.” The upshot is that the administration has set a dangerous precedent regarding the control it intends to exercise over the circulation of American-made software.
This time the government disagreed with Anthropic. tomorrow it could be with anyone else.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
