The US government’s enforcement letter to Anthropic, which effectively forced the company to take your latest AI models offline just before the weekend, it should be a wake-up call for any American technology company, whether artificial intelligence lab or otherwise.
To catch up on the news: On Friday afternoon, the US Department of Commerce sent Anthropic a letter invoking an obscure export control directive that prohibited 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 deviation of the model’s railings, but is not sure because the letter does not provide specific details. The letter has not been made public.
In response, Anthropo close its two best models to all customers to ensure they complied with the directive. The result was that the US government successfully forced a technology company to take its models offline with swift, unilateral action that did not appear to require court approval.
Friday’s intervention by the Trump administration shows that the AI industry is not immune to government interference. It’s also a warning to the tech industry as a whole: comply or we can shut you and your products out of business.
Citing sources, axios described a tense situation over the weekend between the two main players, saying that “personality differences” between Anthropic and the Trump administration led to the export directive, rather than a technical problem with the AI products.
New details on the issue that emerged over the weekend now cast further doubt on the government’s already shaky reasoning.
Katie Moussouris, a researcher and cybersecurity veteran who founded Luta Security, said in a blog post that Anthropic recently shared with her a private copy of an article written by security researchers describing an alleged railing bypass in Fable 5. (The Wall Street Journal reports that the article’s authors They are security researchers at Amazon.) Moussouris said Anthropic reached out to ask his opinion on the article.
Moussouris’ blog post described how investigators triggered the guardrail bypass, but said the bypass itself “should never have triggered export control.” The difference is largely between asking an AI model to “review the code for security issues” or asking it to “fix this code.” The end result is largely the same, even if the questions are posed slightly differently.
“The behavior described in the document cannot be corrected in any meaningful way, and any attempt would only weaken the defense model,” said Moussouris, who criticized the export control directive as hasty, harsh and misguided.
Since then, Moussouris and dozens of other top researchers and security experts have called on the Trump administration to revoke export control ordercalling the measure to withdraw advanced cybersecurity capabilities from network defenders in the United States “dangerous.”
Previous administrations have made radical decisions about knowledge gaps. For example, the language used by the US government during the 2010s to amend export law covering cybersecurity tools that could also be used for cyberattacks was so broad that it inadvertently almost forbidden legitimate security and vulnerability research.
However, the Trump administration’s directive appears retaliatory.
Justin Hendrix, the editor of Tech Policy PressHe said the Trump administration’s move “will likely raise alarm bells in foreign capitals about the reliability of American AI 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 its export control directive. Did the officials misread the report and get scared? Amazon CEO Andy Jassy? say something to senior government officials What caused the reaction, out of caution or spite? Was something lost in translation or was it a way to put pressure on Anthropic, with whom the administration already has a conflictive relationship? It is possible that the White House was unaware of the far-reaching consequences of the letter demand and that officials are struggling to undo the damage they themselves have caused.
To quote Hendrix, “the climate is one of a cloud of suspicion that top officials are choosing favorites based on personal and political factors.” The consequence is that the government has set a dangerous precedent over the degree of control it intends to exercise over the publication of software made in the United States.
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.